{"_id":"q-en-vscode-9d44a6a46c7e70592c3cbbe4edc499fd1aacb4d06549a94f11919cc1b0aeaee9","text":"If this key is set, all telemetry information generated in a debug session of the corresponding debug type will be tagged with that key (so that the data is routed to the correct bucket). what would be a good name for that key? (it should match the name used in other places where this key is used).\nthis is probably something you want to use...\nThe functionality for reporting telemetry only supports application insights pipeline, so I suggest calling it something explicit as .\nwe are not using acronyms in our protocol at all, so not a big fan of , I would rather go with . I would put it as a top level attribute under in the . If this is not provided I would ignore the telemetry event so we never mix the adapter telemtery with our vscode telemetry. What do you think?\nyes, put aiKey at the top of . What's about our own debug adapters? Do we get our own keys for them so that their telemetry events go to vscode telemetry?\nWhat we do we with the current adapters telemetry? do we depend on any of them in our current active reporting? If not, let's put them in the standard first party extensions pipeline. Otherwise, we can use the same vscode key, if so I would want make sure that adds standard properties such as machineId and sessionId so the processing scripts can function normally.\nFor Microsoft owned debug adapters, wouldn't we want their telemetry in the same stream as VS Code so that one could look at all activity for the same session? Or are there other good ways to tie that together?\npreviously we were reporting telemetry events of all debug adapters to vscode. With the new mechanism we don't do this any more, so we have to make sure that the node-debug adapter (which is shipping with vscode) still gets reported in the same way as before. So I will set the aiKey."}
{"_id":"q-en-vscode-be4726a31d84ae83f1d505d34fa8f6fcf3e732019f96e874318980db109a194c","text":"opens the current directory, also opens the current directory - not the parent\nRepros on master. Also: opens opens etc.\ni see the bug and the fix is in - i guess a short term work around is opens the parent directory\nyes the fix should be in the next release, adding a slash is a good work around in the meantime.\nI hope something good happens here. cause the way things are going. pretty much saying microsoft has lost all control of windows"}
{"_id":"q-en-vscode-5798151fa7167a731663f33f538352f69e57179e599791d172c98110b6056027","text":"when a link, src or url... contains an underscore character ( _ ) , you can't distinguish it from a misplaced space character because of the default underline below the links.\nAppears to be impractical to solve using in Chromium until it fully adopts the full CSS 3 spec on . What I'll do is use a instead. Tested locally, and it looks good to me: <img width=\"454\" alt=\"screen shot 2016-03-30 at 6 04 54 pm\" src=\"\"\nFixed via"}
{"_id":"q-en-vscode-8644cad47eb155299bb9ad12f53bc70ef56522af28f06684cdb21b4c2ab9435c","text":"VS Code should optionally be able to sign git commits with a given GPG key, as it is a good practice for authorization nowadays.\nDoesn't GitHub's authentication basically do the same thing, since you need to auth with GitHub to push to repos you have access to?\nGitHub auth is NOT related to GPG-signed commits if I know right.\nBasically, I would like built-in functionality for appending the command line argument to , as shown .\nSure, but isn't what it's accomplishing done by GitHub account/email linking? You need to link your email to an account and need credentials to the account to push.\nGPG can be used in any git environment (for instance: GitLab, BitBucket, etc.) I though this could be actual after seeing the .\nthoughts?\nI'd argue that for any given repo, you'd either want to sign all commits or none. Based on that, I suggest to use the git configuration in your repo, , which will make git sign all commits by default. Does that work for you\nI use exactly that, but when I try to commit from VS Code with that config attribute set, I get the following error output in the Git console of Code:\nI misread the issue, I thought you were talking about the vscode repo starting to sign commits, not adding the feature. while looking into signing commits yesterday, this I came across this\nNot just vscode is affected: If you're using an agent: As a workaround (a configuration change affecting not just vscode) it's possible to add to your For example, run Snippet from explaining : From what I've noticed, is now less verbose (doesn't remind me that I \"need my passphrase\" on ; is completely silent on ). I don't know what else this affects and thus I feel bad for suggesting this, but it works. If you're not using an agent: Again referring to Launchpad, you seem to be out of luck.\nThanks for that Does it solve it for you,\nWell, having listened to I'm not signing my commits anymore, but thank you for all the help provided!\n. It is not strictly accurate that signing each commit is \"totally stupid.\"\nit would be nice if vscode gave me the opportunity to sign a commit.\nSigning commits is not \"totally stupid\" and, in fact, has very real security implications. Not all Git code happens to be \"pushed\" to Github or BitBucket. Many places use git internally for all sorts of things and git commit signing is a very needed feature and often a requirement. I am using a gpg-agent and still am unable to sign commits in VSCode, for what it's worth.\nI found the two places that need to be modified for this to work, but wasn't able to setup the vscode repository properly (not able to run tests, linter hangs, unable to run electron) so I just dropped it. I'm leaving the links to the lines here in hope that someone will pick it up, and soon Funny thing is they both do the same thing, and they are both nearly identical, but one is written in async/await and the other uses plain function arguments instead of an options object. Is there a reason for this? Why not just abstract it out as another module and use it through that, would be less hassle maintaining it (and any similar parts that can be solved via the same solution).\n^ seems that git support is divided into two parts - one is git workbench that's in and second is git extension (that's built-in?) Is there any progress on this?\ni dont there is any work going on yet. I will pick this up but not until this weekend\nI like to confirm I am starting work on this. I will a pull request with a WIP title to show progress soon\nHello here, I recently tried committing with set to . However the issue I encountered is related to my name: it contains an accented 'e' (é) and right now VSCode seems to have some encoding problem (at least with git): ! Which prevents git from detecting the correct GPG key.\nHi! Any progress updates? :)\ngot all the changes needed mapped now going to add menus items etc should be done soon just lots of test and things to write\nI have the change almost complete just tidying up and getting ready to commit.\nPull request Added support to GPG signing hope all is oke\nwould love to see this.\nMe too, it was to the June milestone that got me excited but now its back in the backlog.... So I don't know whats going on. The pull request is this one:\nPlease, that comment, which is purely subjective, is very old and it's quoting achieves nothing, but FUD. Impersonating authors on GitHub is easily done and signing ALL commits is good practice. Especially with hardware tokens, such as Yubikeys. I mean no harm with , I am merely showing how easy it is to submit a commit by another. I feel GitHub could do more here too. On the main page, images are small and easily distracted (I could even change mine, I guess ... and my name). But the \"Verified\" means we've signed that commit. I believe it's important, as others seem to have commented also. !\nI verify ALL my commit is important and we do use yubikeys here at my organisation\nGPG signing provides authentication on the commits themselves. What you said in and subsequent comments are irrelevant to the issue. Because without signing the commit, I can easily fake you as the author and submit a fraudulent patch (e.g. intentionally infringing copyright of your rival company) via other channels (e.g. email) to another committer, who will be having a hard time to know whether you actually wrote the patch or not. And if that committer push the CL to GitHub, GitHub will happily link that to your account. See this commit for example:\nHi all, no need to try to convince me anymore\nYEAHHHHH!\nThis error git commit --quiet --allow-empty-message --file - --all git show gpg: cannot open tty `no tty': No such file or directory error: gpg failed to sign the data fatal: failed to write commit object can be cause by password protected gpg key.\nsomething related to the error you seen:\nI got it to work for me, but only after signing a previous commit (~5 minutes prior) manually through the integrated terminal. I have a hunch that the root cause of the error message , is that VS Code is requesting to open the prompt, and failing to do so. If you already entered your password via prompt previously, VS Code signs the commits perfectly fine.\nYou'll probably want to use a pinentry program that remembers the pass code or prompts graphically. (sorry if terse) On Tue, 17 Oct 2017, 08:33 Jeremy Loy, <:\nusually you see this problem when you dont things set up properly for gpg signing. that has been my experience\nIn Linux, I solve /dev/tty problem by adding binary exec of gpg2 path to git global config"}
{"_id":"q-en-vscode-3232970587aa16d34ea60e05b9c6b73f51948efe29a63d0089986357f575787d","text":"Integrated terminal master issue: Open question: Currently pressing <kbdesc</kbdwill hide the terminal, this can be a problem for some terminal programs, for example . Perhaps this should be configurable?\nThinking about desktop terminals, such as iTerm, I can only close them via a global shortcut, an button or running in the shell. <kbdEscape</kbdshould not bubble outside to the workbench.\nMarked as candidate as it's a very simple change that's been reported a bunch of times.\nCherry picked to release/1.2"}
{"_id":"q-en-vscode-d6288e676c13a8d6389645d2d3b2a858f2509dcc63ea670e2116e47797a645a0","text":"I am using VSCode on Ubuntu 14.04 x64 VSCode About: Version 0.10.10 Commit Date 2016-03-07T10:30:20.458Z Shell 0.35.6 Renderer 45.0.2454.85 Node 4.1.1 Steps to Reproduce: a git repo folder a change the \"GIT\" tab typing a git message I am thinking of a counter starting with 50 (The number of character that'll be used as the short message appearing in git commits), it helps write meaning full messages, without have a word cut and completed in the description part, the behavior is taken as is from GitKraken\nMy suggestion would be to, instead of always showing a counter, when the first line's length has surpassed 50 characters. I'd gladly take this as a contribution PR! Wanna give it a try?\nIt seems easy to implement this, can I give it a try?\nAbsolutely, looking forward to it.\nIt unfortunately seems as if is no longer working on this.\nPlease bring it back!"}
{"_id":"q-en-vscode-181f61b6239acba66352b3e89967487fa755c468ed8d30e6341c71c5ee3218df","text":"when i want to compare a local file with a online resource file, i need to save the online file in project folder, and then do compare. it's a little inconvenient. It'll be great if i can copy the content of the online file, and right-click the local file, and choose \"compare with clipboard\".\nAn extension would be ideal for a feature like this.\nfyi\nIt would be great too if also is allowed to compare selection with clipboard.\nIt would be great if we could compare unsaved files...\nI would be open to a PR here, I think it is not a lot of work. Some pointers: register a new text document content provider to be able to open the left hand side of the diff editor to be the contents of the clipboard (pointers: ) register a new action to diff the current editor with the resource that can be handled with from the text document content provider (similar to ) Due to sensitivity of data inside the clipboard it would probably make sense to ask the user for confirmation that the clipboard is accessed. Related to\nHas someone taken up this issue? I am a complete beginner to the VSCode codebase, would I be able to contribute to this issue?\nnot yet, but it seems has started a similar PR in which could be used as input to this one. The difference is that the contents of the left hand side are coming from the clipboard instead of the file on disk."}
{"_id":"q-en-vscode-1258bf932f264a3ab40bb3716a2b2dc7ec7f20005585c52c48abb70395819afa","text":"VSCode Version: Version 1.3.0-insider (1.3.0-insider) OS Version: OS X El Capitan 10.11.5 Steps to Reproduce: on the non-link area of a code lens line that nothing at all happens Would really like it to at a minimum to focus the editor (if not already focused), but we really love it if it move the cursor to the previous line. I find myself VERY often clicking on that area to start typing above it and nothing happens.\nI'm assuming (maybe) that this wouldn't be a difficult change -- so if someone had the time to point me in the vague direction, I would be happy to attempt to open a PR for this. :)\nPR welcome! :) It is in I would extract the code that does the focus: to a closure called inside the method. Then, at you can call and then : The reason focusing is complicated is because IE has some weirdness around redirecting focus when coming from the URL bar and doing a mouse down. And the already is a good indicator that the view zone does not want to handle mouse events."}
{"_id":"q-en-vscode-09f5e117e81d9e5bdec091f4c997089bfb0b4798d3891053b600f2d1bbc6fea6","text":"VSCode Version: 1.5.0 OS Version: OS X 10.11.6 Steps to Reproduce: this snippet to use Then the example Focus is on the newly created terminal and not preserved.\nBug for"}
{"_id":"q-en-vscode-c802ada0b44750ffacdefb120b0549a0bdb3a04531582e8bd12fd7b5f07e79a7","text":"VSCode Version: 1.5.3 (commit ) OS Version: Ubuntu Linux 14.04 Steps to Reproduce: some code that prints the characters to the output window. are incorrectly turned into a newline. F# Code: Output: This has already been reported as , but that bug was incorrectly closed as \"working as designed\" because it was assumed that the code in (I couldn't trace it further back than that). The character sequence is not the ANSI escape for a linefeed/newline; in fact, , since ANSI just uses the single character for that purpose. It's pretty clear, looking at the code, that that's what was the regex was intended to be set to in the lines just above the : it is set to . However, is actually an invalid Javascript hex escape. The specifies that must be followed by two hex digits, because otherwise the sequences (U+0007 BELL) and (U+000C FORM FEED) would be ambiguous. So is an invalid sequence, and is instead interpreted by Javascript as the character followed by the character (the backslash is swallowed). So the correct fix to bug is actually to replace that regex with . Then one could notice that the code immediately below is replacing all occurrences of with , which is a no-op since is . So the regex and the line serve no purpose but to slow things down by spending O(N) time doing a no-op regex search. They can safely be removed without affecting any other code, and this would actually speed things up a fraction.\nfyi\nYup, we should look into this and fix\nmight you be interested into doing a PR that fixes this since you have already done this great analysis? If not, we can follow your analysis to do the changes you proposed. Thanks either way!\nSure, why not? Here's the PR. Do you need me to sign a code licensing agreement of some kind before it can be merged in? If so, let me know how. UPDATE: Oh, I see that MSBOT replied on the PR's comment thread. I'm not surprised that no CLA is required for a commit that removed two lines and none."}
{"_id":"q-en-vscode-51f4c7e71ecebd0fc42cb014aa5cd2b6cf513272a8888e632bceb598640b9941","text":"VSCode Version: 1.6.0 OS Version: windows7 Steps to Reproduce: Create a vscode extension project. Pull off the cable to keep offine. Launch Extension (F5) I cant set breakpoint in my source file. And I find some warn information in Extension Development Host Window : I change the code in Its works. fyi\nSounds like a potential PR fyi\nYeah, PR welcome.\ncan you elaborate why your change is fixing this? If I look at the docs of I can see that host per default is .\ncan`t be resolved when being offline. And fyi\nthanks"}
{"_id":"q-en-vscode-49a6e04313299b638bdf063b63202f9ab857773dfc5e59e34f2b8e381cfb762f","text":"When you hover over a type in a TypeScript file, the tooltip that appears has a very low max width, making it difficult to read long type signatures.\nI would include a screenshot but it's not easy to take a screenshot keeping the tooltip on screen!\n! Errors are quite unreadable if the width does not adapt to the font size.\n! it would be nice if the tip used the available space instead of wrapping at an arbitrary point in c# too\n!\n! Same problem using rust's language server.\nGeneral styling control would be very helpful. Also line-height. Another approach would be to have a bottom panel that displays the full error and the tooltip displays a shortened version.\n!\n! When using the auto import feature it is often impossible to differentiate whats what as the import path is cut off due to the limited width of the suggestions window.\nIs this something we can adjust with custom css? It makes snippet infoormation very difficult to read as well. !\nSame in Python. We have to jump through hoops trying to flow doc nicely. And when it appears OK on Windows, it breaks on Mac... ! And it is inconsistent. Ex, hover is OK ! But completion list decides to wrap. What gives? ! Looks like hover respects markdown while completion item doc is plain string and gets wrapped randomly.\n!\nSomeone say long type signatures? !\nHere is a workaround by hacking the file \"C:Program FilesMicrosoft VS Coderesourcesappoutvsworkbench\" Simply append the code above after the original content of file, and restart VSCode. (Title bar may alert a warning message \"Not Supported\" and hacked changes would be overridden by updating VSCode) !\nIt would be nice to have some control on the width. The above does work but vscode presents the message your installation may be corrupt on restarting.\nInstead of hacking the CSS directly as mentioned above, I recommend using the to just override the current CSS. That way when VSCode updates, you can just reload the custom css file that you have specified (personally, I have a project where I hold configurations for my shell, Spacemacs, etc. and this custom CSS file goes there). Until it's officially supported, hacking the CSS is likely the best route. The suggested widths above were a bit much for me, so I settled on the following values for now:\nWhat is preventing the devs from providing a configuration option for this? This issue has been open for almost 2 years, and it bugs me on a daily basis. Would love to see it fixed.\n(since you seem like the most relevant person to ping) is anything likely to be done about this?\nI would love to hear an explanation from devs too. What are the difficulties of implementing this? I imagine it's settings a dynamic max-width and adding some more lines to add it in Settings. Are there other design blockers?\nWill you take a look at this? In 3 days it will be a 2 year anniversary for this simple feature request.\nAssigned to and as I remember we are going to have some UX improvements for editor tooltips and so on.\nmaybe now is the time to take this on? current iteration tries to solve parameter hints ux and this is not that much different\n! I employed CSS and put together full instructions: [Custom CSS and JS Loader extension][1] macOS VS Code: VS Code Insiders: Linux: : [1]: Restart VSCode Ignore \"VSCode is corrupt\" error You can choose to suppress this notice forever Run command \"Reload Custom CSS and JS\"\njust as a disclaimer, if anyone modifies the CSS/JS of the app, it comes up at “corrupted” in diagnostic checks, and you’ll typically get pushback if you report issues and include diagnostic information. Or, at least, I received pushback when I had custom CSS/JS enabled, and wouldn’t receive help until I turned it off (even though it was harmless/not related to the issue I was having). While this is indeed an okay workaround, we should just keep making our voices heard so that this is natively supported.\n100% agree. This is just a workaround. It's hard enough to read TypeScript lint as it is! :)\nCurrenlty using VSCode 1.30.2, followed instructions step by step, but it's not working. I don't even get the VSCode corrupt error. I'm running Windows 10 (V.1803). Edit: It was my mistake. For anyone who's having trouble, make sure to run Enable Custom CSS and JS from the command palette. Simply restarting won't work.\nThis issue seems to be fixed in 1.32.0-insider (1.32.0-insider) . I hope this is an intended behavior. I want to thank the developers.\nThis change is related to . I confirmed that the max width of is still small as mentioned above. <img width=\"867\" alt=\"2019-02-22 17 06 19\" src=\"https://user-\"\nUpdate: in , only the hovers for problems and the hovers with code blocks become to have larger width. Others have smaller width.\nHello there,your solution is great,but I want to ask how to get rid of the word at the title.As sort of OCD,sufferring from that. thx a lot!\nThis extension was created specifically for solving this:\nThis has been a long standing annoyance, with an issue that's like 2 and a half year ago. It shouldn't take much effort to have a config property that overrides the width/height of the tooltip, surely? Like, 10 mins of effort? <img width=\"525\" alt=\"Screen Shot 2019-07-18 at 6 41 15 PM\" src=\"https://user-\"\nif it's really 10 mins of effort, then I'm sure you can submit a PR for this? I definitely can't do this in anywhere close to 10 mins, so it would be much appreciated!\nSeriously, I want instead of this: ! At least something like this (with syntax highlighting it will be even better): !\nThere's already been a PR. I really don't understand this. I figure most VScode devs develop VScode in VScode, are they comfortable with the current state?\nfyi, it would be great if the hover could be resized\nHow is this issue 4 years old and such a basic thing still isn't implemented?\nIs there any word on whether we could see this soon? The suggestion widget has had resizing enabled since v1.51 in October, 2020. Is this tooltip widget considerably harder to add this feature for or are there additional considerations to be made? Whether it's a click to resize, or a manual adjustment in settings, could we have an update on the status of this addition? Opening new editor tabs just to properly read code annotations for some more descriptive components on my large monitors does feel like an unnecessary extra step.\nmight be very related and close to what you did here:\nHi, Instead of configuring the max width as the title says, could the tooltip hover be resizable ? (Also check for more info on what was done with the suggest widget)\nIt's sad to see that it's still not a thing.. I would really like to adjust the height of hover description tooltips.. The big brother Visual Studio does a much better job already. Let VS Code catch up a little..\nanother possible view for this issue\nIt is 2022/11/11. Christmas is close and I made sure Santa knows what I want!\n\"Excuse me, it's 2023\"\nThis one will be my favorite. Copilot suggestions and TS errors just can't be seen in this tooltip. It's time to fix it.\nIf this is still open and no one is looking at it, I'll take a crack at it. Feel free to assign it to me if that's alright.\nHey all, thank you for your patience. A PR has been made which is currently under work but nearing its completion. You may find it here: Once this PR is merged, you will be able to resize content hovers within Insiders.\nWould you add \"Reset Hover Widget Size\" command in Command Palette? Like we have for suggest widget: \"Reset Suggest Widget Size\". It would be good if we could reset the size to the initial one.\nThank you very much. 1 Is it correct, that this would not stay across different sessions? 2 The default max width is very small, is there any consideration, of increasing that?\nWill this feature also allow us to define the window size of a hover provider in a extension above the current limitations?\nHi if is is decided that the hover widget size should be configurable, there will be a command to clear the persisted sizes. This feature is still under discussion. Hi depending on the implementation, it is possible that the persisted sizes are persisted or not across different sessions. This is still under discussion. Currently the implementation is such that the maximum natural width is the same as before. Hi the current implementation does not provide this functionality. This could be addressed in a separate issue as it is related to the usage of the hover widget by extensions. The current PR that is worked on is the following:\nThanks a lot. And do you consider increasing the default size, that is unfitting to most real world scenarios? The windows is so small, that I feel the need to increase it even on my WXGA Laptop.\nHey all, thank you for the patience. A PR has been merged in order to make the content hover resizable. The change should be visible in tomorrow's VS Code insiders version and in the next stable release which will happen in the coming two weeks. There currently is no mechanism to save the content hover - we are discussing how best to approach it. You may track the issue on persisting the hover content here: we are currently not thinking about increasing the default maximum size. However it is possible that we implement the persisting mechanism as follows: when you increase the content hover size to a specific size, this size is persisted, and it is applied as maximum default size on the next hover appearances. This potential implementation is still under discussion by our team.\nJust want to say thank you you seem to have been tackling a lot of these high-level nitpicks that have been in the VSCode backlog for years, and for that I am very appreciative.\nI can't find a command to clear/reset hover widget size, is this has been ?\nHi currently the command does not exist because a persisting mechanism has not been put into place as of yet. It will appear when the latter is .\nOh okay, thanks a lot"}
{"_id":"q-en-vscode-6ce764baec8ad743ebb780126df5ac7570542fdcd4cafc1243929529f79fb34f","text":"VSCode Version: 1.7.2 () OS Version: Windows 10.0. Steps to Reproduce: a file a sentence of text such as Lorem ipsum sit dolar amet. your keyboard cursor after <kbdTab</kbdExpected: Spaces or tabs to be between and Result: word gets converted into an XML tag and auto closed, leaving us with <Lorem</Loremipsum sit dolar amet. !\nThis is coming from Emmet. To disable this, add: to your settings. Although this is currently the expected behavior, I think we should probably disable emmet expansions in markdown by default because this behavior can be annoying during normal writing. Any thoughts?\nthis behaviour is a consequence of , which has as the goal to not hard code in which language modes emmet is enabled for. As a heuristic uses the grammar's scope name hierarchy as a hint. This heuristic has worked pretty well. The scope of the markdown grammar is . This means Markdown is treated as a specialization of and since emmets supports , is enabled in Markdown as well. According the emmet doc there is no special syntax profile available for Markdown So a quick fix would be to: hard code an exclusion for markdown in #isEmmetEnableMode change the scope name of the markdown grammar, so that it doesn't inherit from html, not sure what the consequences of this are. do you have other ideas how to improve the support for emmet in Markdown and avoid the surprise described in this issue?\nHello, I see the same way as you suggested above. But, perhaps, we can also disable Markdown in the settings, as it was for . If you decide to disable Emmet inside Markdown, it is better to first option. But my 2 cents: Markdown is a lightweight markup language and it presupposes the existence of HTML constructs within yourself. Thus, it seems to me, to disable Emmet inside Markdown is not correct. P.S. Small news: the creator of Emmet had already started work on the second version aka 2.0.0\nthanks for the feedback. I agree with your point that disabling emmet without some way to re-enable it, if someone really wants emmet is better than to totally disable it. Glad to hear about Emmet 2.0.\nCould just be in the default setting for ? That would stop most users from triggering emmet unexpectedly in markdown files, while still allowing users to opt back in to emmet for markdown if they wish\nthe idea is good. It has a small drawback. If someone defines in the settings, then they will not profit from this fix. However, if someone if familiar with this setting and then they know how to disable emmet in markdown, if they are annoyed by it. Bottom line I'm fine with this fix.\nOk, I've merged in which adds to the default list of excluded languages. Not a perfect solution, but it should stop many users from hitting this behavior unexpectedly"}
{"_id":"q-en-vscode-5e7e08a14faa525f594b7bc37537a21ab55d1ee07eff706c4f24e1bd30d78b8f","text":"VSCode Version: Code - Insiders 1.8.0-insider (, 2016-12-06T16:53:46.634Z) OS Version: Darwin x64 15.6.0 Testing The current message seems to general: \"For better performance exclude folders with many files.\" \"better performance\" is to weak because the typescript server basically shutdown and won't provide any of the advanced features. \"folders with many files\" is vague, it should mention JavaScript/TypeScript files that the user does not work on (like packages/libraries). could also mention that this happens in the js-, that would set the right expectation for the 'Configure Excludes' button and also inform the user how this can be done manually. Needs more thought: \"Too many JavaScript/TypeScript files: To enable advanced editor features in JavaScript/TypeScript files exclude large folders with source files you do not work on.\" /cc\nAlso, the status bar button might benefit from being specific about the language it refers to."}
{"_id":"q-en-vscode-24da38d9a5adcb48f948a10cfbaa6444c16c1762e26d2ed6a76470769b1bdcfd","text":"VSCode Version: Code - Insiders 1.9.0-insider (, 2017-01-20T14:58:38.006Z) OS Version: Linux x64 4.8.0-34-generic Extensions: TypeScript file: ! Press enter: !\nverified."}
{"_id":"q-en-vscode-baa33a969db93309ad24ace857b60cb24d4e2707f2969450e49218068e53549e","text":"Particularly on Windows, should be called to terminate the process. This will ensure the sockets are closed and the pty agent is shutdown.\nClosing in favor of"}
{"_id":"q-en-vscode-83fe556b46d99a255f5e92d534ca95c090415bf928d61e22a548f1eba17c4596","text":"<!-- Do you have a question? Please ask it on --- VSCode Version: 1.10.2 OS Version: Mac OSX (10.11.6) Steps to Reproduce: VS Code integrated terminal a relative path (eg. ) Description: Let's say you have a build system that outputs a stack trace on a build error or failed test. If this was run from the integrated terminal, the file paths in the output should be included as clickable links which can be clicked to open the file in VSCode. However, this only works if the path is either an absolute path (e.g. ) or a relative path that begins with \"./\". Many build systems and tools output file paths relative to the directory the command was run, without the leading \"./\", Karma for example. So the output would be something like: , not . The first is not matched by VSCode as an openable path, but the second is. I believe that what is determined to be a path comes from the Regex in \"\" Is there a reason why relative paths of this form (e.g. no leading \"./\") are not included?\nThe reason was because until 1.11 links aren't validated so there would be heaps of false positives. Performance was also a factor but it shouldn't impact that much really now that I think about it. Now that links are validated maybe we should be more lenient when linkifying.\nI'll assign this to April but if someone wants to do a PR before then you're welcome. You will need to: Modify the regex to allow paths that are not qualified with , , etc. Make sure that the new path matcher does not conflict with previously matched links and does not create links with adjacent characters. Eg. we want to avoid this:\nNice. I'll write a PR...\nThat would be awesome :+1:\nFixed in"}
{"_id":"q-en-vscode-8176b5ed55b97e1de970908ac64217ec71792d99eeae9b2b5668070c1707f6c8","text":"VSCode Version: Code - Insiders 1.11.0-insider (, 2017-03-17T08:14:15.805Z) OS Version: Windows_NT ia32 10.0. Extensions: ; Steps to Reproduce: a terminal and run \"bash\" with a text file containing some lines of text ctrl+v (or right-click) to paste some lines of text What's expected: The text should be appended before the cursor, and the existing text should be there after the cursor. What happens: Existing text disappears, and it seems some extra newlines are when pasting text. Note: there is no such problem when doing the exact same steps with This happens on insider builds, no idea if it happens with standard builds as well.\nMore information. Here is what happens when I try to paste this text function startagent { echo \"Initialising new SSH agent...\" /usr/bin/ssh-agent | sed 's/^echo/#echo/' \"${SSHENV}\" echo succeeded chmod 600 \"${SSHENV}\" . \"${SSHENV}\" /dev/null /usr/bin/ssh-add; } In the middle of a file containing this text: existing file with some text Here is what happens with (that's what expected: text is pasted in between existing text): ! Here is what happens with VSCode: text is pasted with extra newlines plus existing text appears in the middle of the pasted text !\nAny update on this bug ?\nThanks for the ping. I've validated it's definitely an upstream issue. I tried to fix it a couple of days ago but didn't get anywhere. I'd guess it's related to node-pty or winpty as different shells appeared to handle it differently. It also seemed somewhat inconsistent. No plans to invest further in fixing it just yet.\nThis is probably a bug with VSCode's JS xterm module, not a node-pty or winpty bug. I placed these two lines in an untitled editor window: I then set the environment variable , started VSCode, and copy-and-pasted the two lines above into VSCode's integrated terminal. This appeared in the winpty trace: The clipboard has CRLF-delimited lines, because we're on Windows. However, in the VT/100 input format, CR means Return, and LF means Ctrl-Return. Something (probably the JS xterm) is copying the clipboard contents into the pty, and that something should normalize LF and CRLF into CR. I described this bug in detail the last time I saw it:\nthanks! This is likely the culprit then which replaces all end of line chars with on Windows:\nYeah, I think your change will fix the issue. The function looks like it might also need a change. Maybe something like this?\nFYI I think you may have run into these issues before? It would be a good idea to test these changes in in tomorrow's Insiders build.\nI've certainly experienced aspects of this issue from the Python extension when running selections of code in the terminal.\nI was unable to reproduce the problem anyway, so have no way of testing this. But thanks for letting me know :-)"}
{"_id":"q-en-vscode-283d65110775a09080cf52d8bb48f0b24fdeeed3528f1cf19929463ce8467b94","text":"Chrome-debug supports .js files with query parameters (). But if I load one with \"Open loaded scripts\", the query parameter is stripped. This happens here: since fsPath doesn't include the query parameter. Any idea how to make this work? I don't think URI is expecting to have a URL with another scheme.\nThat normalize code should only be executed if the source could not be found - otherwise should not be and the should be passed to the adapter. Do you know why is that? As for the call - I am not married to it and we can change it to whatever uri massaging techinique would work for you and not break the existing behavior. If we change this I would prefer to be in debt week -april\nIt tries to get the source from , but it looks like that is only populated when it breaks in there, right? Plus it's a different format, the script from the quickopen starts with but the ones in don't.\nI don't know about replacing , would do it if that isn't too hacky.\nYes, is only populated once there is a call stack that contains that source - so if it breaks. Yeah that feels like a total hack. Why don't you experiment with different solutions and create a PR where we can comment and try out different approaches if needed?\nThis regressed from adding back the in DebugContentProvider.getTextContent. is that needed?\nAlso,\nI think opening the file by putting its url into a uri is very fragile, to say the least. I have listed some problems I've noticed while implementing a Loaded Scripts Explorer for the Firefox debug adapter in . My suggestion would be to add the ability to open files using the numerical that the debug adapter assigned to them or using the object that is used for identifying the file in the messages between VS Code and the DA. Whatever solution you adopt, keep in mind that VS Code should be able to determine that the file opened by clicking on a stackframe and the file opened by clicking in the Loaded Scripts Explorer are one and the same - otherwise setting breakpoints in a file opened through the Loaded Scripts Explorer will become problematic.\nI've addressed this issue by making sure that the flow -uri -preserves the data so that the initial and resulting objects are identical. The \"path\" attribute of the source object can contain anything, e.g. a file system path or a full URL. So if a Uri is constructed in the following way: the debugContentProvider inside VS Code will retrieve the content by sending a \"sourceRequest\" with a \"Source\" argument to the debug adapter identified by the debug session.\nVerify that Rob's Chrome-debug case from above works.\nIt is escaping the query parameter, but works anyway using the new loadedSources API, because then we can rely on sourceReference instead of a string path.\nI have also seen this (with the Firefox debug adapter) and couldn't find any user-visibile ill effects. In particular, the source url is displayed correctly in the tab opened from the Loaded Scripts Explorer (i.e. the query parameter is unescaped there and other escaped characters in the url remain escaped).\nAdded \"Verified\" label since both comments from above confirm that this works correctly now.\nIt does rely on me updating chrome-debug, in case anybody else tries this. I'll do that this week."}
{"_id":"q-en-vscode-dcbb752dcef949bb324e6994ecb1b58fed66f30b65ddf33b8ccfe3b3b4a464f3","text":"Variable substitutions that are available for (see ) do not include currently selected text. I am looking for something similar to $(CurText) that is available in Visual Studio Arguments for External Tools (see ). Please add a variable ${curtext} which would be replaced with the currently selected text in the editor window. If no selection is available, it could use the current line at cursor position.\nNice idea, but I think it's too late for implement that in the current milestone...\nFixed by in\nVerifiers: what happens if all text in a 300 MB file is selected...\ngood point. Maybe we should put a limit on this. Or only to copy the howl line at most (even though the line could also be 300mb I am fine with that) thoughts?\nSelecting lots of text, at some point I get . If there's a predictable limit where this will fail, we should just cut the string off there. Unless only sending part of the text could be dangerous or unexpected for the user. In any case I think that sending multiple lines of text could be useful.\nthere is not a predictable limit since the variables are used in different scenarios. If somebody has an idea for a limit I am open to PRs"}
{"_id":"q-en-vscode-abca1b7e345251a6aa9ed8c632f901f1d4393ca7ec52c2a6c314ce707f5c5e07","text":"Platforms Tested: Insiders 5-30 Code 1.11.1 Stable Windows 7 OS X Sierra (should be all platforms) Problem: Output panel drop-down does not track current active output if output channel is activated later , typically when an extension output Channel is active (eg Git). Output select box will show the default 0 index output (Tasks) Steps to reproduce: Open code with git repository/workspace select Git output Channel close code reopen code Git output will be active but drop-down shows Tasks Fixed with PR always update active select box selection with setOptions call on any output channel update Tested on Windows 7, OS X\nFixed via"}
{"_id":"q-en-vscode-d6fc19b69b473140b029ba8bce157de61018372bbedce63d12c06446b34603dc","text":"Testing Steps to Reproduce: integrated terminal and open the find widget the integrated terminal area very skinny bug The find widget starts overflowing into the explorer area: <img width=\"420\" alt=\"screen shot 2017-06-28 at 1 25 51 pm\" src=\"https://user-\"\nAfter , no more overflow but we now clip the find box: !\nI guess we want it to be tiny like in the editor: <img width=\"221\" alt=\"screen shot 2017-07-07 at 3 02 45 pm\" src=\"https://user-\""}
{"_id":"q-en-vscode-7f6aa7c3a38f736a1a34910fa0568796d9a62ff793bf7d5fd71a75265c2959ee","text":"I noticed that the terminal no sends Operating System Commands (OSC) like setting the window title. Right now the task runner filters them but this gets more and more complex and I am actually not an expert with this. So I would like to request that the terminal provides API to notify data without control sequences. Currently I used these two regular expressions to filter them:\nThere are two unknowns to me right now with this: Various control sequences can do , so where is the best place to gather this data? It might be worth seeing if adding listeners to will give what you need. How do this efficiently. Because a lot of data can stream through the terminal, we should probably only enable this when it's explicitly requested (ie. only fire the event if something is listening). It should probably also be batched. PR welcome upstream.\nFrom a problem matcher point of view it would be enough sending the line. We could sent it when a linefeed is . Then there is still the problem of a line content changing. We once discussed the idea of a line identifier to signal that. I agree that sending single characters might not be a smart idea in terms of performance.\nIf it's just on line feed then many characters could go missing, eg. \"abc\" <move cursor up a row\"def\" <linefeedwould result in \"def\" only. Maybe this is not something we care about, but it could lead to some obscure bugs if people relied on this new API.\nThis is a good example why something like a line model would be helpful: After looking into the issue it looks like that pester constantly 'repaints' the whole screen hence the terminal sending the same line onData over and over again. For me it is very hard to tell whether I have seen that line and whether I have processed it already (the same content). What would be cool if the terminal could provide a line model where is line has an id (number) and changes are signaled on that model instead of sending raw data.\nI actually an ID concept to the buffer when doing the selection stuff () but I ended up not needing it and pulling it out, instead the selection listens to when the buffer is trimmed and adjusts the indexes of its own model. It really didn't feel very good adding it as it introduced much more complexity to the already complex circular list and the overhead of an addition object for every entry - which is a lot when you need to create tens of thousands of these things in an instant. There were other issues that come up with this like IDs being interleaved due to programs having the ability to delete/insert lines so there's no way I could think of to optimize it nicely, leading to a slower, more memory hungry terminal. It would probably be easier and more performant to just deduplicate these errors on your end. For example you could record a hash for current problems and check it before adding new problems, or perhaps even better, record the raw terminal line of the error so you don't need to parse it again for duplicate problems.\nI implemented the deduplicate yesterday but it feels like a hack on my side :-) I need to create a unique key for something that doesn't necessarily have one. I still think having a line bassed model comparable to what the editor exposes would be a nice thing. Me giving the raw data is comparable to the editor signaling raw cursor moves and text changes. Instead I get a nice model with text edit events. And I am not interested in color information like I am not interested in this on an editor buffer.\nJust for the record. I can not filter on the line itself from the terminal without knowing if I have parsed that line already. Matchers are multiline aware and there could be two equal lines in one output on purpose since they difference is made by a line above. Me knowing this is implementing a line model on top of it :-)\nisn't the point of programs rewriting lines that the lines do change and that they may need to be reparsed? Here's one idea that might work for both of us: exposing some data stream you can listen to as in the original comment, but only ever pushing data if it's on the last line in the terminal. Would this cover your use case? For the PS case you would get multiple lines for the last line but maybe this could be more clever somehow in recording when the cursor moves and only reporting the line a single time (not by using IDs but by only reporting once after a call inserts a new line.\nyes, it would be cool to be able to reparse a line when it changes. As I tried to outline in this needs to have some sort of model to do it right. I can't simply decide things on the content of a line. So instead me implementing a line model on top of ANSI control sequence characters I was somehow hoping that the terminal could expose its. I my simple mind the terminal works like this: program -Output with control characters -line model -view model (does the wrapping, could do folding, ...) -DOM. I think the last line suggestion will work for parsing problems. However when exposing this as API in the extension host I think it might be a little to limited.\ncurrently there is only the line model, no view model. I did a big write up on introducing one to properly support reflowing lines here This has been a pretty long discussion across a few issues but to clarify where we are now, here are my concerns: Extra memory overhead in the buffer: I'm currently in the process of trying to push down the terminal's memory usage as I need to store more data against each character and memory usage right now is approaching unacceptable limits imo Here is a snapshot of a 160x24 terminal with 5000 scrollback filled which is not out of the ordinary, I typically have two of these running just in the window with Microsoft/vscode opened: ! Extra processing when data/lines are to the buffer: back in December I made a big effort improving the terminal's performance by attacking it in a number of ways. The main take aways from this experience was that trying to write a high performance terminal in JavaScript/DOM is very hard because processing and displaying data is so expensive. I can now safely say that is the fastest JS terminal out there by a long shot and I want us to maintain this status. A proper view model that can track lines that changes would amount to a significant amount of processing in the worst case which would be generally unnecessary for other use cases. Some possible solutions: a new event that you can listen to that fires when a line feed is to the last line, ie. only fire an event once per line and ignore lines being updated or lines being inserted. Pros: It doesn't involve storing any more data or doing anymore processing, other that pulling the string from the buffer. Cons: it probably isn't sufficient to cover all use cases. an ID to each buffer line () and fire an event when there is a line feed against a line or the cursor moves away from it (specifics here would need to be figured out). Pros: It's a more complete solution and covers more use cases. We need an addition JS object and number associated with each line, this probably doesn't add up to that much data as it's on the line rather than the char. Cons: Additional complexity within the data structure. I'm not so sure just yet when exactly we should fire these events. The IDs will be out of order in some cases which may be confusing to consumers. These events may still fire when the line did not change (eg. if the program/script erased and rewrote the line). a proper view model that throws events only when a line changes. Pros: I think this is the ideal solution from your perspective. Cons: All the cons of number 2. The amount of string comparisons needed here is out of the question in the worst case with my current thinking, unless this was disabled by default and generally did not run on a lot of data. Commentary on the solutions: doesn't look like this is sufficient for problem matchers could work, but I'm not sure it helps much with the issue in behind an option on terminal initialization is probably the best way to go, but it's also the most difficult and time consuming to implement. Work on this is blocked on the following active PRs: & , and it should be done in a way that accommodates future line reflowing\nthanks a lot for the detailed explanation. Having solution 2 will for sure make my life easier and will not make things worse on my end. Regarding option 3: I am not interested in the view model since I need lines unwrapped. The line model is what would work best for me assuming that the line model doesn't have any in there to make lines visually wrap .\noption 3 including reflow is required in order to get unwrapped lines Not sure when I'll have time to work on this, pretty swamped atm. It might be a good opportunity to collaborate if this is high priority for you?\nI will be out on vacation until 2. August. I think we need to address this sooner than later since we can't assume that problems always fit onto one line. An interims solution could be as follows: can we introduce a custom control sequence (e.g. a operating system control sequence) that describes the new lines inserted to wrap the text. I could then filter them on my end since I would be able to distinguish them from normal newline. Would this be something you can add in July then I would catchup with it for the July build ?\nI actually expect you to get the unwrapped lines currently, just wraps the lines when they are going to overflow , that's the only wrapping it knows about. On Windows though I'm not sure rows are ever marked to be wrapped (not 100% sure), at least on PowerShell. This is just how winpty works."}
{"_id":"q-en-vscode-2863ea5b18199b7351b1de33b471f5bcceda4e5ca2caaf5ec9fe04157dba9342","text":"Would be great to see a tooltip on links, showing the target. Or popping it up in the status bar or elsewhere, like Chrome.\n: Could you please explain the preview functionality bit more with screenshots of Chrome? I am new to open source community and this project. I use vscode often and would love to work on this if someone can guide me a little.\nLower left corner of the Chrome window when I hover the link to your name: ! But I think a normal tooltip would be easier.\nAh yes. I understand what you are talking about now. To verify, where the link is leading to. Seems like a beginner friendly feature. I can pick it up this weekend if anyone can nudge me towards the right direction. :)\nSorry for the delayed response. It seems that tooltips are only shown if the element has a attribute, but not for links that only have a . I wasn't able to find if there is an electron setting to enable tooltips for links without . I think we may have to implement this ourselves. To get started, take a look at this PR that a similar feature to another electron app: Here's where our webview code lives: I believe you should be able to attach a handler to the webview handler. Please let me know if you have any questions about the code or want to discuss potential designs\nI m happy to receive help! :) I was facing issues while trying to build and fiddle around with the source code. Will definitely look into this PR and the code you pointed out this weekend. Thanks for taking your time to post this info. I ll post any questions or designs which I have.\nHi , I will create a PR to add attribute to every .\nFor verification: this was implemented as tooltip when you hover over links in a webview"}
{"_id":"q-en-vscode-65b1961dbb461729afd8f7d9690c50a869984ddb402dbd471f5e3aec41d67cce","text":"<!-- Do you have a question? Please ask it on --<!-- Use Help Report Issues to prefill these. --- VSCode Version: latest OS Version: Win10 A little bit more space between last line from terminal and the blue line below would be nice !\nI 1px border on the bottom would be fine, opening to PRs. Who ever wants to pick this up, you need to make sure the size of the terminal is calculated correctly given the new margin. Relevant files: , .\nI'll pick this up. pretty much all of the other terminal issues are already taken, or out of my realm. just let me know somebody's not already working on this ;-)\nsounds good\nPR posted"}
{"_id":"q-en-vscode-e5f62ff3ed6ffcf657fa342f47cd1adfaf4ff9f2f6c62648651beb34b6d8ef35","text":"FYI I commented out the test. Please fix it and enable it.\nThe test is failing because the DOM is no longer filled with nodes so these tests can't work the way they did previously. Is it possible with Spectron to pull a DOM node object, including any custom properties it has? If so I could attach a reference to the object to the terminal element which would allow querying the terminal for this sort of information really easily.\nYes, you can access the DOM in Spectron and we do it in multiple places. For example refer to . This will allow you to execute java script code on the renderer process that has access to DOM and let you return the model you want. But please be aware that, Since it runs in a different process than the test process, console log does not work.\nPR with fix is out , will merge in after release is branched off."}
{"_id":"q-en-vscode-ad748d2c7885aae18593096248671fa2ea86757587487ef00e465f9d7c6413ad","text":"I admit that it's a niche scenario, but I think that what \"Open Changes\" shows for not-yet-staged changes to files that have been detected as renamed in the index is incorrect, at least inconsistent with how changes to files in the index are shown that have not been renamed. Steps to Reproduce: an empty directory and initialize a git repository in it. that directory create two files and with identical content: This exact content is somewhat important, as I've confirmed that it triggers git's \"rename detection\" (checked with git v2.14.1). and commit both files. the first line in both files. to . \"\": ! does the right thing for both files, that is, it shows the changes between HEAD and the staged version: : ! : ! the first line in both files once more, but don't stage these changes. still shows the right thing for (in this case the changes between the staged version and the file in the working tree): ! As the content of both files is identical both in the index and the working tree, I would have expected the exact same result for , but now the left side is just empty instead of showing the staged version: ! VSCode Version: Code 1.16.1 (, 2017-09-14T16:26:20.018Z) OS Version: Linux x64 4.12.13-1-ARCH\nHehe great catch! Would you like to to it?\nSo is this issue still open and up for grabs? If so I would like to try to help\nJust look above your comment, there's a PR for it already:\nFixed by"}
{"_id":"q-en-vscode-0f81fc2192f5ef8056a6be34f7a9e51e96ea361a92368717e5c9b1a67637f18f","text":"a .ts file tab to activate the \"import statement\" snippet Expected and actual: completes to a .tsx file (TypeScript + JSX) tab to activate the \"import statement\" snippet Actual: completes to Expected: completes to Is there any reason why the snippet is inconsistent between .ts and .tsx files?\nYes this is unintentional. I've marked this as help wanted in case anyone want to take a look. Here are the relevant snippet files:\nThanks! See I was also thinking it would be useful to add this to .js and .jsx extensions. What do you think?\nThanks for the PR! The change should be in the next insiders build Yes I think adding the same for js/jsx also make sense. Did you want to submit another PR that copies these snippets over to js/jsx?"}
{"_id":"q-en-vscode-57a9f64b254c68b9266597db911bac4b1697c673092e4bbf998b5ecf83ca22b5","text":"VSCode Version: Code 1.15.1 (, 2017-08-16T18:07:25.676Z) OS Version: WindowsNT x64 10.0. Extensions: ExtensionVersion ------ overtype0.2.0 python0.7.0 theme-dark-plus-contrast0.1.101 cpptools0.13.0 monitor-file1.0.1 (4 theme extensions excluded) Steps to Reproduce: there is such line in a *.bat file rem set SERVERVERSION=5.7 cursor on it, and uncomment with \"Ctrl+/\" expected: set SERVERVERSION=5.7 actual result: REM rem set SERVERVERSION=5.7 Looks like uncomment is unnecessarily case-sensitive. It should uncomment despite it is 'rem' or 'REM'. I don't care whether with next \"Ctrl+/\" it will add Upper-Case 'REM' instead of 'rem' - it is not problem. <!-- Launch with to check. --Reproduces without extensions: Yes/No"}
{"_id":"q-en-vscode-8f700d054040da68f93ea70aad931eb058224e384f8fc1cd983f31a324a5400f","text":"VSCode Version: Code 1.17.0 (, 2017-10-05T06:07:51.906Z) OS Version: Darwin x64 16.7.0 Extensions: ExtensionVersion ------ vscode-markdownlint0.10.1 xml1.9.2 code-settings-sync2.8.3 sonarlint-vscode1.1.0 html-snippets0.1.0 Bookmarks0.16.0 markdown-preview-github-styles0.0.4 path-intellisense1.4.2 gitignore0.5.0 vscode-eslint1.3.2 githistory0.2.3 jquerysnippets0.0.1 tslint1.0.9 auto-rename-tag0.0.14 Angular22.7.0 HTMLHint0.4.0 cpptools0.13.1 csharp1.12.1 indent-rainbow0.7.2 tsimporter1.2.14 java0.11.0 vscode-open0.1.0 stylelint0.29.1 code-spell-checker1.4.7 code-spell-checker-spanish1.0.1 bootstrap-3-snippets0.1.0 JavaScriptSnippets1.5.0 Steps to Reproduce: new CSS file to add -bdrd with vendor-prefixed as mention in emmet documentation adding a hyphen before. But not show any suggestion or auto-complation list. Expected result: Emmet Documentation: I'm not see in his documentation if there is a way or option to enable this behaviour: <!-- Launch with to check. --Reproduces without extensions: Yes\nYou are right, expanding the abbreviations so that all vendor prefixes are included is not supported in Emmet 2.0 Have you used the command? We could just overload that command to add missing vendor prefixes or have a new command to do the same.\nI'm not using it, but I think is better add it to a new command.\nNote to verifier: The vendor prefixes should be attached as documented in when using the command."}
{"_id":"q-en-vscode-6d0e3209813ec3ef92f7cc426dfd5c4f95758c31ad2b924c4602f424a3b0b0d3","text":"VSCode Version: Code - Insiders 1.19.0-insider (, 2017-12-01T06:07:42.474Z) OS Version: Darwin x64 16.5.0 Extensions: ExtensionVersion ------ beautify1.1.1 vscode-caniuse0.5.0 jshint0.10.15 tslint1.0.24 php-debug1.12.1 php-intellisense2.0.4 php-pack1.0.1 plantuml2.3.3 vscode-duplicate1.2.0 sublime-keybindings3.0.3 debugger-for-chrome3.5.0 laravel-blade1.10.0 Steps to Reproduce: a HTML file with references to javascript files over the path of the JS file in any <scripttag 'll get the tooltip \"Alt + click to follow link\" Well, there isn't an Alt key in my MacOS keyboard (stock Apple keyboard). So saying \"Option + click to follow link\" will make more sense for Mac users. Not a big deal though. Thought I'll just point out. :) <!-- Launch with to check. --Reproduces without extensions: Yes/No"}
{"_id":"q-en-vscode-1751b666f0ef016713ce3e5ebc0edf0f87c1e130308bad928ffb22b3cee78546","text":"VSCode Version: Version 1.19.1 (1.19.1) 0759f77bb8d86658bc935a10a64f6182c5a1eebaOS Version: macOS, 10.12.6 Steps to Reproduce: TypeScript React, and Reproduces without extensions: Yes.\nFor the case: where the user has just typed , returns false since returns . This in turn is caused by having the following language ranges for the string : where = and = . These language ranges cause to take as the current word, which does not match against the known word patterns\nAny update on this? The Ctrl-Space works and provides the correct autocomplete options for props with associated type. Would be nice if I didn't have to do that each time.\nHave the same problem, but wonder about: \"... autocomplete options for props ...\". Is VS Code capable to autocomplete my custom from an imported component?"}
{"_id":"q-en-vscode-d79b982faa6aabce964692bb48c4cf29159ee39a529551404b4a8102f80dad81","text":"In , the comparison functions , , , and have misleading descriptions. For example, says \"Check if is before this position.\" I think it would be more clear to say something like \"Check if this position is before \", as that is when is returned. Likewise for the other three."}
{"_id":"q-en-vscode-f16d9876d169a3ef5780822c398ced7a1bd7a5ca62da6f4778b94f9bfc07d9a8","text":"It would be nice to invoke current line codelens actions without moving hands of from keyboard to mouse. It could be done via context ctrl+. where codelens actions are or from ctrl+p, or even new shortcut key which opens context menu which contains only codelens actions. They are sometimes available if extension implements them that way that they are to menus/context etc but this probably could be handled more generic way in vscode itself?\nHi! I just opened a PR with the ability to trigger codelenses with the keyboard. Hope you find this helpful as I do :D I appreciate any suggestions and comments from people coming to this issue"}
{"_id":"q-en-vscode-eacaa740b6edc1c7ac518042e55993a6c501170d7ddf47fb4270a8a7318c156a","text":"With latest release (1.21) you can split terminal infinitely. This breaks badly the terminal area. Probably the IDE should limit this to a defined maximum or something?\nI considered adding a max value but you would always be able to break it then by sizing down. As a mitigation we could disallow creating terminals if the desired room isn't there but the real fix is proper overflow behavior.\nThanks , I would agree that proper overflow behaviour is the way to go. Thanks!\nI think the easiest way to improve this is to only allow terminals to be created if there's a reasonable amount of room. This can exist with our current lack of overflow behavior as it's an improvement on the current and will prevent 100 terminals from being created..\nSome terminal can be resized to a small size if they are not important. In some case, showing multiple terminals is just to notify the user whether they are running. Currently the minimum size of terminal is too large.\nI suggest the minimum size of a terminal should be an option in Settings."}
{"_id":"q-en-vscode-6f896230f9a9dafa8af58dedc226d948c3ed38395a7098edebd10440f5837460","text":"Issue Type: <bFeature Request</bWhen using a touch screen, I want to be able to interact with VSCode in an intuitive manner. When I touch the minimap, I expect the document to scroll immediately on touch-down, not on touch-up. Mouse-up makes sense for a mouse, but not for a touch. When I touch and drag the minimap, I expect the minimap to scroll with my finger. VS Code version: Code 1.22.1 (, 2018-04-06T02:26:57.615Z) OS version: Windows_NT x64 10.0. <!-- generated by issue reporter --\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.487 --- <!-- score: 0.464 --- <!-- score: 0.449 --<!-- potentialduplicatescomment --\nI haven't PR'd vscode before, if anyone who cares about this issue has tips on getting my PR ^^ merged and released I'm happy to put more work into it. I find it very frustrating not having touch-scrolling on the minimap, as I'm so used to it working in old-school Visual Studio!"}
{"_id":"q-en-vscode-08086a885ea729908af389c97d9e6c7dfedad50e36c4842725720aac22b2468a","text":"<!-- Do you have a question? Please ask it on --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.23.0-insiders x64 commit OS Version: Windows 8.1 x64 After update to latest version window is empty (only main menu bar is visible). Renderer process consumes 16-20% CPU but nothing is changing. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes\ncan you run \"\" from the command line and see if any output is printed?\nThe output is here:\nadding for the (maybe unrelated) spdylog stacktrace that shows up. can you share a screenshot how VSCode looks after it started like this?\n!\ndoes it make any difference if you run with \"\" from the command line?\nThat literal looks weird. It looks like the user has a problem in one of his variables, possibly . Whats the value of your HOMEPATH variable?\nVisually nothing is changed. In console also nothing. is not presented. Username is , but profile folder named as .\ndoes it make any difference if you run \"\" from the command line?\nYes. Window is opening like normal new profile. My profile is broken. Nice. ! UPD: Erased my old profile. Nothing. Same situation. So, trouble is Cyrillic symbols in path to profile folder.\nto verify that can you use the argument with a folder that has the Cyrillic characters and see if you get the same results?\nYes. Launch with Cyrillic characters in path give same result.\nit would be easier for me to reproduce if you could post the exact Cyrillic letters that you are using in here as text so that I can copy it.\nHere it is: .\nI can reproduce and the reason the workbench is not coming up is due to spdylog crashing very early. To reproduce run \"\"\nJust to confirm, you could reproduce by using cyrilic in a path provided to ?\nI can. Command: Result:\nyes\nI can confirm I have an identical problem. My username contains the character 'ø', which now makes the insiders build fail to start. Changing the to a path which does not contain that character fixes the issue. However, the Windows context menu action does not work, as the argument is only provided on the shortcut. The stable release (1.22.2) works as expected. EDIT: I fixed the context menu issue manually in the registry, so everything is back to normal for now :)\nI pushed a workaround for now to not crash so badly when we fail to instantiate the spdylog writer. Turns out we already had code in place to protect from failing loads thanks to Alex, but the output channel was failing very early on. fyi maybe you want to quickly check my changes via The real fix probably needs to go into spdylog, pushing to May since this was already happening in stable.\nThis will break Output channels when not writing into file. When user switches between output channels, they will become empty as there is no way to retrieve content now\nI was falling back to old way of writing into channels if fails. I removed this fallback this milestone . You can get that back. But I would really want to remove this fallback because for further enhancements to output channel which depends on spdlog.\nthis is not a real fix, I would expect some regressions. Previously VS Code would not start at all without my workaround. Do you think the current behaviour with my fix is worse than before? I think its clear that we will work on a fix for May.\nIMO The better workaround would be to bring back my fallback . I was doing the same as you except that I fallback to old way of writing into Output channels. This will not have any regressions as it was there in last milestone. When we fix it in May, we can remove the workaround.\nSorry, I am confused. May I know if this issue is due to Loading of spdlog library OR spdlog fails to create a file\nI see the workaround is if spldlog fails to create logger/file. So I think it is better to use my fallback.\nPushed the previous workaround I was using. As said, this will not break output channels if there is an issue creating spdlog logger.\nI don't see issues running dev with a user-data-dir containing a non-ascii character (eg ) any more. I do not understand if this is because of your workaround or if the new spdlog version actually fixed the issue. Moving it to you in order to check if the issue is fixed already, since I don't understand how to repro.\nI too do not see any issues after removing the workaround I before for the crash. Hence removed the workaround. To verify: Open VS Code using an user directory that has non-ascii character (eg ø) Verify that VS Code starts normally and the output channels are properly written Verify that there is no crash\nReopening this as the crash from spdlog is not fixed.\nI still see following error in the console with our log files Edit: This is seen in the console even with the workaround. Workaround just mitigates it by falling back to old output channels. Logs will not work though.\nUpdated node-spdlog library to support in file names. Thanks to for hinting it. With this crash no longer occurs with non ascii or special characters (tested with all reported characters). Added a smoke test to launch VS Code with special characters in user data dir."}
{"_id":"q-en-vscode-b9a4facb6b9e507bbf71d276826ad983db7a2e240bb94994931cad2f296584d6","text":"I did this on accident, but it took a little effort to figure out how I did it. Drag file from Finder onto Explorer Get scrollbar in Explorer to be visible, like by dragging near the top/bottom of the Explorer The entire explorer becomes highlighted when you hold the file over the scrollbar Drop file onto the scrollbar The entire explorer stays highlighted. It stays highlighted through interacting with the explorer and even switching to another viewlet. Only can be reversed by dragging another file onto the explorer. !\nActually a much easier way to repro is to two-finger scroll while dragging the file to get the scrollbar to appear.\ni'm able to reproduce it. will try to fix it\ncouldn't find where to start. help is appreciated\nI gave it a go in this PR:"}
{"_id":"q-en-vscode-8ad5afbabd1bf964bae7ee00be6958460382c249eb3b1f453a4645336b2f0d1b","text":"Have file named in explorer Copy/paste it Window hangs This is from We shouldn't attempt this when the number is = Number.MAXSAFEINTEGER. It loops forever here: This bug will happen whenever incrementFileName fails to modify the file name.\nI guess only interesting thing about this what we should do in this case as we need to create some new file name? Add without changing the original number?\nthat sounds reasonable"}
{"_id":"q-en-vscode-eaccda67f2116422d2ebfdf25128f96ce9fba2d92594e42fc2b0cf68791166d2","text":"Issue Type: <bBug</bJSON syntax highlighting fails on the following array (item 1, \"response\"): { \"request\": [24], \"response\": [152, 26, 0, 0, 227, 147, 4, 0, 0, 0, 0, 0, 64, 233, 245, 229, 2, 43, 60, 0, 0, 0, 0, 0, 0, 0, 0, 0, 64, 233, 245, 229, 2, 43, 60, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 65, 3, 0, 1, 0, 100, 0, 12, 1, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 70, 32, 32, 32, 32, 82, 67, 73, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 221, 41, 12, 1, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 48, 49, 32, 32, 32, 82, 67, 73, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 52, 41, 12, 1, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 48, 50, 32, 32, 32, 82, 67, 73, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 40, 12, 1, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 48, 51, 32, 32, 32, 82, 67, 73, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 69, 32, 12, 1, 204, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 48, 52, 32, 32, 32, 82, 67, 73, 32, 84, 48, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 231, 70, 12, 1, 205, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 48, 53, 32, 32, 32, 82, 67, 73, 32, 84, 48, 53, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 1, 206, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 48, 54, 32, 32, 32, 82, 67, 73, 32, 84, 48, 54, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 221, 41, 12, 1, 207, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 82, 67, 73, 48, 55, 32, 32, 32, 82, 67, 73, 32, 84, 48, 55, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 2, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 83, 70, 85, 50, 32, 32, 32, 32, 83, 84, 79, 67, 75, 32, 70, 85, 84, 85, 82, 69, 83, 44, 32, 68, 69, 67, 61, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 65, 0, 12, 2, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 83, 48, 49, 32, 32, 32, 83, 70, 32, 84, 83, 32, 76, 69, 86, 69, 76, 32, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 3, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 69, 70, 51, 70, 32, 32, 32, 32, 51, 45, 89, 32, 69, 88, 67, 72, 65, 78, 71, 69, 32, 70, 85, 78, 68, 32, 78, 79, 84, 69, 83, 32, 70, 85, 84, 85, 82, 69, 83, 32, 1, 1, 1, 2, 1, 0, 0, 28, 12, 3, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 69, 70, 51, 48, 49, 32, 32, 32, 69, 70, 51, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 3, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 69, 70, 51, 48, 50, 32, 32, 32, 69, 70, 51, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 3, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 69, 70, 51, 48, 51, 32, 32, 32, 69, 70, 51, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 67, 79, 77, 70, 32, 32, 32, 32, 67, 79, 77, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 4, 1, 0, 0, 0, 12, 8, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 67, 79, 77, 48, 49, 32, 32, 32, 67, 79, 77, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 8, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 67, 79, 77, 48, 50, 32, 32, 32, 67, 79, 77, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 174, 70, 12, 16, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 70, 32, 32, 32, 32, 77, 72, 73, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 4, 1, 0, 0, 0, 12, 16, 22, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 67, 32, 32, 32, 32, 77, 72, 73, 32, 67, 65, 76, 76, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 157, 32, 12, 16, 23, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 80, 32, 32, 32, 32, 77, 72, 73, 32, 80, 85, 84, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 0, 0, 12, 16, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 48, 49, 32, 32, 32, 77, 72, 73, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 171, 0, 12, 16, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 48, 50, 32, 32, 32, 77, 72, 73, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 16, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 48, 51, 32, 32, 32, 77, 72, 73, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 16, 204, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 48, 52, 32, 32, 32, 77, 72, 73, 32, 84, 48, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 16, 205, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 48, 53, 32, 32, 32, 77, 72, 73, 32, 84, 48, 53, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 16, 206, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 48, 54, 32, 32, 32, 77, 72, 73, 32, 84, 48, 54, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 16, 207, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 77, 72, 73, 48, 55, 32, 32, 32, 77, 72, 73, 32, 84, 48, 55, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 16, 250, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 77, 67, 95, 77, 72, 73, 32, 77, 72, 73, 32, 84, 77, 67, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 20, 6, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 83, 79, 77, 67, 32, 32, 32, 32, 83, 84, 79, 67, 75, 32, 67, 65, 76, 76, 32, 79, 80, 84, 73, 79, 78, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 2, 32, 32, 12, 20, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 83, 79, 77, 80, 32, 32, 32, 32, 83, 84, 79, 67, 75, 32, 80, 85, 84, 32, 79, 80, 84, 73, 79, 78, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 2, 32, 32, 12, 20, 172, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 83, 79, 77, 49, 32, 32, 32, 32, 83, 79, 77, 32, 83, 89, 78, 84, 72, 69, 84, 73, 67, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 20, 250, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 77, 67, 95, 83, 79, 77, 32, 83, 79, 77, 32, 84, 77, 67, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 73, 66, 70, 32, 32, 32, 32, 72, 73, 66, 79, 82, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 0, 0, 12, 24, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 49, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 50, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 51, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 204, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 52, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 205, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 53, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 53, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 206, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 54, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 54, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 207, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 55, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 55, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 208, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 56, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 56, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 24, 209, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 48, 57, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 48, 57, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 210, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 48, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 48, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 211, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 49, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 212, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 50, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 213, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 51, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 214, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 52, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 215, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 53, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 53, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 216, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 54, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 54, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 179, 71, 12, 24, 217, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 55, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 55, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 218, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 56, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 56, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 219, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 49, 57, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 49, 57, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 220, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 50, 48, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 50, 48, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 71, 7, 12, 24, 221, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 50, 49, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 50, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 24, 222, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 50, 50, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 50, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 23, 71, 12, 24, 223, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 83, 72, 50, 51, 32, 32, 32, 72, 73, 66, 79, 82, 32, 84, 50, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 27, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 68, 74, 73, 70, 32, 32, 32, 32, 68, 74, 73, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 81, 6, 12, 27, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 68, 74, 73, 48, 49, 32, 32, 32, 68, 74, 73, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 27, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 68, 74, 73, 48, 50, 32, 32, 32, 68, 74, 73, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 221, 41, 12, 27, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 68, 74, 73, 48, 51, 32, 32, 32, 68, 74, 73, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 27, 204, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 68, 74, 73, 48, 52, 32, 32, 32, 68, 74, 73, 32, 84, 48, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 34, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 70, 32, 32, 32, 32, 72, 83, 73, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 4, 1, 0, 0, 0, 12, 34, 22, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 67, 32, 32, 32, 32, 72, 83, 73, 32, 67, 65, 76, 76, 32, 79, 80, 84, 73, 79, 78, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 2, 0, 0, 12, 34, 23, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 80, 32, 32, 32, 32, 72, 83, 73, 32, 80, 85, 84, 32, 79, 80, 84, 73, 79, 78, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 2, 32, 32, 12, 34, 170, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 73, 83, 84, 68, 32, 32, 32, 32, 73, 78, 68, 69, 88, 32, 83, 84, 82, 65, 68, 68, 76, 69, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 222, 142, 12, 34, 171, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 73, 83, 84, 71, 32, 32, 32, 32, 73, 78, 68, 69, 88, 32, 83, 84, 82, 65, 78, 71, 76, 69, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 34, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 48, 49, 32, 32, 32, 72, 83, 73, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 83, 80, 12, 34, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 48, 50, 32, 32, 32, 72, 83, 73, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 34, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 48, 51, 32, 32, 32, 72, 83, 73, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 84, 12, 34, 204, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 48, 52, 32, 32, 32, 72, 83, 73, 32, 84, 48, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 34, 205, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 48, 53, 32, 32, 32, 72, 83, 73, 32, 84, 48, 53, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 179, 71, 12, 34, 206, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 48, 54, 32, 32, 32, 72, 83, 73, 32, 84, 48, 54, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 34, 207, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 83, 73, 48, 55, 32, 32, 32, 72, 83, 73, 32, 84, 48, 55, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 49, 32, 12, 34, 250, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 77, 67, 95, 72, 83, 73, 32, 72, 83, 73, 32, 84, 77, 67, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 35, 22, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 88, 72, 83, 67, 32, 32, 32, 32, 88, 72, 83, 32, 67, 65, 76, 76, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 0, 0, 12, 35, 23, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 88, 72, 83, 80, 32, 32, 32, 32, 88, 72, 83, 32, 80, 85, 84, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 0, 0, 12, 37, 5, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 88, 72, 72, 83, 32, 32, 32, 32, 88, 72, 72, 32, 83, 80, 79, 84, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 2, 2, 2, 0, 0, 0, 0, 12, 37, 22, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 88, 72, 72, 67, 32, 32, 32, 32, 88, 72, 72, 32, 67, 65, 76, 76, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 0, 0, 12, 37, 23, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 88, 72, 72, 80, 32, 32, 32, 32, 88, 72, 72, 32, 80, 85, 84, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 0, 0, 12, 38, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 70, 32, 32, 32, 32, 72, 72, 73, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 4, 1, 0, 0, 0, 12, 38, 22, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 67, 32, 32, 32, 32, 72, 72, 73, 32, 67, 65, 76, 76, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 2, 0, 0, 12, 38, 23, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 80, 32, 32, 32, 32, 72, 72, 73, 32, 80, 85, 84, 32, 79, 80, 84, 73, 79, 78, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 2, 82, 82, 12, 38, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 48, 49, 32, 32, 32, 72, 72, 73, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 69, 69, 12, 38, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 48, 50, 32, 32, 32, 72, 72, 73, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 38, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 48, 51, 32, 32, 32, 72, 72, 73, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 38, 204, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 48, 52, 32, 32, 32, 72, 72, 73, 32, 84, 48, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 38, 205, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 48, 53, 32, 32, 32, 72, 72, 73, 32, 84, 48, 53, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 78, 73, 12, 38, 206, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 48, 54, 32, 32, 32, 72, 72, 73, 32, 84, 48, 54, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 38, 207, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 72, 72, 73, 48, 55, 32, 32, 32, 72, 72, 73, 32, 84, 48, 55, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 78, 12, 38, 250, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 84, 77, 67, 95, 72, 72, 73, 32, 72, 72, 73, 32, 84, 77, 67, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 32, 32, 12, 51, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 86, 79, 76, 70, 32, 32, 32, 32, 86, 79, 76, 32, 70, 85, 84, 85, 82, 69, 83, 46, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 32, 32, 12, 51, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 86, 79, 76, 48, 49, 32, 32, 32, 86, 79, 76, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 51, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 86, 79, 76, 48, 50, 32, 32, 32, 86, 79, 76, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 175, 70, 12, 60, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 70, 88, 67, 70, 32, 32, 32, 32, 70, 88, 67, 32, 70, 85, 84, 85, 82, 69, 83, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 1, 1, 1, 2, 1, 0, 0, 0, 12, 60, 201, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 70, 88, 67, 48, 49, 32, 32, 32, 70, 88, 67, 32, 84, 48, 49, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 71, 6, 12, 60, 202, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 70, 88, 67, 48, 50, 32, 32, 32, 70, 88, 67, 32, 84, 48, 50, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 60, 203, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 70, 88, 67, 48, 51, 32, 32, 32, 70, 88, 67, 32, 84, 48, 51, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 221, 41, 12, 60, 204, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 70, 88, 67, 48, 52, 32, 32, 32, 70, 88, 67, 32, 84, 48, 52, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 60, 205, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 70, 88, 67, 48, 53, 32, 32, 32, 70, 88, 67, 32, 84, 48, 53, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0, 12, 60, 206, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 70, 88, 67, 48, 54, 32, 32, 32, 70, 88, 67, 32, 84, 48, 54, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 2, 1, 2, 1, 0, 0, 0, 0] } VS Code version: Code 1.23.1 (, 2018-05-10T16:04:33.747Z) OS version: Linux x64 3.13.0-40-generic <details<summarySystem Info</summary </details<details<summaryExtensions (16)</summaryExtensionVersion ------ vscode-markdownlint0.18.0 xml2.2.0 vscode-docker0.0.27 gitignore0.5.0 githistory0.4.1 python2018.4.0 cpptools0.17.4 title0.1.0 vscode-icons7.23.0 bash-debug0.2.0 annotator0.11.0 code-spell-checker1.6.10 mako0.2.0 cmake0.0.17 gitblame2.4.2 vscode-todo-highlight0.5.12 (1 theme extensions excluded) </details<!-- generated by issue reporter --\nFor performance reasons, we skip tokenization for lines that are longer than 20k. It would be good if the user can see a message when that happens.\nI would like to take this one if you guys think this is a but I would need some guidance. I already this: but I'm getting this error , so, I'm not sure how to import correctly into this file in order to use the method. cc:\nThis code runs on the renderer side, while is only for the extension host process. We have to wait for to add his thoughts. There's something similar when we open very large file. Check out the for that.\nis API available only to vscode extensions and should not be used in the UI code. The UI code should use and inject itself with it. A possible fix would be to prompt (as suggested above), but prompt only once. It would be very annoying to prompt 20 times if there are 20 lines over 20.000 chars.\nPR tackles this.\nTo verify, open a file with a very large line. A message should be shown, explaining why the large line is not tokenized.\nI tried on the production app (May 2018 (version 1.24)) and I didn't get the notification, I guess it hasn't been deployed there yet, right? Where should I verify?\nWe would want that someone else than the author of the change would verify that the issue is fixed. So, there's nothing for you to do here, since you are the author. Someone from the vscode team will pick up this issue and verify it. (The change will make it to the stable channel in the following couple of weeks. For now it is available in the insiders channel at )\nVerified !\nHow about using field for this kind of information? Lines that long only occure on minified files or base64 data images which doesn't actually need any syntax highlighting so this notification is completely useless in my opinion. How to get rid of it?"}
{"_id":"q-en-vscode-e0df8542f9d24ab6d9aa48145b5a24211f925f072e586d1767c83ff82921ac2e","text":"Learnings from NVAccess testing: I'm not sure how to get to the \"details\". Steps: Visual studio code editor. \"Ctrl+Shift+X\" key navigate to the extension side view bar option. Tab key navigate to the \"More\" option and select different options( \"Show installed Extension, Show outdated extension, Show popular Extension,Etc\") an extension and using \"Tab\" and \"Arrow\" keys navigate through the extension details window on the RHS\nWhat do you think of a \"Read more\" button for each item? It will get a little noisy. Any other ideas?\nOr just update the aria label on the extension item with \"Click to read more\" ? Because hitting Enter when the list item has focus, does open the details\ndefinitely prefer making changes on the code side. The label should probably read like “View extension details” so it’s a little more explicit."}
{"_id":"q-en-vscode-71b53177e9d583173d324936726979a363db2c0d35ff5aeb032679aab1a4d4aa","text":"This feature request is about accessibility. This is only for removing 1 or 2 Breakpoints from a list of, say 30 breakpoints across multiple files in the Breakpoints List. Right now, there is no Close Button in the Breakpoints List when debugging a program in Visual Studio Code. The only way to remove 1 or 2 breakpoints right now is to: on the Breakpoint from the Breakpoints List. the red dot on the line highlighted by VS Code. the Breakpoint is now removed from the Breakpoints List. I would like to request that we can simplify this process by using the middle scroll wheel click on the mouse, so when you click on one of the Breakpoints with the scroll wheel button, it does all of the 3 steps mentioned above. And, it can act as a Remove Single Breakpoint mouse click action. Just like the middle scroll wheel click on tabs to close 1 tab out of many other tabs.\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.647 --- <!-- score: 0.595 --- <!-- score: 0.583 --- <!-- score: 0.563 --- <!-- score: 0.555 --<!-- potentialduplicatescomment --\nyou can already use \"delete\" or \"cmd + backspace\" on the mac to remove the breakpoint. That way you can easily remove multiple. I do not understand what this has to do with accessibility since all these actions are already accessible. A middle click to remove a breakpoint would be consistent with our middle click to remove a tab. Thus I accept prs that fix this Code pointer Simply add a listener in the list for the mouse click, check if it is the middle one and if yes remove the breakpoint. If nobody volunteres I can look into this in the future\nI am not aware of using \"delete\" keypress to remove the breakpoint. How is this triggered? If I press \"delete\", it would remove a character in my source code.\nfocus on the breakpoints view in the debug viewlet F1 Debug: focus breakpoints\nWouldn't it be more user-friendly just to add a close button for breakpoint list item? For example: !\nIt's more about middle scroll wheel click to remove breakpoints, and I think what you pointed out may be a separate issue.\nrelease notes?"}
{"_id":"q-en-vscode-0c84edbaeb2e8f321aa4dd6f966cb1a1dd559773fd2f8287ac78d05c6b455595","text":"Issue Type: <bBug</bThe cursor after pasting, I'd expect it to be at the end of the pasted text: ! VS Code version: Code - Insiders 1.26.0-insider (, 2018-08-02T05:12:16.245Z) OS version: Darwin x64 17.7.0 <details<summarySystem Info</summary </details<details<summaryExtensions (4)</summaryExtensionVersion ------ 999-pack10.12.6 ssh0.0.4 github-issues-prs0.9.0 vscode-open-in-github1.7.0 </details<!-- generated by issue reporter --"}
{"_id":"q-en-vscode-cee7451470a934bcb43da11082f2371d0080b61906cd62e93b7db20ce852d903","text":"I would like to make a pull request that changes the markdown snippet for links to use HTTPS instead of HTTP. Microsoft seems to encourage HTTPS by default in the latest .NET Core & browsers are also encouraging it. To me HTTPS sites seem to be the default for most websites these days.\nSure, please open a pr. I didn't even know we had a snippet for links :)\nFixed by"}
{"_id":"q-en-vscode-5807bc0c8996c95adb7af39dd65fe648fe7d1807d6ca8c55c949414fa38e561d","text":"Issue Type: <bBug</bI use oh-my-zsh and set integrated terminal /bin/zsh cmd+, to show settings tab terminal(ctrl+shift+`) settings terminal integrated: Font size option. say we have default value 12. . now value is 1 and terminal gets so small 1 so now value is 11. and terminal is gone. expected to see current terminal cursor. i type lots of enters or ctrl+c, then I can see my terminal again. i saw my terminal again, it looks like change font size 11 -1 -12 doesn't make terminal to disappear. VS Code version: Code 1.27.1 (, 2018-09-06T09:14:32.385Z) OS version: Darwin x64 17.7.0 <details<summarySystem Info</summary </details<details<summaryExtensions (7)</summaryExtensionVersion ------ gitlens8.5.6 code-runner0.9.4 beautify1.4.2 Go0.6.89 vetur0.12.6 vscode-docker0.2.0 java0.30.0 </details<!-- generated by issue reporter --\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.481 --- <!-- score: 0.459 --<!-- potentialduplicatescomment --\nalso vertical scroll of terminal get somewhat screwed even after I see my terminal again by typing enter or ctrl+c\nthis seems to be related to the usage of from xterm: a similar problem with an explanation of a solution can be found here: In short, after the fontSize changes, the terminal renders before the xterm field is updated, so the render uses previous values (corresponding to the previous font size values) Related: ,\ncan some settings only commit the change on blur?\nIf is an invalid font size for that setting, can the terminal ignore it or set a minimum in the setting schema?\nafter thinking about this some more I think this is totally on my end as something weird happens when changing the setting frequently. There is already a minimum value which is why VS Code doesn't crash"}
{"_id":"q-en-vscode-98970aca17f80031b7af5038e4ac8ddf64e231e1e93732c0fc47c9a6ef1c48e4","text":"Repro: Add keybinding: Find a quick fix, and hit alt+enter quickly. The first action is triggered. I was seeing this here (notice Promise not TPromise): !"}
{"_id":"q-en-vscode-eb71d82ccdf1ca5dc5552812ddc6ee06d49d24c9f17340bfad7656fa4d7354c6","text":"Env: In Insider version, the API doesn't respect the option. The same code works in the normal vscode 1.28.2.\nThe code here converts the filters from the extension API into the electron API, but the function is supplied with the electron API format so this is not necessary and is causing the break.\nI think this is something that we should ship with 1.29\nWe believe this is a regression, adding candidate label"}
{"_id":"q-en-vscode-a46fe737b4008f322fbf4743dea0b270cd7a557c6c5f346f3fc3a1c59a7f6373","text":"<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.28.2 OS Version: Microsoft Windows 10 Pro, 10.0. Build Steps to Reproduce (2 problems): a source code file over an IPv6 URL and notice the faulty incomplete underline: ! Click the link: the default browser opens without accessing the URL <!-- Launch with to check. --Does this issue occur when all extensions are disabled? Yes\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.463 --<!-- potentialduplicatescomment --"}
{"_id":"q-en-vscode-984cd81dac3a391b79e42311d01dbf23bf6e3cf7ef50bf89a82f4387da913ce5","text":"macOS run in fullscreen mode open the process explorer the prices explorer window also opens in fullscreen mode...\nHi Gonna take a look at this. It seems to be happening for 'Report issue' also.\nHi Can I help you somehow to get this merged?\nYep! Sorry for the delay, your change looks good. I'll merge it now. Thanks!"}
{"_id":"q-en-vscode-c7f6c717d4c04a79b6b415e0dcba2d1eefb102b447e1f0486991d1fddf2b7377","text":"Set color customizations for the peek widget: 29: ! 30: ! Candidate or 1.30.1 maybe?\nGrrr, another regression... I am for putting this into 1.30.1\nI have pushed a fix to master and I have created we think having for for is good enough, otherwise the PR can be merged for"}
{"_id":"q-en-vscode-7faf64986485648d6d9a0039632745086c443495c2ac63b1afbd603a4106a61c","text":"While creating a user snippet, the rule must be like this: If I want to show a lot of description, I must write a huge string which includes many escape characters like this: It maybe a good idea that option receive a list value, like the :\nThe Version infomations: Code Version: 1.30.1 (user setup) Commit: Date: 2018-12-18T18:12:07.165Z Electron: 2.0.12 Chrome: 61.0.3163.100 8.9.3 V8: 6.1.534.41 OS: Windows_NT x64 10.0.\nCan you guide me on resolving this issue?\nsure, checkout , and"}
{"_id":"q-en-vscode-8f202a480169a94a44169c515a2d7247ae1332c0a8761aff1cf73f20c612f77a","text":"Ability to update built in extensions\nThis is a pre-request to enable built-in extension to on-board to\nDisabled updating builtin extensions in stable"}
{"_id":"q-en-vscode-d6c82a40ac80128c6037465fe145129c099abcf30908bed0cf043b53057c1597","text":"1) use yo code to create an extension 2) Replace the with the code below 4) Press F5 to run the workspace 5) Open a project for a node module that has a nodemodules folder with many modules. 6) Execute the ´Hello World' command --the call workspace.findFiles returns file matches from the folder, but the nodemodules folder is excluded. Running the same extension in stable properly excludes matches in the folder. Tagging as a candiate for February since the change of behavour impacts NPM script running and now proposes scripts from the node_modules. The code for\nWould be fixed by ."}
{"_id":"q-en-vscode-a9bc22c02632ac9fec4ae46d644c23321bdfb643cb7facec00348f14c17a160c","text":"Open a folder Create untitled file with some content Quit vscode Open vscode Two windows are opened, one has the folder +untitled file, other is an empty window + untitled file. So the same untitled file is opened in two windows.\nI could only repro this with an empty window: Code via finder all files/folders an untitled file, add changes +q Code via finder <img width=\"490\" alt=\"screen shot 2019-03-07 at 5 42 14 pm\" src=\"https://user-\"\nMoving this to as the last who touched this code.\nfor some reason the  array has multiple entries for the same window?\nTemporally until update is not yet, for prevent backup window add to settings \"window.restoreWindows\": \"none\"\nWhen set , then every close & start VSCode will multiple windows (1 -2, 2 -4, 4 -6, and so on). Fix is needed ASAP.\nthis change is not in the recovery build yet pending right?\nPushed.\nI saw this by downgrading from 1.32.2 to 1.31.1, then upgrading back. But I don't think we care about that.\nI can confirm that this issue is fixed in the latest 1.32.2 release! Thanks! But another one what I noticed: VS Code one existing document Optional: Open a new tab or new window with a new tab and type anything and do NOT save this new file VS Code -Open VS Code Now all tabs are correctly restored! VS Code a File from Windows Explorer with right-click and open in VS Code All tabs should re-opened (Tabs/files from Step 2., 3. and 7.) Only the unsaved file from step 3. and the one from the Windows Explorer, step 7. are opened. And the last session (int this example, file from step 1. is lost! Is this wanted or can this be configured?\nConfirm fix and also above bug with 1.32.2 and 1.32.3. Could You report this as new issue? Now session is partially broken :-(\nthanks. Please see"}
{"_id":"q-en-vscode-2ea5245f7f128551d9d321749c985ffdd7866c58d0590bf479b83fb7076a70ad","text":"Issue Type: <bFeature Request</bHere's what I have tried using as a reference: In the settings page, \"command\" doesn't show up as a link, just plain text. It would be nice if command links were supported like regular URL links. VS Code version: Code 1.32.3 (, 2019-03-14T22:46:48.547Z) OS version: Darwin x64 18.2.0\nI assume you want this in the settings UI.\nThat's correct. A clickable command link above the field.\nSure, out of curiosity, what sort of command would you use this for?\nThe property/setting I'm trying to use this for is a directory path containing some directories. And the command is a refresh command that will pick up the latest setting value and do some parsing of the contained directories. I know I could use to detect changes and trigger the command automatically but I want to allow the user to refresh manually.\nAre you avoiding because it might trigger while the user is not done typing, or is there a different reason?\nYes, that's the main reason.\nI would like to second this. My use case is: My extension involves authenticating with a server, and I have built a web-view for this. This authentication process is too complicated to handle through native VSCode settings (I want to validate the server, there might be 2FA, etc, etc) However, since users have an expectation of seeing their settings in the settings UI, I would like to display the server that the user has authenticated to in the settings UI as a read only property. This requires me to have a description which says \"You can modify this value by doing X\" And overall, it seems like a better user experience to have that description be \"you can modify this by clicking \" and having open my webview, rather than: \"you can modify this setting by selecting the X command in the command palette\""}
{"_id":"q-en-vscode-e719d7af96ec0d5310dd9fb7e82bd7c8f45aaf1e4fc5f4c542a75178c1728473","text":"Maybe you've already thought about this but it would be useful to be able to create a new folder in the simplified file dialog. This could be a button in the title bar or a confirmation step when a folder that doesn't exist was typed. Creating a folder and opening it, and creating a folder and saving a file in it, are common workflows so it would be useful to be able to do those in one step.\nFor the meantime:\nHey thanks for the mention! This was a feature I personally needed for quite some time - we all basically do those steps mentioned by on a daily basis so I decided to code an extension for it. Even though the extension already works great, I am actually working on it right now and adding a new feature which will allow it to create new folders using the Simple file dialogs - regardless of user setting, see . If anyone's interested, I think I'll manage to finish it by tomorrow.\nI've been busy with work and other projects but after a long wait, I worked really hard and managed to finish my npm module for Visual Studio Code that implements the aforementioned behavior and shows a custom QuickPick that allows browsing and creating folders from the Command Palette, even nested/recursive folders. The extension that implements this package is available at repo. Not perfect (yet) but it works.\nAlso create new file in Open File dialog would be useful.\nSaving is good. Opening hasn't been implemented -- should there be a new issue for that?"}
{"_id":"q-en-vscode-bf2dc64c1fb69862e064c402aa4e9cb579e1e7e28b996f51407120f6d2923530","text":"When picking a Formatter to run for a file type, it'd be nice to see what extension is providing each formatter. This can make it easier to choose which formatter I want to use, or to open an issue to request that the extension use a more obvious name for their formatter :) (in this case, the Ruby extension accidentally provides the legacy formatter as well as the \"current\" formatter at the same time). When in a situation like this, I'm not sure which to choose: <img width=\"621\" alt=\"Screen Shot 2019-04-07 at 6 20 49 PM\" src=\"https://user-\"Maybe have the extension name de-emphasized like the \"Configured Tab Size\" text in the indentation dropdown?: <img width=\"615\" alt=\"Screen Shot 2019-04-07 at 8 29 28 PM\" src=\"https://user-\"\nWe use the 'displayName' or 'name' (when not having the former) but it seems that there are multiple extensions with the same name... We should be more precise in that case."}
{"_id":"q-en-vscode-94ff3f958f41c70833a1c5cfd665b2bee74f62456e2105d8f440803b4f9ddd8f","text":"Description Currently, when editing a file, the command <kbdctrl</kbd+<kbdshift</kbd+<kbda</kbd(toggle block comment) generates the following comment placeholder . I wonder if it might be also wise to include this as an auto-closing pair so that when users attempt to write the closing is automatically like other languages (ex: Typescript).\nThis issue is caused by an extension, please file it with the repository (or contact) the extension has linked in its overview in VS Code or the for VS Code. See also our guidelines. Happy Coding!\nI'm not quite sure as to who or where the owner is for the extension. Looking at the I noticed multiple past updates to support additional features. Isn't the proposed feature request present in the issue similar to those past updates?\nOh, you are right. This part is us, not the extension. Sorry for the confusion...\nLooks like a good change to me. Feel free to make a pull request.\nNot able to reproduce ctrl+shift+a not having any effect.\nI assume you are editing a file, but sorry the keybinding I mentioned is for Linux, on macOS for example the keybinding is . The easiest way to test is to simply run the command from the menu 🡒 :)\nGot it, in windows it is Shift+Alt+A.\nI forgot to mention I won't be able to submit a PR myself due to my company policy. If anyone would like to take on the issue, please feel free to do so :)\nIn that case, I am more than happy to make a pull request for this. Thanks for the information. I will submit one soon.\nthank you!"}
{"_id":"q-en-vscode-c63cf426bf0e3e8e08c951c7f1810efa7920f6df253898ca98be83285d194851","text":"a TypeScript file open (with some code in it) the outline to jump to the beginning of a method something somewhere the cursor jump to the beginning of the previously jumped-to function as soon as you stop pressing keys for .5 seconds Yes (both in insiders & stable) Version: 1.34.0-insider (user setup) Commit: Date: 2019-04-26T13:39:29.623Z Electron: 3.1.8 Chrome: 66.0.3359.181 10.2.0 V8: 6.6.346.32 OS: Windows_NT x64 10.0.\nIt seems that the tree send a selection change event after updating children (which happens because markers changed). I will look into a workaround, e.g being but it seems like a tree issue <img width=\"777\" alt=\"Screenshot 2019-04-29 at 10 33 24\" src=\"https://user-\""}
{"_id":"q-en-vscode-2902d485cbeeebc1dd3e0385433a14c1b2f0ef9b775a1f43f1b8c2e2d81352f5","text":"Extensions smoke tests are failing\nThis was caused because of two reasons has changed their search strategy and looking for extension by name is broken - label of an extension has string which was a bug introduced recently. Fixed them here -"}
{"_id":"q-en-vscode-e27f845ea830c92fd6c26eb1d08c279539d6a8f688b9421904c20abacb1d7e90","text":"Issue Type: <bBug</bI noticed when I install extension by clicking on the green \"Install\" button from within the VScode extension page, as opposed to the \"Would you like to install...\" notification/chat bubble in the lower right corner of the editor, the chat bubble stays there, asking me if I want to install the extension I just installed. Would be great if the state was in sync and the bubble would be dismissed when I installed the extension, no matter how. <details<summaryScreenshot after installation</summary<img width=\"2032\" alt=\"Screen Shot 2019-05-06 at 17 54 50\" src=\"https://user-\"</detailsVS Code version: Code 1.33.1 (, 2019-04-11T08:22:55.268Z) OS version: Darwin x64 18.5.0 <details<summarySystem Info</summary </details<details<summaryExtensions (19)</summaryExtensionVersion ------ project-manager10.5.1 solargraph0.19.6 bracket-pair-colorizer1.0.61 vscode-eslint1.8.2 es7-react-js-snippets2.2.2 gitlens9.7.1 prettier-vscode1.9.0 rails-snippets1.0.8 rest-client0.21.2 vscode-codemetrics1.18.0 vscode-language-babel0.0.22 ruby-rubocop0.8.0 sublime-keybindings4.0.0 debugger-for-chrome4.11.3 JavaScriptSnippets0.2.0 ruby0.22.3 code-settings-sync3.2.9 change-case1.0.0 JavaScriptSnippets1.7.2 </details<!-- generated by issue reporter --\n, I would like to try this , Some starter and code pointers would be helpful :)\nIt depends on how got this notification. can you please let us know how you got this?\nIt shouldn't depend on that. I got it by clicking install button in VS Code Extension Marketplace website, in the browser.\nThis notification comes from here -\n, I checked the code but just wanted to discuss on one thing . Is that notification needed in the first place , clicking on install button on top will do the same job , Also removing that notification will keep the two scenarios consistent and same (Normally opening that extension page and opening the extension page from browser) We can remove the notification prompt to fix the inconsistency .\n, were you able to look into this ? I can remove the notification bubble and send a PR, please confirm :)\nYeah it look unnecessary for me. Please remove it."}
{"_id":"q-en-vscode-aec580cc18592d155e07bfe296b7f54d87b8f4394176af85c8793e6d21e43c5f","text":"<!-- Please search existing issues to avoid creating duplicates. --Step 0. Version 1.35.0 with extensions disabled. Step 1. Open a file with changes committed in git Step 2. Use ALT+F3 to navigate to the next change in-line Step 3. Click Revert Change Step 4. find the line number you were on and move your cursor there, then hit ALT+F3 again <!-- Describe the feature you'd like. --I would like vscode to remember where I was whenever I revert a change, then return me to that place, so that I can see the change disappear and so that I can proceed using hotkey to the next change.\nWe might just have to set the cursor in that position, should be simple. Would you like to submit a PR?\nI can take a look at this.. As this will be my first pull request here in vscode repo, I will require some help. Do let me know a good starting point for this..\nHi, I would like to work on this issue. I'll submit a PR soon.\nDo we want to fix this for 'stage change' functionality too? It behaves now the same way as 'revert change' - it starts searching from the beginning, instead of placing cursor in staged line position.\nSounds good."}
{"_id":"q-en-vscode-3d1e7ffabb0213f52e6a751eba03a55df2b2f6f5b7550ee4be750d5a0ef48989","text":"Support rendering text decorations in the minimap, beginning with search results. We can add a new option to decorations, similar to in : The new optional property could look like: The find widget can then set this property for its decorations, and the minimap can read the decorations from its text model."}
{"_id":"q-en-vscode-bc8a89b4a9cc1a6fae90d2a7b361e46719eff29f5212c6a44d8cb32abe1ec55b","text":"I am a bit confused about this change, isn't it that command line arguments are being passed to Chrome/Electron anyway? Why is this needed? Originally posted by in\nWould you like to submit a PR that reverts the original PR?"}
{"_id":"q-en-vscode-3f8a221e4a452fb8bea9fd82d072124cba16c0d735ee41b01002b248523b4cb2","text":"When installing the extension completes, the message is shown: \"Completed installing the extension <publisher.<name. Instead, the extension displayName should be shown. VSCode Version: 1.36 OS Version: Windows 10 !\nCurrently, is displayed, it would be better if is shown."}
{"_id":"q-en-vscode-c4207f9b44eee2212b93595564da87daae33a678d9d528a41f678b8d2e2cdd2e","text":"Apparently resolving a string like will work but it will leave the last curly brace. We should disallow this case or throw an error, etc. See\nStill seeing some odd behavior (although this may be expected) Repro MacOs : new terminal Bug In the terminal, see: This does not happen if I use a setting such as\nLooks like that's a side effect of 'breaking' the path. Nested variables are just no longer interpolated, it's left as:\nare those side effects you listed above intended side effects of the change? If so, I see the same thing on codespaces and we can just leave this issue verified. Otherwise, should the issue be moved to the next milestone?\nYes, it's intended that variables that have nested segments are no longer replace. Previously it would have resulted in a malformed string."}
{"_id":"q-en-vscode-bf2eb3a7e1ada1426a619ae5e87192cf3eea187bdb314491de8e880a1cfa7c28","text":"Are there any commands in VScode to navigate \"up\" or \"down\" (a folder, directly) in a selection in breadcrumbs? <img width=\"659\" alt=\"Screen Shot 2019-07-13 at 7 04 55 PM\" src=\"https://user-\"\nfor what the command id is but pressing ArrowLeft does the job for me\nThe command is actually . It will collapse a node. If already collapsed, it will select the parent. Should we implement a command to simply select the parent?\nI can do this and have a try\ni make a PR but not very clear about how primary work.\nCommand is , you'll be able to configure a custom key binding when it's out."}
{"_id":"q-en-vscode-4b002998e2343db9beed6d12c42bbae5342f51a4d28ca4ae2306d8e408ccd116","text":"Testing Please see the attached recording ! Notice how 'bb' is only shown after adding 'ccc'\nGood catch. The \"fast update\" only applied when list DOM element is in focus but not when list item is in focus."}
{"_id":"q-en-vscode-f2a85e31b52d4ab998dd1143c023e3db1035ed3609af4d980644e8662513cefc","text":"Issue Type: <bBug</bIf you have keyboard with multimedia buttons using them in Explorer tree triggers highlights/filtering. focus to Explorer tree Play/Pause media key that filter for was applied ! This keyboard have only one multimedia key, but I expect that other might be affected too. VS Code version: Code - Insiders 1.37.0-insider (, 2019-08-01T05:37:54.848Z) OS version: WindowsNT x64 10.0. <details<summarySystem Info</summary </detailsExtensions: none <!-- generated by issue reporter --\nI would like to have a go at this.\nCan you still reproduce this? Think it may have been fixed in commit . I know I used to be able to reproduce this but I can't anymore.\nthanks for checking this issue. Unfortunately I'm still able to reproduce this issue wits latest Insiders build. ! Version: 1.40.0-insider (user setup) Commit: Date: 2019-10-22T05:27:32.053Z Electron: 6.0.12 Chrome: 76.0.3809.146 12.4.0 V8: 7.6.303.31-electron.0 OS: Windows_NT x64 10.0. Reported key is (which looks like expected key code for this button) ! UPD. While looks good media key might \"leak\" somewhere on\nIssue related. I am thinking that this key filtering (see below) could be simplified massively. Instead of blacklisting all of these edge cases, we could just check if is a single character. I believe this should only allow key presses that we want. Then we can just add the exceptions such as . If we remove L660 and modify L664 to be: Key filtering:\nI thought about this. But you're going to hit issues with unicode characters, in which one visible character actually has length 1. Let's just keep filtering out these cases one by one, it's fine. Simply add .\nCannot reproduce in 1.39. Could someone who can verify the fix in the latest Insiders?\nFix works as expected for me. Version: 1.40.0-insider (user setup) Commit: Date: 2019-10-31T05:42:33.138Z"}
{"_id":"q-en-vscode-894f01d006d8040416c04fcf14fcaa6d39364fea1072689d538b1c71c9d6d7fc","text":"Reported here: Nice example:\nSorry didn't get the chance to get this done, pushing it out to September.\nNote to self: remember to revert"}
{"_id":"q-en-vscode-6590fc903ddbb871d74a94d60824be0ff9f0132bc40589cc25ea25a2205770ad","text":"VSCode Version: 1.37.1 OS Version: Fedora 30 I'm extracting the CSS Language Server for usage in 3rd party IDE (Eclipse IDE). I noticed that if I start the CSS-LS without passing a element to initialization, the LS will fail because of . One of the important lines in the stack trace is .\nSubmitted PR for this issue.\nFixed by .\nVerified in master by removing the dataPaths"}
{"_id":"q-en-vscode-5008901ef0bd3e1e2b211b765d763bb1dc87237e04186a574d4f4bc24c595e6f","text":"Issue Type: <bBug</bSteps to Reproduce: a project/folder mouse into Editor and then move it out from right edge of the window Activity Bar mouse into Side Bar and then move it out from left edge of the window VS Code version: Code 1.37.1 (, 2019-08-15T16:16:34.800Z) OS version: Darwin x64 18.7.0 <!-- generated by issue reporter --\nThe issue seems to be in , but I am not sure what the difference is between that edge of the editor and the others. For some reason the mouseLeave event is not being fired on the right-hand side. It seems it thinks the is the same as the initially given - which means the leave event is never fired. I'm happy to keep digging into this if that's ok with you."}
{"_id":"q-en-vscode-c59ea12bb4d903c0831262f61dd4b2011be3c8b0c25414c946772da469ee8403","text":", Replace string functionality in search does not work properly for hyphen and underscore separated variables , Initially figured it out from this Example: Replacingto with preserve case turned on replaces correctly to Replacing to with preserve case turned on replaces incorrectly to instead of PS : Additional Info from the code I found out there is a bug in function. The whole idea to call the replacement function again at was to reuse the already existing replacement logic ( First char capital , full caps and full small) but there was a bug as the hyphens and underscores were not even called because of wrong ordering in the condition checking logic.\nThanks for the PR, can you please phrase this issue in the form of a bug report (ie describe what to do, what happens, what is supposed to happen)"}
{"_id":"q-en-vscode-318f37c03b2fdb709168bfc4494ab2e11b939fa0f105c89349a79ed16a14b135","text":"First, thanks for making VSCode soooo awesome! Wouldn't be as productive as I am without it Currently, VSCode suggests installing the to have support for TypeScript. However, is supposedly going to be deprecated sometime this year. ! Change the suggested extension to ESLint instead. This way, we're providing a solution to the community that is future-proof (at least for the foreseeable future) I'd be happy to open a PR if you were to accept this proposed solution. (it looks like this would be a matter of replacing the extension id here: )\nRemoving it sounds good. We already stopped recommending tslint for general ts users: Probably just remove the whole typescript section for now. Setting up eslint for TypeScript is not simple enough at the moment to show on the welcome page in my opinion Can you submit a PR that does that?\nAgree with you there. Definitely!\nI noticed you the \"help wanted\". I'm planning to submit a PR. Can we change the label indicate this issue is in progress or taken?"}
{"_id":"q-en-vscode-9dc13ef188a12fd1e9d36d36fa531684de42f09668283b2c2048e06389f69714","text":"! ! Version: 1.39.2 autoReveal always moves the filename entry to the middle in file explorer window when we select a file on tab. if the filename is not visible in the explorer, this is ok. if the filename is visible in explorer, it should not move the filename to the middle in explorer, just highlight the filename where it is, right there in explorer. thanks!\nI've taken a stab at this in as it's something that annoys me, hope that's alright"}
{"_id":"q-en-vscode-67ee5005535b2dee1f905472dadf438fd96363a412bcbbb6a5c29ca524430ed1","text":"I'd love it if the description would specify that this setting only pertains to the command as explained in , as opposed to the SCM interface, which would be the expected behavior according to the current description.\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.564 --<!-- potentialduplicatescomment --\nFeel free to submit a PR for this instead. Here's where you want to edit:\nCool, thanks!"}
{"_id":"q-en-vscode-086deae9b71bd3e7ee03eb563ab9a11941d5638cc5c014b8a2f2659d6cd3364a","text":"Babel 7.7 introduces a : , which shares the schema with . When editing a , one expect to see JSON validations same as . I tag it as feature request as Babel 7.7 is not yet released.\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.92 --<!-- potentialduplicatescomment --"}
{"_id":"q-en-vscode-09ceea76ba7003edf983f966a44f1220c3db857db7ca744b34cc91e46232539f","text":"Issue Type: <bBug</bReopening of steps the intergrated terminal. a command that generates more than one screen of content: to bottom run emacs and then close emacs using ctrl+c ctrl+x Bug Terminal scroll position is sometimes reset to top of buffer ! VS Code version: Code - Insiders 1.40.0-insider (, 2019-11-06T16:56:25.487Z) OS version: Darwin x64 18.7.0 <details<summarySystem Info</summary </details<!-- generated by issue reporter --\nCan't repro some things that might help diagnose: Debug VS Code to see what's getting hitting hit. Calls to this when exiting would be particularly interesting Try repro in the demo Enable escape sequence logging and capture the entering emacs logs and exit emacs logs for both the case where it's fine and the case where the scroll is off.\nSeems to happen most reliably when I run emacs as the git editor, but as the screenshot shows it can also happen for normal runs. From debugging: Normal case called with called twice with not called by Bugged case called with called with called again called by with\nThat's definitely the problem, could you get the stacktrace for both of those?\nStacks look the same. Here called with : ! : And with : ! It seems like a timing issue. Between the call to which sets and , someone else is setting . Here's one possible call sequences I see opening emacs in the bad case (true/false is the value of ): ! And from the good case: ! For what it's worth, I also currently see this more often in insiders (vs OSS)\nI don't use emacs or something, just run some tests which output a lot of stuff, then hide and unhide terminal - after a couple of moments I am on top of terminal. Scrolling it with a scrollbar to the bottom doesn't bring me to the last line, so I press enter\nYeah, I have to press enter to get to the bottom of the terminal. I can confirm that this is an issue with Linux and Windows (I see this on both), so it doesn't appear to be specific to any OS. This also happens with both native and remote development, so it is not specific to remote development (which I use a lot). For me, it starts happening after about an hour of use, and then becomes really frequent (almost every time I open the terminal). Its causing me to form a habitual response to just always press enter when I open the terminal :frowning_face:\nAnother repro:\nThe case seems to be a symptom of the TerminalPanel onDidChangeVisibility callback not triggering a setVisible call on the active terminal instance. The code to re-sync the the Terminal viewport and scrollbar is in the TerminalInstance setVisible method. I'm not sure if it's the same issue as the Emacs variation of the bug.\nI'm also having this issue. When I toggle open the terminal on WSL and PowerShell, the terminal is scrolled to the top, even with all extensions disabled. I'm not using emacs - it happens reliably regardless of the type of content in the terminal. Just focusing on the terminal doesn't have this behavior, only toggling it does.\nUpdate: for anyone looking for a temporary fix, I used the extension \"multi-command\" and set up a shortcut to scroll the terminal to the bottom when opening it. Since I needed to first focus the terminal after opening it, I had to set up two cases, toggling when it's closed and toggling when it's already open: \"multiCommand.commands\": [ { \"command\": \"multiCommand.toggleTermClosed\", \"sequence\": [ \"\", \"\", \"\", ] }, { \"command\": \"multiCommand.toggleTermOpen\", \"sequence\": [ \"\", ] } ]`\nverified with steps from"}
{"_id":"q-en-vscode-bdbe74a5d08107a8b0de08ffcc2827c57a905bff82be3f73473540155c72786e","text":"Issue Type: <bBug</bTypeScript references code lenses the VS Code Code base, open the line: After type a comma then and re-type again Bug Suggestions take forever to come in Computing the references in TypeScript can potentially be expensive. This is caused because the method is not being as aggressively canceled when the document changes as I expect it to be (and as I believe it used to be) We can workaround this on the TS side but it seems like unexpected behavior /cc VS Code version: Code - OSS Dev 1.41.0 (Commit unknown, Date unknown) OS version: Darwin x64 18.7.0 <details<summarySystem Info</summary </details<!-- generated by issue reporter --\nAdds a TS specific fix. Whenever the document changes, we now aggressively cancel all existing code lens requests against the TS server for that document\nI believe nothing has changed here - at least not as far as my recent memory goes... Cancellation happens after 250ms which isn't aggressive but should also not be perceived as overly slow. Are those 250ms the delay you were seeing?\nthe delay is more like several seconds. It could be because we now provide references code lenses for object members, which adds a lot more references for TS to compute.\nhm... but the delay/debounce on the by this... Does work happens on the extension host thread (not the tsserver-process) so that the cancellation doesn't come through until the event loop is free again?\nOk, so after more digging, I don't think is related to the debounce. Instead it seems some calls to are not ever canceled when the document changes. Here's the basic sequence I see on the extension host side: The issue is that the first batch of resolve requests (40-49) are not cancelled when they should be Resolving individual TS code lenses can take around a second in some cases\nNice find and kudos for digging through that code"}
{"_id":"q-en-vscode-0ef75b3ae9a96bc2cc7c908e5306b5d52d1d75fedfcffbbc5906ba57a83b7d70","text":"Please add an option to disable incremental naming. By disabling ... the end effect should be to effectively overwrite existing files when pasting. The current options of and are helpful; however, not useful in my context. For myself and those users who commonly do file operations in the host OS file explorer, the expected behavior when copying files/folders in the VSCode explorer is to similarly have these overwritten.\n<!-- -96bd-47a8-a0e8- --This feature request is now a candidate for our backlog. The community has 60 days to upvote the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our . Happy Coding!\n<!-- -a7ce-58b9-b1f9- --This feature request has not yet received the 20 community upvotes it takes to make to our backlog. 10 days to go. To learn more about how we handle feature requests, please see our . Happy Coding\n<!-- -a7ce-58b9-b1f9- --This feature request has not yet received the 20 community upvotes it takes to make to our backlog. 10 days to go. To learn more about how we handle feature requests, please see our . Happy Coding\n<!-- -c9e0-7adb-d31b- --:slightlysmilingface: This feature request received a sufficient number of community upvotes and we moved it to our backlog. To learn more about how we handle feature requests, please see our . Happy Coding!\nWe are open to a PR which adds the option to the setting. Code pointer\nTo verify set and copy and paste a file to ensure its name isn't changed and if a file of the same name exists in that directory it asks to overwrite the file"}
{"_id":"q-en-vscode-759949766640196ce8609ca3a227858e9a5dfc7d2effcd4e5bfea3ce215e67d8","text":"<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: code-server2.1692-vsc1.39.2-linux-x86_64 and a few previous OS Version: Fedora 31 Firefox 70 @ Windows 10 1909 I am running VS Code remotely via code-server. When I use Find function to search something, then press Esc to exit search field and try to insert a new line by Enter, the search field is opened instead of inserting a new line. It works fine in Chrome though. +F, type a text, press Enter Esc to exit search field, eventually press End to deselect found text Enter field is opened again instead of inserting a new line <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.482 --<!-- potentialduplicatescomment --\nA similar issue was reported for the monaco-editor. The cause and fix should be the same, so I'm linking it here:\nYes, monaco playground behaves the same on my side.\nI wonder if this issue is reproducible when running VSCode in a browser. We allow to run VSCode in a very simplified form in a browser following these steps: setup to be able to run VSCode out of sources: run from a command line it should open a browser with VSCode running If you are not able to set this up, we also provide a hosted version of VSCode for Web at: https://vscode-web-test- Does the issue reproduce in that setup in the browser?\nI just tried the web version and it reproduces there, too.\nI have tested the playground on azurewebsites and the issue is reproducible. !\nThanks for testing\nThis problem happens to me too, when using codercom/code-server:2.1688-vsc1.39.2 in Firefox Linux. Enter starts working again when you close the Find box enough times by clicking it instead of pressing Escape."}
{"_id":"q-en-vscode-e585ddeba652daa41182440f9890d68f68f1a8e9d37c5b1a940c52e260538de0","text":"VS Code has support for . It would be nice to also have support for \"normal\" region folding.\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.732 --- <!-- score: 0.506 --- <!-- score: 0.503 --- <!-- score: 0.456 --<!-- potentialduplicatescomment --"}
{"_id":"q-en-vscode-eace274717908d9a1cf27b03af1a9b22158f8e0b01a8018eba21b311693c945a","text":"The setting \"breadcrumbs.symbolSortOrder\" cannot be enabled for a specific language, i.e. \"[markdown]\": { \"breadcrumbs.symbolSortOrder\": \"position\" } is not possible in the settings. It makes sense to sort by position in markdown files to see headers and subheaders as they are. However, one might want to display symbols by type when writing code. Thank you in advance!\nThanks for creating this issue! We figured it's covering the same as another one we already have. Thus, we closed this one as a duplicate. You can search for existing issues . See also our guidelines. Happy Coding!\nActually not a duplicate but should be straight forward to implement as filter-options can already be defined per language\nCould you point me towards how to get the current resource from the editor model inside ?\nThe  has a reference to the text model\nto verify, have this setting and a file with this content (symbol should be sorted alphabetically) Now, paste the same content into and make sure symbols are sorted by position"}
{"_id":"q-en-vscode-2788b7860a6fe84c04ff446d0b67506f64215cbf780377e8f0d6662eff0ba10b","text":"insiders, install remote SSH extension the Remote view container all the views via context menu menu on the title -the views are not there and they can not be brought back -the views are back -I would expect the hidden state to be preserved fyi !\nAssigning to as remote viewlet has special treatment.\nOriginal decision was that none of the views in the filter-view should be hide-able. Looks like I missed a spot!"}
{"_id":"q-en-vscode-d9c611348efd100569cc7cce7629399b011130480e9173ddad712cbc676ea2d2","text":"VSCode Version: 1.41.0 OS Version: Not applicable Steps to Reproduce: to row 231 that the word 'must' is unexpectedly repeated. Does this issue occur when all extensions are disabled?: Yes\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.555 --- <!-- score: 0.539 --<!-- potentialduplicatescomment --"}
{"_id":"q-en-vscode-8a3d93bac0e85e9030db26940647015adc55ed5f2a0273dee84238268fbc38a0","text":"As mentioned in , the goal is to improve the first impression of semantic highlighting as well allow themes to opt-in to semantic highlighting. give themes a way to opt-in to semantic highlighting enable semantic coloring by default only the built-in themes. All other themes can be enabled by the theme author or in the user settings. show a notification when semantic coloring shows up the first time. The notification will point to the docs with background and information on how to configure the feature. fix the bugs found related to wrong token classification and color lookup. no longer show semantic color in imports\nI like this feature. I think an individual config is better than a global one because it will allow us to use non-maintained themes that have semantic highlighting disabled and themes that have it enabled, without having to set the global config every time we change from a theme that uses it to one that doesn't\nIt's a good default setting\nWhat are the reasons? It would be nice to know what most theme creators think, not just a few. Maybe some of us would like to use it A configuration could be to enable or disable that feature:\nChanges pushed.\nTo verify: create TypeScript file . If semantic color is enabled, , and all get the function color Open a built-in in theme, semantic color should be enabled verify that the very first time the semantic colors are show, a notification dialog is shown Install a theme from the marketplace (e.g. ), semantic color should be disabled. try and semantic highlighting should be enabled verify\n~Now that semantic highlighting is opt-in on a per theme basis, is there any way for a user to override that? I use Monokai Pro, which hasn't been updated in quite some time, and the developer seems fairly inactive now. The semantic highlighting was working great on 1.43, and I'd like to get it back now that it's been disabled in 1.43.1.~ ~I tried adding~ ~to my user config, but that didn't work.~ Edit: The above actually did work, except that semantic highlighting doesn't work on imports. I see now that this is the intended behavior. But why is semantic highlighting no longer available for import statements?\nMake sure to also remove\nSorry, I was able to get semantic highlighting working by setting the config in my previous reply. But I'm wondering if it's possible to re-enable semantic highlighting on imports. On 1.43, variable names in import statements in Javascript would be highlighted based on their type. But on 1.43.1 with semantic highlighting enabled, import lines are no longer highlighted, as noted in your original message."}
{"_id":"q-en-vscode-a1e6f4a9999f2d4c8067490d0315b859841126d54a3ca1fd42e9cb87ba8eadf1","text":"Testing Why is not there a contribution point for timeline provider? Following are the advantages I am not sure how I have to activate my extension if I only contribute timeline provider. Having a timeline contribution point can allow me to activate my extension when user starts using the timeline my extension contributesAlso helps in displaying the feature contributions in extension editor view\nand I talked about supporting this originally but felt that it was something that could be later.\nHm, not sure... Somehow timeline feels like master-detail, e.g you open an editor and the timeline shows its contents. With that usage-pattern it feels wrong to have an activation event. Then there is the 'Open timeline' command which is different... You could that command as activation event and all is good\nOn what event shall an extension contributing a timeline should be activated?\nFor instance these and\nshould we support an activation? Rather than a really specific point just for the timeline?\nYeah, that would make sense\nTo verify ensure that your extension that contributes a timeline can activate properly when the timeline is revealed via the activation"}
{"_id":"q-en-vscode-e26c10db1492448a903a8260435bc166d3d24de9c0106b181366898660e98e5c","text":"Issue Type: <bBug</bI moved all views in panel to sidebar and VSCode crashed the first time I pressed ctrl + j to show panel. I can consistently repro this issue. VS Code version: Code - Insiders 1.44.0-insider (, 2020-04-03T23:31:33.481Z) OS version: Windows_NT x64 10.0. <!-- generated by issue reporter --\nTo Verify: all panels to the sidebar (panel will hide) Ctrl+J that the panel opens to an empty view e.g. (empty Terminal View) and that you do not crash. We will iterate on a better UX next milestone.\nI made this change as the safest option for now. If you have a better solution that warrants another fix, please take over. Otherwise, will roll with this.\nYah looks the safest fix for now. Seems panel is not ready to handle no active panels. Is there a follow up issue for this created?\nbut the panel does handle no active panels if you run the command instead of (<kbdCtrl</kbd+<kbdJ</kbd) !\nYeah there are some issues with it and it is not working always. Hence we did a safe fix for now. We are working on this with proper UX for next milestone."}
{"_id":"q-en-vscode-9fd932f7bd69a8995d71967120fe6a03d3017c928fe98a76869946a723a4cb4b","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.44.0 OS Version: PopOS! 19.10 Steps to Reproduce: Not necessary, it's just a typo._ <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No Yes. The description string for is , should be . I know this is a very minor cosmetic issue, but I thought I bring it up anyway."}
{"_id":"q-en-vscode-25bdf6104af2e1fe18d7d66c97ef98b3d2496a4e5e25d097d7777aee00da980b","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.44.1 OS Version: Linux 5.6.4-arch1-1 SMP PREEMPT x86_64 GNU/Linux Steps to Reproduce: a symbolic link to the VSCode starting script with a space in the name. For example: \"My VSCode\" from the terminal Error is printed to the terminal: All in the starting script () should be enclosed in double quotes to avoid this error. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes\nI can take this!\ngreat!"}
{"_id":"q-en-vscode-2be968533643682327ab6173eb0b62d9cdaeecc862712f2cddfe5fc968066c26","text":"This is somewhat the opposite of and the acknowledgement that computing additional text edits for all items upfront is too demanding. This issue is only about and the goal is a pragmatic solution, similar to the one that we have for . Idea accept an unresolved completion proceed with insertion as usual set up a content change listener resolve the items once resolved check if additional edits overlap with \"concurrent\" changes. when they overlap, do nothing otherwise proceed\nThe first version for this is now merged to master, it's a pragmatic approach that's better then not resolving but worst than (costly) upfront resolving. Items to followup on [ ] telemetry and follow up issues [ ] - TS issue for having incomplete items [ ] [ ] java for returning very few (resolved) completions, e.g doing the right thing\n/cc - My understanding is that Java resolves all items upfront, e.g compute additional text edits during \"provide\", and that it is quite expensive and therefore Java only returns a small amount of items. Is that correct? Could Java return results faster when additional text edits are computed during resolve (and therefore only for selected items)?\nYes. I think so! is the dev who was working on Java completion. I believe he can provide more insights toward this topic. // cc\nthis is exactly the behavior that used to have, ie compute all the completion items first (fast), then the text edits were computed during resolveCompletion. We were asked to not do that as and changed to return a small subset of completion items (with incomplete=true), with full text edits computed upfront. The user experience is definitely not as good as with our initial implementation.\nTo be clear, computing any Java textedit upfront is costly when returning 100+ items (and we can return several thousands items). We'd prefer if the async insertion of textedits covered all textedits, not just the additional ones.\nThe requirements discussed in are still true and there are no plans to change that. There is (which might be derived from ) and . The former () is inserted at the cursor and must be inserted synchronous to accepting a completion (we cannot block typing while waiting for that). The latter () is meant for things like import-statements and we now want to relax that, e.g it can come late and we try to apply those edits despite the user having typed more.\nWhen exactly does run, is it only when an item gets highlighted in the menu? Can't tell from the docs. I only need to generate for the completion item the user ends up applying -- doing it for other items is unnecessary work that's bogging down the extension host, and I will have to implement yielding. I assume no one would need to generate before a completion item is applied? Is there any kind of completion applied event we can hook into to work around this for the time being?\nThat's exactly what is about . Today, there is no guarantee that is ever called because its original design is to provide data for the suggest details widget. However, we see that (amongst other things) are computed during resolve and leads to subtle bugs (because don't always call resolve or always await it) Well, is the problem if you do not generate them upfront. select suggestion synchronously inserts the string starts for additional text edits continues to type quickly and all over the place -call eventually returns but it is unclear if the edits are still valid because the state changed many types in-between Solutions: Won't work because in JS you cannot wait for another \"thread\" or \"process\", the world always keeps turning Implement a full -solution because this is two entities \"typing\" concurrently in the same document. However, it will feel awkward because the user doesn't expect to be such a scenario. approach that tries to apply additional edits if possible but inserts the \"main edit\" (insertText) always synchronously. () \"if possible\" is currently defined as \"no edit has occurred before the first additional edit\". That works well if you use additional edits for things like imports or prefixes (which is the indented use-case) but won't work if additional text edits change the main edit, like appending a call-operators with arguments etc\nI mean JS can't block but ignoring typing while awaiting is possible -- I was imagining that VSCode would keep the completions menu open and ignore any typing until is resolved and all text edits applied for the accepted item. Obviously this would be bad if resolving takes a long time, but the API has cancelation tokens already, so the UI would just need to show a spinner/cancel icon, and also cancel if the user presses ESC or navigates to a different editor, for a decent UX. I think this would be less confusing for users too; it's the only way it would be clear when VSCode is completely done applying the completion. I think many users would be confused if imports usually get but not when they resume typing too soon. FWIW I can syncronously generate the additional text edits I need, so I'd be happy if there were an API to synchronously generate them when the completion item is accepted. I know not everyone would be able to generate them sync though.\nAlso I know my questions may have seemed redundant, but I hadn't seen anyone on this thread point out that it would be ideal to avoid ever computing additional text edits (or even the primary insert text) for unselected completions. Since can get called before any item is accepted (I assume, since it seems to be for showing additional details in the menu), I'd think any solution via could still result in unnecessarily generating additional edits that are ultimately discarded.\nThe golden rule is that \"you can always type inside the editor\". Ignoring typing is not a good solution and will bug users. Just think of quick suggest and auto accept characters - you don't want the editor to stop typing just because an extension is taking a little longer (+50ms). To be precise you should have written \"typing at or before the offset at which additional edits are going to happen\" I honestly think the majority can generate them synchronously but this happens in the extension host process or the language server process, and all of that might happen on a different machine and there is no way to synchronously transport that into the UI.\nre with the current changes you should be able to have a \"simple insert text\", say , and compute an additional text edit, like , that goes in front of it so that you eventually have . Right? Or am I missing something?\nthat might actually be a good idea ;-) In practice though, we might need to revisit the way the principal TextEdit is computed (i.e rewrite a bunch of stuff from upstream JDT)\nI get the desire to be able to always be able to type inside the editor and agree that's an important principle. The only reason I think this case is an exception is that: The user would probably want to wait for all edits to get applied anyway. Seems unlikely that people would want to get the main edit without the additional edits By making the operation cancelable, you can still go back to typing immediately if you so desire. I certainly wouldn't want to block the user from typing without recourse (as you know, I feel frustrated that happens currently when using Vim extensions) On the other hand, I guess it's unlikely that the additional text edits would overlap whatever the user continues typing...\nI hadn't thought about auto accept characters though, that's a good point. That's something I don't have experience with. Also considering that VSCode could pop up a warning if it has to bail on inserting additional edits, it's not a big deal if it ever happens\nI remember that we have some new data points regarding the time consumption in different stages of code completion in Can you share that data and see how we can tune it with this change?\nFor Java language server, below is a sample of call tree when triggering completion. (time value is larger than expected because of extra overhead from JProfiler) ! Total time of completion stage mainly consists of three parts: competion candidates in JDT to CompletionItems (calc textEdits, additionalTextEdits, etc.) --- calc an incomplete list of only 50 items by default (the full list can contain thousands) --- here calculating import statements costs a lot of time server-side code snippets (not much related to this topic) I did a simple experiment on java language server, with vscode-insider built on 2020-5-15. I was using spring-petclinic, which is not a big one, but with some dependencies. Type a initial letter, e.g. \"O\", and observe perf of completion and resolve. Do multiple times for statistics. Calculating additionalTextEdits async in resolve stage does decrease the time cost a lot. textEdits and additionalTextEdits are calculated during completion stage. completion: ~150 ms resolve(avg of top 10 items): ~10 ms Now I simply move a bunch of textedit calculation from \"completion\" stage into \"resolve\" stage. completion: ~100 ms resolve(avg of top 10 items): ~17 ms !\nI just realized that can return a thenable, but I assume that would delay the appearance of the entire menu?\nClosing as we haven't seen issues with the new pragmatic approach and the plan is to stick with it.\nso how do servers know that clients support lazy-loading of additional textedits? Surely this should be a client capability advertised by vscode (that needs to be specced in the LSP)\nis the right person for that and you should probably create a follow up issue in the LSP land\ncorresponding LSP feature request created here please add more details if I miss anything.\nThanks\nTo verify: use TypeScript have a suggestion that will import a symbol make sure that suggestion isn't first or focused in the UI use the mouse(!) to select the suggestion make sure the import is there make sure undo, redo works as expected"}
{"_id":"q-en-vscode-cb843a5ed26154e74d3b563dbd122c1e9af1f503040ba8bbf5ebaf661580df80","text":"ERR LinkedMap got modified during iteration.: Error: LinkedMap got modified during iteration. at LinkedMap.forEach (file:///Applications/Visual Studio Code - ) at g.cleanUp (file:///Applications/Visual Studio Code - ) at g.saveState (file:///Applications/Visual Studio Code - ) at file:///Applications/Visual Studio Code - Code is here:"}
{"_id":"q-en-vscode-cf365a15576edae35f742d52eca205a3671d61ddbca495c64b93d31d9a6c2fd4","text":"see gif open GH notebook with a few cell that reference a variable run peek references and select entries notice how notebook cells are being highlighted, like \"Milestone Work\" get selected !\nI can repro in Insiders but not master, I'll check again tomorrow\nThis happens because of this listener: Tho, not yet sure why the selection keeps changing... Maybe some inner list event (from the references list) bubbles up to the outer list\nThat listener is fine, it's the nested lists\nHey do you have any thoughts on handling this? We have a list inside another list. The inner list handles the touch event properly. The outer list then handles the event and picks up the wrong index because it finds the data attribute inside the inner list in We can Make sure the event is fired on the outer list with the correct element, by fixing to check whether it's in the right list Not fire the event on the outer list at all, when we see that it originates in a different list Add some property to the list event that says it originated in a different list so the consumer can decide In this particular case, I think any of these are fine\nThinking more, I don't think that the event should ever be handled by two lists at all. Meaning we either mark the browserEvent as handled, like with preventDefault/isDefaultPrevented, or we make the list check every event to be sure that it originated in this list.\nThis sounds good to me. This also sounds good to me.\nrelated when I do peek references, I see markdown cells being selected? ! Notice I'm looking for under-discussion but \"Configure\" is being highlighted as one of the references\nFrom Good catch, I didn't notice that\nThanks for the quick handling!"}
{"_id":"q-en-vscode-b683ac1573953ce4d5fe1bcea18edffa26df02739619ba8a11d81ceb11185f70","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --Currently the VS Code Git commit message box automatically synchronises with the contents of .git/MERGEMSG (provided you have not entered your own commit message yet into VS Code). I would like to request support for VS Code to do the same with .git/SQUASHMSG (produced when running ). This would provide better integration between VS code git related VS Code extensions which often have the ability to perform a squash merge via the extension GUI but do not typically replicate any of the existing VS Code git interface (commit message box and changed file list).\nI wouldn't mind taking this on, and I've looked through the code so I have a pretty good idea of what's required to implement this. Just one question, though — is there any situation where and are both present? If so: which one do we think should take precedence as the 'default' commit message?\nmy suggestion would be to follow what git-gui does since it's bundled with git. Looks like it preferences MERGEMSG over SQUASHMSG (see )\nThat makes sense, and that's what I was thinking too. I'll put a PR in shortly!\nI got this to partially work by: a file with some content like \"HELLO\" a local branch file sometimes gets filled with content, like the following it seems like the content of the file is not constant for me, and changes once I do a squash Reopening this for discussion. I also might be doing something wrong. Bonus bug: when chaning the content of the SQUASHMSG the scm input does not reg resized and ends up looking broken (picture attached). So I would recommend that we revert this PR and polish it a bit more. !\nIt worked fine (and always displayed the intended (?) squash msg), but maybe it's because I was only squashing another branch that only had one commit? I concur that it seems like the contents of SQUASHMSG aren't always consistent though, now that I've tried with a branch with more than one commit. EDIT: It seems like it displays all of the squashed commits in one message, prefaced by As for the bonus bug, is that issue exclusive to SQUASHMSG or does it also happen if you change the contents of MERGE_MSG?\nYou're not supposed to manually create or edit the file. This is a file created by git during the command. When you run this command you will have a commit pending... and the input box should be prepopulated with a template commit message. Can you try this again? Thanks for the input box issue though, will try to repro.\nThis was the intended behaviour behind this feature request. I'm not expecting VS Code to do anything magical here with SQUASHMSG. It should just be taking the entire contents of the file as other graphical git tools do (and as a command line call to does if SQUASHMSG exists). I do not think the PR needs reverting.\nAdded another commit, since the can exist but just have no content. In that case we should take . Please verify once again, now the feature is clearer.\nSorry for the confusion, just was not very clear to me. I will verify this later today.\nAdding verified label since the git commit gets nicely preopulated with the squash message."}
{"_id":"q-en-vscode-56926236e2ff478ac1bda76c1146e1b2eb5032208e70e0784150e21a56cef01d","text":"VSCode Version: 1.47 insiders OS Version: Windows Steps to Reproduce: on the \"Add item\" button in and press enter. ! Does this issue occur when all extensions are disabled?: Yes"}
{"_id":"q-en-vscode-440d87acb074cf59e2a29ebdb1f07227df7e3c47169f85422776fbe1eebfa496","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.47 OS Version: MacOS Steps to Reproduce: a local project server (f.ex angular) with a long filename (some 16+ characters) <filename.<extension. a debug session (f.ex firefox), and check the debug console output. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes !\nI don't understand what the screenshot is showing. Please show the whole debug console. I do not understand what gets concatenated in the screenshot. What exactly are you expecting and what is the erroneous behavior?\nattached photo of whole debug console - error occurs on last line. Behavior: supposedly filenames (some 16+ characters) will be shortened with (...) instead of showing the whole filename and line number. Expecting: the whole filename and line number where a console log is being outputted. !\nYes, I agree this is an issue. The problem is that we set a to 150 . The ideal solution would be that the elements gets shrinked first, and that it has unlimited size. Adding help wanted as I am open to review a PR which fixes this. If nobody tackles this I can also look into this in the future.\nHi I can give this a try.\ngreat, thanks!"}
{"_id":"q-en-vscode-84e2b29eeb7b1f6cf5c171ad9f3163f245bea1de4811bd4d8bc5ac6e1cdb2b63","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.47.2 OS Version: Windows 10 (2004) Steps to Reproduce: a dotnet project & to a problems pane and check \"Hide Excluded Files\" and open problems Expected: problems pane not to show problems in generated code in folder. Actual: all problems are still visible under problems pane but as expected not visible in the explorer. ! <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: No, the \"problems\" are generated by an extension. So cannot reproduce this way. Related issue:\nAFAIK the feature in Problems to exclude items uses its own filter, which you must enter in the field to the left of the filter button. I don't think it pays any attention to\nThis comment suggests that suppose to work: While I can use the Filter on the left, it is separate from \"Hide Excluded Files\" option. They don't interact.\nDeferring to\nClarification with regards to Problems are always visible not only when the file is open in the editor. I opened one in the screenshot just to show the error, but others that aren't open in the editor are still visible in the Problems pane too.\nOh alright, I realized I misunderstood so that is why I deleted my comment. I apologize for any confusion. Lemme see if I can figure something out today\nHi upon further investigation it appears that this feature uses a different list. I believe that it goes off of this list found in settings. But I do also acknowledge . could you maybe give me some input here? Maybe it can use both the global setting and the option together? <img width=\"981\" alt=\"Screen Shot 2020-08-07 at 10 13 42 AM\" src=\"https://user-\"\nthis also happens with php language server. is there a of some sort, like has? I have to disable altogether because doesn't stop vscode to fill Problems pane.\nI wasn't able to reproduce this issue when I tested just now on 1.52.1. Problems reported on a file whose name matched a glob in disappeared from the panel when I selected the \"Hide Excluded Files\" filter. They reappeared when I deselected that filter option, or when I removed or disabled the entry. If the problem still exists for you please provide simple repro steps.\nVery much appreciated for validating the issue again. Please provide exact steps to repro.\nHey I did a repro on a clean machine and I know exactly what the issue is now. It has to do with how glob pattern is applied in the problems pane. If you apply it on folder like this - it does not work for problems pane, but works in explorer view. If you apply it on all files inside the folder - it works for problems pane and explorer view. I was able to reproduce this behaviour in typescript solution too, so was a red herring. !\nThanks for the repro. Will investigate in February.\nPushed a PR with a possible fix"}
{"_id":"q-en-vscode-30d3e92cbda311ce4420166dda5c84f13369957e8045ce2e39e955b0e315d744","text":"Issue Type: <bBug</bWhen : one mouse wheel scroll in extension view: moves 100px one mouse wheel scroll in settings editor: moves 100px When : one mouse wheel scroll in extension view: moves 100px, with slight animation one mouse wheel scroll in settings editor: moves 15~25px, without animation I'm using a Logitech G502 Lightspeed mouse, without any scrolling tweaking softwares. ! VS Code version: Code - Insiders 1.48.0-insider (, 2020-08-05T05:30:25.429Z) OS version: WindowsNT x64 10.0. <details<summarySystem Info</summary </detailsExtensions: none <!-- generated by issue reporter --\nI don't see this, any idea\nI believe this setting is ignored for macOS. Duplicate of\nActually not a duplicate. Interesting. I need to get my hands on a mouse to see if I can repro.\nCreated a PR that fixes this:"}
{"_id":"q-en-vscode-fee729438a322534546b37650e250b9794183eb254a0648e5a893344e898e9b7","text":"Issue Type: <bBug</brepo your user tsdk to use this version, for example: load the VS Code workspace (making sure you are using your user TS version) a file in and try getting the document outline Bug Big delay. This is because two full TS Servers are being spawned instead of a syntax only server + full server (see memory usage in screenshot): <img width=\"1079\" alt=\"Screen Shot 2020-08-11 at 3 13 33 PM\" src=\"https://user-\"VS Code version: Code - Insiders 1.48.0-insider (, 2020-08-11T10:13:04.551Z) OS version: Darwin x64 19.6.0 <details<summarySystem Info</summary </details<!-- generated by issue reporter --\nSeems like this was caused by\nFor verification, try the following cases: Using TS 3.9.7 Using TS 4.0.1-rc Using TS Using an old TS version like 3.5 (older versions don't support the syntax server at all) All should load the document symbols quickly even while the loading spinner is showing You can use the same global install in all cases, just make sure to reload VS Code each time\nAssigning this to july milestone since it is a candidate.\nfollowing your steps I have verified that the outline get loaded quickly, however I still see two tsservers spinning up when I open the process explorer. If I understood correctly this is by design, since only one of them has a large memory usage. I have verified for TS: 3.9.7, 4.0.1-rc. ts 3.5 Thus adding verified label !"}
{"_id":"q-en-vscode-1f55725d3d3625dfc643244c916d02b1331fc45814b9a425cb9da3e77fdf2af5","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --In markdown mode, Dimmed Monokai theme: does not render bold / italic. no special color for link / image title. !\nThe markdown used in the screenshot:"}
{"_id":"q-en-vscode-183e4dd6f91c728d2bde6b7a637f46abc68372f2acc75752679b72cd7353fd93","text":"The usage of as seen here only works because there is there is special code to return a url-string (instead of a path, see ). For loading extension relative resources the and should be used. That is the new, web-ready replacement API\nfyi\ntried to adopt this. When will be available. I can't find it in .\nOpened for ESLint\nIt's in the API since 4 months:"}
{"_id":"q-en-vscode-2097c1ed4b4720227571e4f0017c86b63f34aa3c40f37a554622aef073f22cbe","text":"Since we don't know when links are no longer cached by , we cannot drop the objects. Upstream part:"}
{"_id":"q-en-vscode-d8d936b13b3acbb45607408043406fdd5172e77353d58a403187577b49a1df3b","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --The repository picker feature allows to pick github repos after selecting the GitHub provider. Asking for option to default the provider to GitHub so that user can have one less click when creating a codespace. it lists all repos where you are owner, is it possible to list repos of organizations you are in as well? A side question is how does the api authenticate, is it with the stored git credentials in git extension? Thanks, Jason\nSure, I can accept a PR here. But let's call it . Note that multiple providers can have the same name, tho that is a hypothetical issue, so no worries. The code that gets the repositories lives in the GitHub core extension: Maybe by querying for you'd get those repos? If so, we could add another option to the command . Authentication is implemented via the GitHub Authentication core"}
{"_id":"q-en-vscode-d0ceb1d2bb8c2c54722542b33941e4cd5dae1c375b8c11d1379be40a3c95fe78","text":"Below is a recording of a Control-R reload when no workspace is open and one non-builtin extension is contributing to to the Welcome view of Explorer. ! Observe how we start with Explorer's usual \"NO FOLDER OPEN\" view. Next my extension contributes its text and buttons. Finally the built-in git extension contributes text and a button. Since built-ins get higher priority the content from my extension jumps downward, perhaps while I'm reading it or trying to click one of its actions (link, button). There's no progress indicator when the content is being fetched and assembled, and it's hard to know how long to wait before concluding that no other extensions are going to contribute. VS Code version: Code 1.49.0 (, 2020-09-10T13:22:08.892Z) OS version: Windows_NT x64 10.0.\nSimilarly on the Source Control view, it starts with a message saying \"No source control providers registered\" and stays this way long enough for me to worry that something is broken or missing. Eventually the contribution from the git extension arrives. If a progress indicator isn't , maybe at least the initial message could say something like \"Please wait while we look for source control providers\". Then if nothing contributes (e.g. the git extension has been disabled,) it could then display \"No source control providers registered\".\nThis will be very hard to achieve. The problem here isn't actually the extension point, for which we could definitely add a progress indicator, but the activation of the git extension. That git section won't appear until the git extension is activated and all the potential repo locations have been probed... We simply can't add a progress indicator for that in the core unless we add some more API.\nOn this, we could potentially add a message.\nThat might help.\nClosing this until we get actual user feedback that it's needed, sorry"}
{"_id":"q-en-vscode-5d8d684e0ddc6fe6c4200376a869a59639e328e4a78e0b75f318641ddb67e502","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.48.2- -1.49.0- OS Version: Ubuntu 20.04.1 LTS Steps to Reproduce: -dev not installed and install code1.48.2- --VSCode works apt update; apt-upgrade VSCode to 1.49.0- --VSCode fails with: /usr/share/code/bin/../code: error while loading shared libraries: cannot open shared object file: No such file or directory install libgbm-dev --VSCode works Expected: upgrade checks for dependency and installs missing library\nnew Electron 9 dep needs to be declared here?\nI didn't know about this dependency configuration in the codebase, yes adding should be sufficient. It is a new dependency for chromium 83 to use memory buffers in gpu code path.\nThis bug has been fixed in to the latest release of ! you can help us out by confirming things are working as expected in the latest Insiders release. If things look good, please leave a comment with the text to let us know. If not, please ensure you're on version of Insiders (today's or later - you can use in the command pallete to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nTested this by upgrading from 1.48.2 to 1.50.1"}
{"_id":"q-en-vscode-f1909379a13d16df149b23bf1bb7a2e135c2232f10c77d159dfcb04b7cbe715f","text":"It does fire for folder on move and delete. According to the reason lies here: The code goes directly to the file service. Since we correctly report move and delete for folders I think we should report create as well.\nSome background: I want to adopt this in LSP but specing that this only works for move and delete of folders sounds a little silly. So I want to make sure that there is either a very compelling reason or there is some ack. that we will fix it in the future.\nWe currently have no related method such as for folders that the explorer could use here: If we support this, I guess one question in terms of API breakage is wether it is fine to let and emit events for folders too while they currently only emit for files. If we think this is fine, then I think this would be a relatively simple change to make.\nI don't consider that breakage, it's a bug fix esp. given that move and delete already fire events for both\nVerification: write an extension that registers to and and verify you see both events when you right click into the explorer and create a new folder.\nThanks for fixing!"}
{"_id":"q-en-vscode-9119167c4f8294f035fc1a9a9e00cf232793584e8138e0db4ba49afa3400a45c","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --This would prevent needing to convert the to a string only for it to be parsed back into a later on, and is friendlier to developers.\nproposed or finalized?"}
{"_id":"q-en-vscode-e9b4bfa229478cd12572af832214a1f3fb195e76e290258671308855c2a6257c","text":"Update: reproduces in web + desktop Steps to Reproduce: a Codespace until it is loaded various activity bar items that open menus (gear menu, accounts, main menu, home indicator) = somehow the accounts menu only opens for me after a delay, as if something was resolved when clicking? !\nThis still reproduces\ndoes this still repro? maybe its imperceptible to me?\nYeah still seems slightly delayed to me, can we not check the code and see if a longer running promise prevents the menu from showing?\nlooks like we fetch session info every time, so this probably adds some delay. maybe this can be cached and updated separately\nYeah that seems like a reasonable ask since it goes to the extension host and back, today.\nThis issue has existed since 2 years! But only recently have I encountered this."}
{"_id":"q-en-vscode-553e1ae93832a9eb05964b687b7920ffe6b60544ff0f65d9ab0d0bf7173ea2ee","text":"The find widget's update history is wrapped in a delayer promise which can cause an unhandled promise rejection when the widget is getting disposed of if the delayer's completion promise hasn't been cleanup up. This can happen if the editor is closed before the find widget has time to finish closing. This change addressed the issue by handling the rejected promise returned by . <!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.51.1 OS Version: macOS Steps to Reproduce: new editor. -F to open the find widget. to close the find widget and quickly close the editor before the widget has time to close."}
{"_id":"q-en-vscode-47f30727aa358653f75b2494291faddb6ea3cdf0b740273c2ae6cbcc86bacffc","text":"We have been seeing a bunch of remote issues where the remote extension host burns a lot of CPU. I think it is caused by our port scanning code, which executes even if the setting is false. See Possible issues: / Other: ? ? ? *\nLooks like we need to work on our indentation and ensuring that we break lines once they go past the edge of the screen for the sake of readability\nI could reproduce some increased CPU usage of the remote extension host using the project linked to i.e.\nIn insiders, you can now temporarily disable the port scanning code by adding this line to your .bashrc file, then restarting the VS Code server on your remote:\nOnce we have an insiders build I'll start commenting in each of the issues asking folks to try out the workaround in insiders.\nKeeping open for now so that the issue can be found more easily.\nClosing since the fix is available in insiders.\nVerified that cpu usage was reasonable when running the petclinic app"}
{"_id":"q-en-vscode-689bc5587c18b2ff1a79e2c33e01854720caa3dbd52c94444ecc4bbc8680e727","text":"VSCode Version: 1.52.1 OS Version: Windows 10 Steps to Reproduce: a WebView tab, such as the Markdown preview. click on it in the Open Editors Section, and select \"Select for Compare\". click on any other item and select \"Compare with Selected\". ! Does this issue occur when all extensions are disabled?: Yes\n//cc not sure we have a context key to use here already\nCode pointer We would need to use that context key for all the compare context menu actions in the Open Editors view\nisn't this just a matter of using for this command? We seem to be using this also to limit commands such as \"Copy\"?\nyou are correct, I checked and this is indeed the context key we needed here.\nWill have to revert this due to\nPR:\n/verified"}
{"_id":"q-en-vscode-0575e6505ade5bfd84e4fd2fddad3a69259a1a17238ee2f864ffaa07a465970d","text":"Testing Create multiple editor groups Ensure that the selected tab is not the first one in the editor group Start dragging the sash so that there is not enough room to wrap tabs Notice how the name of the tab that is visible jitters as it switches between the name of the first tab, and the name of the selected tab !\nYeah I was not able to figure out what triggers this and was even suspecting a browser bug (though didn't follow up). //cc\nI see this being related to: is called we enable wrapping we disable wrapping right after because we figure out the space is too limited somehow the scrollbar is now in a bad spot and needs a proper reset to avoid flicker\nI was able to find the origin of the bug in : wrapping is activated even when there is no space due to the actions right to the last tab. In the next call, it is disabled again, because the action's size is considered in that case. Constantly switching between wrapping and not wrapping makes it behave weird. I'll make a PR."}
{"_id":"q-en-vscode-190a888961b615eef7b7c2a70383fbc8ff9aaedf1d1f374c08c8b1524225e770","text":"Issue Type: <bBug</bRepro a terminal both terminals, run some kind of script that outputs text that horizontally fills the terminal resize the left terminal to make the right one smaller Bug For a half second during the resize, the content of two terminals end up overlapping https://user- VS Code version: Code - Insiders 1.53.0-insider (, 2021-01-26T05:12:23.287Z) OS version: Darwin x64 20.2.0 Remote OS version: Linux x64 5.4.0-1036-azure <details<summarySystem Info</summary </details<!-- generated by issue reporter --\nI think this is just a problem with the dom renderer?\nOh yeah, I had set. There's a slight amount of overlap with canvas renderer (specifically the sash border seems to overlap with a few characters of the terminal for a few hundred ms at most), but it is much better\nShould probably use on each pane, that should fix."}
{"_id":"q-en-vscode-11d52f54d243a5d9a68a23db0a2bc53f4c8b9280b1585c86dda390644c9633c7","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Also please test using the latest insiders build to make sure your issue has not already been fixed: --<!-- Use Help Report Issue to prefill these. --- VSCode Version: 1.52.1 OS Version: Windows 10 1909 Steps to Reproduce: then press to run extension any folder with some code. code lens show, then hover one of codelens. Actual Result: ! Is this an expected behavior of vscode extension? Or is there any configation to show it?\nI believe this is expected behavior, however forwarding to so he can close\nYes. what tooltip to you expect when hovering over a code lens?\nAs , show maybe suitable I think.\nAh, I see. That actually makes sense and would be a simple change\nCan I take a look at this issue and try to work on this?\nThis is where the dom-nodes for codelens are created: Using the title attribute on them should do the job\nIs the native tooltip sufficient for the fix? Windows & MacOS :-------------------------: <img src=\"https://user-\" width=\"400\"Windows | MacOS :-------------------------:|:-------------------------: <img src=\"https://user-\" width=\"400\"| <img src=\"https://user-\" width=\"400\""}
{"_id":"q-en-vscode-32d048a044e14f0dff11cf04aedd5d5cefb178cfa35cac3431fa113c2a20da08","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --- VS Code Version: n/a OS Version: macOS 10.15.7, npm 7.5.0 Steps to Reproduce: to the and read the instructions in the \"Integrate\" section: a terminal and run the command as indicated: . Expected outcome: installs the freestanding JSON language server. Actual outcome: reports the following: I have attached the cited log file. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: n/a <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported direct to the extension publisher. The 'Help Report Issue' dialog can assist with this. --\nSorry, that was a typo in the README. Please try\nSuccessfully installed using , thank you very much!"}
{"_id":"q-en-vscode-43ccf1ff568d3824b6d318a47879c0181ecdd22a7bae6fc98d1be589f30d2511","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --- VS Code Version: 1.54-Insiders OS Version: macOS open a GitHub issue notebook cmd+shift+o navigate down only code cells are highlighted https://user- <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported direct to the extension publisher. The 'Help Report Issue' dialog can assist with this. --"}
{"_id":"q-en-vscode-23938cb97387e4fb862bcce07d8b46e9f1ae12cd7e5854b51900e16763411404","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --- VS Code Version: 1.55 OS Version: MacOS Big Sur The unit testing readme instructions for running unit tests from node is missing the --ui tdd flag and doesn't work without it. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported direct to the extension publisher. The 'Help Report Issue' dialog can assist with this. --\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.453 --<!-- potentialduplicatescomment --\nThis is about\nYeah see pull request sorry. I just lumped it into something else."}
{"_id":"q-en-vscode-88350a42f9f541f0cf2750e47dc3294e1fa74ecb2ea736f26f6876886e44e4fd","text":"The call was removed in\nIt's weird, seems like in stable we should be leaking terminal processes then but I see them shut down immediately after closing the window in stable..."}
{"_id":"q-en-vscode-469729638394f349dcc49a42438b90e527e8c963a26301debedcbcbd4521c4d2","text":"On Windows: a local empty folder an empty a with the following contents: the Markdown preview the link File opens with a strange input, as if its entire path is the file name. Also, if you open from the Explorer, it now appears twice in Quick Open: once normally and another in the weird long format. !\nwindows? local/wsl? Doesn't repro on macOS so I suspect an issue with how URIs are being created from the markdown.\nThis is local Windows, updated the steps.\nThere is some weird encoding going on with that URL, here is what it should look like: And here is what I get:\nI'm working on this issue. I've logged out how the Markdown Preview renders the paths to HTML with \"Click Here Absolute Path\" containing the same link as in the issue: Markdown: HTML Output: The paths are correct though. Plus I think the link should be given as a relative path (ie. ) instead of an absolute path. Giving the absolute path here () is not correct. Btw, I've tried on other IDEs for comparison, they all don't open the absolute path link properly (or open nothing at all). The relative path link always works. Not sure how would you want to tackle this now?\nI have a hunch that maybe the issue is somewhat related to\nThanks, for the pointer. I misunderstood the issue at first..."}
{"_id":"q-en-vscode-259030ef976ff6eaa275d8109d488f1a36d9993f2127759c8edfd102556029f7","text":"skipped in\nSeems to only flake on Windows\nCouldn't reproduce locally\nLet's try re-enable in debt week"}
{"_id":"q-en-vscode-d77133b6169516e818508357d37a777ee15ec18977ee9a706ad14b11eb4c4be2","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --- VS Code Version: main branch, commit (28 march) OS Version: Linux Component: notebook When a user adds a new cell in a notebook, this created cell is automatically selected. This selection happens in the class (): However, when the user undoes and redoes this action, this new cell is not selected any more. This may be considered as an issue since undo/redo must be an identify function. I think that this issue takes sources in the method () where the current selection operation to do after the cell creation is . If I replace this arrow function by: and I removed the block of it works as expected (build, tests, integration tests OK). Steps to Reproduce: To reproduce the issue automatically, adds this test in (executed by ): I will open a PR to detail the proposal. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes (in development mode). <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported direct to the extension publisher. The 'Help Report Issue' dialog can assist with this. --\nI created a cell, and then undo/redo a bunch on that cell and the selection isn't always the recreated cell\nany chance you can do a screen recording of the reproduce?\nCan't repro this anymore. Marking as verified. Don't know what happened before. Although there is some screen cheese on many undo redos it's very brief."}
{"_id":"q-en-vscode-541725f4084fa3aebff54fe47b6456fcb61a9e2a363c955e49bf6712c81f9383","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --- VS Code Version: 1.55.0 OS Version: Windows 10 Home 20H2 .867 Steps to Reproduce: run latest version of VS Code, 1.55.0 the list of default WSL profiles and you will find something like \"Ubuntu (predet.)\" instead of just \"Ubuntu\". profile is impossible to launch, because there is NO distro named \"Ubuntu (predet.)\". This happens because , at line 178 of latest snapshot I'm checking, does this: . It should remove other locales, too. I know this is difficult to do because... well, there's plenty of locales. So, in order to get the distribution name, another way should be used. I think it's safe to remove any last element within parentheses from the distro name as returned by , but really another method is available which I think it's safer, but I'm not sure how to write it in TypeScript. I have the following code already written in Python, for a system maintenance script I use in my own system, where I retrieve the installed distribution names from the registry: It's a partial code but well, I think it's self explanatory, and I can provide further information if needed. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported direct to the extension publisher. The 'Help Report Issue' dialog can assist with this. --\nSame happens to me with the new version 1.55.0 One temporal solution that I found is to disable the autodetection of the distros and include each one with the profile setting:\nhi, at first glance it looks that adding a -q argument to the 'wsl -l' command would be safe... regards\nThanks for the suggestion and the fix. Still, I'll keep an eye on this, just in case Microsoft decides to change command line interface. But so far, so good, and it's simpler than having to read the registry\nJust encountered this issue today (after updating Docker for Desktop, but not sure if this has something to do with the issue).\nVerified in"}
{"_id":"q-en-vscode-2cc856f740782963019fedcda4ea658c2e9868c575554b8454755b828d96ec0c","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --- VS Code Version: 1.56.0-insider OS Version: MacOS Big Sur 11.2.3 Steps to Reproduce: Settings and and editor it manually. a rule: editors can't be edited anymore. <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported direct to the extension publisher. The 'Help Report Issue' dialog can assist with this. --"}
{"_id":"q-en-vscode-0d29eb0029a7c980c02d3e8be2d990cf255e12dd35a62b67ef1bdd21e2d96ce4","text":"editor hover over any link correctly uses and ignores ! VS Code version: Code - Insiders 1.56.0-insider (, 2021-04-16T11:25:09.662Z) OS version: Windows_NT x64 10.0. <!-- generated by issue reporter --\n(Experimental duplicate detection) Thanks for submitting this issue. Please also check if it is already covered by an existing one, like: <!-- score: 0.57 --<!-- potentialduplicatescomment --\ncan you explain the issue/feature to me a bit more? The way I am understanding it right now, is the behavior I am seeing, so I think I am misunderstanding something. I have this in my settings. This is what I see. Where/when am I expected to see the blue text? <img width=\"939\" alt=\"Screen Shot 2021-05-07 at 6 43 02 PM\" src=\"https://user-\"\nwhen you mouseover the link, like here: !\nthanks for clarifying, submitted a PR for this."}
{"_id":"q-en-vscode-c6c1790dd1e77f83faf11aaf68f2008df2f34693617bf842295ae43a40155726","text":"If you \"refresh\" an item (with children) by calling and the item is not expanded, the method still gets called. should not be called until the item is expanded, otherwise potentially expensive operations can be run for no visible reason (since the children are hidden).\nThis could result in unexpected behavior. If an extension currently relies on to always update the children then there would be out-of-date tree items.\nHow would this result in unexpected behavior? If the tree item has not loaded its children, why should calling refresh on it force its children to load?\nI'm concerned that some extension authors have come to depend on the current behavior to keep their internal tree structure updated. I suppose we can try it and see if there are any complaints.\nI think a change like this would need to happen in the , as at the custom tree view level we have no knowledge of whether a node is currently expanded or not. Is this something you'd be interested in doing?\nNot sure I follow. What describes is exactly how the AsyncDataTree works: children aren't resolved until the user expands the parent the first time.\nI am looking for a way to call on the AsyncDataTree such that if the tree item passed in to is collapsed, it will: \"invalidate\" the children or the tree item and not actually call on that tree item rerender the tree item instead of: calling for the tree item rerendering the tree item\nOh I guess I was wrong, what I said only applies to the next level. But the first node is called with will always have its be called. We can maybe add an option here, or even change the default behavior as I see no reason for collapsed tree nodes to need a call.\nany update here? This issue causes GitLens to request way more data than desired in multiple cases (before the user expands them that is).\nThank you!\nThanks for the nudge!\nMaybe it's easier if you verify this one.\nwill verify.\nFYI, I tested GitLens with this change (in today's insiders) and it looks good to me!"}
{"_id":"q-en-vscode-1b8c42d4394fd18a3090754bf9b1f1b0a218c467cc1642e61519fcec78528f8c","text":"If the knows which languages a user tried in their first day, it could unlock more specific questions to further improve the language onboarding experience. How though. tracks the edited count for languages that run surveys; which are the languages we care most about right now. Only issue is that the storage key isn't exported, so the doesn't know about it. any general thoughts on the idea, and best practices on how to make the storage keys known to the ces survey code; so they can be read and attached as params to the survey URL?\nThe CES code already has access to the same . Can't you copy the maps the surveys to storage service keys over to CES and read them from there?\nSure, was mainly worried that this creates some copy/pasta. Will take that approach for an initial pull request draft and gather feedback.\nNo worries. We can also extract the common code to a new file, sibling to those two.\nare there verification steps for this?\nSorry that they are so complex and time-intensive. I can also verify that I see survey responses from insiders that include the language data. a fresh VS Code Insiders instance , ensure that you are in the treatment group a Java file and save it. VS Code open for 1 hour to trigger the new-user survey popup and click to open the survey that the URL it opened contains\nOk, see you in an hour!\nThe treatment service does not seem to like me, for some reason. Unmarking as verified for someone else to give it a shot. Note that you need to repeatedly open VS Code windows with new user-data-dir's until you get into the treatment.\nAlso attempted this but no survey after an hour. Confirmed I did get an initial\nwhen the experiment enables the survey you should also get a event? Did that work?\nTesting again\nWorked! [redacted]&javaLang=1"}
{"_id":"q-en-vscode-a10309cc7a6b7da8591654ce38eeabd2897dab5ab5a163aca9c826814644cffd","text":"When changing the default editor we show a in the quick pick to change this. The gear action by itself is not clear that it will immediately change the default and assume there would be another screen to confirm. ! Change the flow to match how you select different formatters to make it clearer/explicit. In the formatter case, there is a \"Configure Default Formatter...\" that is clear and indicates that user input is required on the next step. The same can be applied to default editors. ! cc"}
{"_id":"q-en-vscode-eb2e75d4a3e4463386e8ce868780bf31424749a940adcc473341a87a707ca2be","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --- VS Code Version: 1.56.0-insider OS Version: Windows 10 Steps to Reproduce: new blank notebook cells only are missing bottom padding: ! <!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --"}
{"_id":"q-en-vscode-966ba5ff57b3e9a507ca555651d9b56ffb7295d00a9eda2b98964e0146001a74","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.56.2 OS Version: Darwin x64 19.6.0 I was confused by the QuickPick API as and were commented \"This can be read and updated by the extension.\" while wasn't. Then I was trying to dispose QuickInput and create a new one to update items, but after I read the source code I found can be directly updated by extensions\nverification: look at the PR. It updated a comment."}
{"_id":"q-en-vscode-aa2e718fa516c3ea8453d0a08df8c23d6d4afea3bd9974eacc1ff983d87dd4b0","text":"Issue Type: <bBug</bI have a multi-root VSC workspace with three folders: A/ ./subfolderOfA In A/, there is a file and in there, there are the following settings: When I edit some file somewhere inside A/, it seems like this JSON schema file path is only properly resolved If I remove folders and from my multi-root workspace. Shouldn't this work? VS Code version: Code 1.56.2 (, 2021-05-12T17:13:13.157Z) OS version: WindowsNT x64 10.0. <details<summarySystem Info</summary </details<details<summaryExtensions (10)</summaryExtensionVersion ------ bracket-pair-colorizer-20.2.0 xml2.5.1 tortoise-svn0.1.1 python2020.9. vscode-pylance2021.5.3 jupyter2021.6. cpptools1.3.1 code-spell-checker1.10.2 open-in-browser2.0.0 vscodeintellicode1.2.14 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nHi Unfortunately, I am not familiar with your planning process but I wanted to ask if this bug will be solved in the near future?\nSorry, I don't know when I will find time for looking at this.\nIt's honestly quite a shame that this feature is missing. Unless you want to modify the json file structure to include the \"$schema\" property you are stuck having to open the given sub-folders directly which is a significant detrement to the DX when working in multi-root workspaces.\nHi Sorry to pick this up again but, according to the motto \"New Year, New Luck\", is there a chance to have this and fixed soon?\nIt seems like relative schema path resolution does not work on Windows, even when not in a multi-root workspace. Schemas always try to resolve in rather than the workspace folder. !\nNote: relative path can not be used in user settings. But they should work in workspace files, where I could reproduce the problem, thanks !\nNow on macOS, unfortunately it seems like json.schemas in the workspace settings are ignored. <img width=\"1605\" alt=\"image\" src=\"https://user-\"Changing to something invalid does not result in any diagnostics being reported, either. However if I move into my user settings and change the schema path to be absolute, it works.\nNow I get in the Problems view with the following setup: In a non-multi-root workspace:\nI created a separate issue to investigate as the error comes from to the empty which is unrelated. If you remove the you should see that the - schema is used\nTo verify: clone open workspace look at the workspace settings: all files are associated with schema . Note that this is a relative path to the workspace file open both files and check that code completion gives back property . You can also click on the JSON language selector ( next to JSON in the status bar) to see the schema that is associated look at the folder setings: is associated with schema . Note that this is a relative path to the folder open both files files and check that code completion gives back property , but only when in folder1. You can also click on the JSON language selector ( next to JSON in the status bar) to see the schema that is associated\nI tried to verify this using build but the schemas are not used. Here is what I see in the UI: <img width=\"706\" alt=\"image\" src=\"https://user-\" please advise which build I should use for verification.\nLooks like it is still not working on Windows due to to an URI encoding issue.\nI tested as described in with VS Code 1.76.0 on Windows. I observed the following for and in VS Code: no schema-defined code completion is offered there is no schema offered when clicking on the language selector I then tested folder-only settings by opening VS Code to . I observed the same as above for . FWIW, describes another repro with only folder settings (not workspace settings). Is there a known workaround to get validation working using on Windows for 1.76.0?\nthe fix for Windows is coming:\nis now also fixed (and verified)"}
{"_id":"q-en-vscode-6c885d6767e8766bffa9b0c62cb256f38d0cc9e844e489f771d4f2f91cc5b721","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.56.2 OS Version: MacOS Catalina 10.15.6 Steps to Reproduce: VSCode Version any typescript file (for example in ) /Ctrl + Click on any native dom method, for example , there should be a error message popups ! I thinks this is caused by the wrong type definition file path in below code:"}
{"_id":"q-en-vscode-f10a08ae24640811ce708bd7e3cad42fdb146c2c86371b00a9776a9264d66fbb","text":"The recently remote indicate tooltip helps users to learn about the limited extension functionality when in virtual workspaces. That allows us to reserve the banner for information on the restricted mode and avoid the complicated sentences when restricted mode and virtual workspaces are on at the same time. While it is interesting information that is limited extension functionality when in virtual workspaces, it's IMO not that crucial to warrant a banner. Also there's nothing the user can do about it.\nFYI"}
{"_id":"q-en-vscode-1f90303c64f30e0b19fea8701836013282baf91b0f53e5234782594f82935b42","text":"Since profiles are removed, the jupyter extension will show their \"run above/below\" commands unless we put them back in the cell toolbar by default, so we should just disable this setting until we have a jupyter profile available that disables it"}
{"_id":"q-en-vscode-a5840eab3d3b2d5b756383939d428c580fa4b4adc338149692282fec6e9a86fe","text":"Not sure if this is a German only problem <img width=\"440\" alt=\"capture\" src=\"https://user-\"\nThis is probably a generic issue. Command should not get translated. Is there a way to ignore a substring of the localization string so that it doesn't get translated? I guess we could use interpolation...\nThis is a generic problem. Here's pseudo language: !\nAny specific verification steps? I tried changing my display language (to German) but the workspace trust page doesn't seem localized yet: !\nVerified this by looking at the language packs in the vscode-loc repo. It's currently in English but translations take some time."}
{"_id":"q-en-vscode-2f1fba3c2a4748113196c418a144be909f45663ba712680ca406cacfa67d01af","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: Insiders OS Version: Mac Basically when i have a statusbar item (extension that contributes statusbar items), VS Code re-creates all of the statusbar items for all cells. This causes ALL statubar items in all cells to flicker (as its destroyed and re-created) & we also end up with duplicates. /cc Flickering UI Duplicate statusbar items See below (from latest VS Code Insiders, can also replicate in latest main branch) Note how we end up with two language pickers in statusbar. ! After the fix (re-use instead of re-creating & only create what we need): !\nThought I had squashed all of these. Thanks for the PR!\nThe power tools don't seem to provide line numbers anymore so I tried switching the language mode around to repro and couldn't so I'm not sure if that's good enough for verification\nI think it's sufficient to mess around with it and not see any dupes. And there were random reports of dupes earlier and I haven't heard that in awhile."}
{"_id":"q-en-vscode-3c9f4b9c19bd0d1ed688ec35a4d5f6ef0e78fac9c6b301e3b87a768a0f268d33","text":"Version: 1.58.0-insider Commit: Date: 2021-06-16T05:18:57.982Z Electron: 12.0.11 Chrome: 89.0.4389.128 14.16.0 V8: 8.9.255.25-electron.0 OS: Linux x64 5.11.0-18-generic After starting up VS Code I got a modal dialog telling me about an unusual line delimiter. It didn't tell me about the file this is about. ! After accepting to fix, I saw where the problem was ! I find the dialog too heavy, especially on startup. I would need to see the code to find out what's the issue here. E.g. in our case, this is in a test, so might be intended Maybe have a warning triage next to 'LF' in the status bar? When I click ignore for this file, how can I unignore the file?\nI agree this is pretty weird. The prompt comes from a code editor contribution which executes only when a code editor widget is instantiated and shows the problematic file. I think that this normally works fine, but in this case I believe the problematic file was opened, the dialog was shown, and then the \"Release Notes\" editor opened under the prompt and changed the active editor from under the prompt. I agree with the feedback here, but I'm not sure what to do. Maybe we could make the \"Release Notes\" editor wait until all confirm dialogs have been closed before showing itself or that the workbench starts directly with that editor after a release? Regarding converting this to a passive indicator in the status bar: I don't think that would be wise. Having this unusual line terminator will lead to semantic tokens, jump to definition and all kind of language features to become off. Here is the background:\nMaybe instead of a modal dialog this could be shown somewhere in the editor area? Almost like the trust banner but instead of showing to the top of the window, showing right above the editor but still inside the editor? ! As for release notes, I think we are investigating to move away from using an editor and using something else but I cannot find the issue now. Arguably this can happen in other cases too I guess, e.g. when I open a file which loads slow and I manage to open another editor over it?\nWell, easiest fix is to add the file name into the dialog, I would argue then it doesn't really matter which editor is active.\nPR is up to change the dialog just a bit (): !"}
{"_id":"q-en-vscode-b00afb299877fd218ce9e2286bafdb35d9e70ccb66e562ef2bd54f703ba96320","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --The comments API doesn't have the concept of resolved/unresolved threads, although most code review tools distinguish them. The current API allows us to surface this to the users, see However, this seems pretty limited. I find it difficult to identify all my unresolved comments. It's very easy to miss some of them. If the API had a boolean for unresolved comments, we could surface them in multiple ways: In the Comments panel, surface the number of unresolved comments, similar to what is done with Problems: ! In the Comments panel, highlight the unresolved comments (or put them in bold), so that it's obvious to the user what comments they still have to go through. In the Comments panel, have a way to hide/filter the resolved comments. In the Editor, consider consider using a different background color for unresolved comments. I currently use a different label for unresolved comments, but this is very subtle. In the gutter, use a different symbol for unresolved comments (e.g. or a diamond of another color). In the Editor, provide a way to go to the next unresolved comment (an alternative to ).\nLet's discuss the API change. IMO, the information should be part of a . On many levels, Comments seem similar to s. A diagnostic is a comment provided by a tool such as a compiler. A comment is typically provided by a human. Diagnostics have a severity field, which indicates how important is the problem. We need an equivalent for Comments, something that tells the user if an action is expected (e.g. they need to resolve it) or not. Once the user addresses the comment, the importance is reduced. simplistic solution is to add a boolean, e.g. . to , we can go for a two-state enum instead: to be more general: . We might consider having more fields here. In a code review, multiple users interact with the comments. The current VSCode user is not necessarily the person who needs to do something. For example, the user has just updated the comment and they wait for a response. So the third value could be used to mean \"it's not my turn\" or \"acknowledged, I'll come back later\". The benefit is that the user can use the list of \"unresolved\" issues as a TODO-list and publish the comments once it's empty.\nThank you very much for your proposal. and me have quickly discussed this and we believe this proposal makes good sense. We plan to start looking into it next milestone. Though no promises yet when this will land, we will know more in December when we start our investigation.\nAll of the proposals in have merit. In a perfect world I would choose one of options 2 or 4; however, we have users of the comments API who's comments don't have a resolved/unresolved state. Option 3 circumvents that, but then we have the issue of what to show in the UI: a GitHub PR comment isn't high or low priority. Another proposal, similar to how comment reactions work:\nI see than left comments on the commit (). I like the approach of having an object for the state (instead of the proposed boolean/enum), it gives more flexibility and lets user create their own states. In my proposal, the information was attached to a thread, not to a single comment. Do we have use-cases where we want this information per comment? We might also want to control the gutter icon:\nGitHub comments are resolved per-thread, so no, no cases were we want it per comment. I'll move it to the thread.\nAfter much discussion, we have decided to proceed with the following: The main reasoning behind this is that resolved vs. unresolved comments should look the same to the user across a GitHub PR extension, an Azure Dev Ops PR extensions, and any other PR extension. There original concern around using \"resolved\" and \"unresolved\" is that maybe not all comment provider have those states (maybe they have \"viewed\" and \"unviewed\" states for example). To alleviate this, we'll see if we can avoid using \"resolved\"/\"unresolved\" language in the UI. If additional states are needed later, we can always add more to the API.\nI have a basic implementation of this working, with and themable colors. Before I finish up the pull request to send it out, I was wondering what the best practice is for providing defaults for these colors? In the mocks from we use a yellow for unresolved comments that is inspired by visuals of our internal tool. I am not sure if they match an established VS Code palette.\nawesome, looking forward to reviewing it! Generally, we try to avoid introducing a new color entirely for something like this. The best practice would be to look through the existing color values where we use yellow then use the same yellow, just like is done here where uses the same color as : We'll also do a UX review of this, so the color may end up changing anyway.\nI went for new colors as the comments are conceptually something different than e.g. editor warnings. Currently the comments in the editor use (default blue), so a theme, an extension, or the users themselves would not be able to style comments independently of peek views. I think it makes sense to base the default on another color like as in your example, so that themes which omit the new colors receive a reasonable default, but it seems like we would give up a lot of flexibility and also clarity if comment widget colors directly used something like internally?\nYes, basing the color on another is what I meant! I agree that we don't want to directly use another theme color.\nIf anyone is eager for this API to be finalized then please comment and let me know how it's working for you. Otherwise, I will plan to finalize later in 2022.\nThis would be very nice to have. We are using Github and especially larger PRs are hard to manage, if there are many comments (in Github as well as in VSCode)\nSo is there currently a way to filter resolved/unresolved comments in the sidebar? If not what can I modify currently to provide a visual distinction between resolved/unresolved?\nthere's a color difference on the icon, but that's currently it. This is the feature request you're looking for"}
{"_id":"q-en-vscode-372c921c8a1c4c34e84800215c2b9edada7e37d254f890c153e876b402cf3de2","text":"According to the documentation, if calling show(true), the terminal will be made visible but it won't capture the keyboard focus. I also remember in the past this worked. For some months already I noticed that regardless of how I call show, the focus is always taken. Anything changed? Is there already an issue that you tracking for this? Am I doing anything wrong? Related to this, if we won't manage to fix this behavior, how can we make key bindings for commands to be recognized even when the terminal has the focus? For example, when having focus in the terminal, if I press F5 debugging will start. The terminal will not try to write something for \"F5\". But if I run another key combination that I defined for some extension command, the terminal will react to the key combination (will try to write something) and the command will not be triggered (unless focus is somewhere else). We have this issue in the CMake Tools repository and the user would be happy if we fix this problem either way.\nHey, I would like to work on this issue. Could you give me any pointers? Where should I start?\nthis one's a little involved as it needs an extension to trigger the issue, you'd be better off finding an issue with the help wanted label.\nOkay, thanks!\nTried this in an extension after making sure I had a terminal open: And the focus moved into the terminal.\ndid you have focus in the terminal before running that? preserve focus returns the focus to whatever was focused before. I'm unable to reproduce the problem using the exact code you mention above. Could you please include a gif? !\n!"}
{"_id":"q-en-vscode-7c304ccf8b4d6718dd269aa75efaa6678785255d4f24578ef7ef4bb2873e5e91","text":"In some cases, when the extension host is unresponsive, a window reload can lead to the new extension host process not being able to acquire the lock for workspace storage, since it is still held on by the old (unresponsive) extension host process (). There are even greater concerns that the current way to resolve lock contention (i.e. generating a fresh workspace storage folder e.g. ) is problematic if extensions store very important data in these folders. e.g. suppose a lock cannot be acquired, the workspace storage folder is changed to , the extension writes very important data to , VS Code is closed and reopened the next day, when the extension will end up using as the workspace storage folder, missing out on the data written in (). As a low-risk band-aid, we can limit the acquiring of locks to only the remote extension host, where we have initially observed problems with multiple extension hosts sharing a single workspace storage folder, bringing the local extension host behavior back to what it was in 1.57.x and in the past.\nTo verify: open a remote window. look at the local extension host log and check that it says look at the remote extension host log and check that it says\nThe remote extension host log does not say , but instead . I guess that counts too. The local extension host reads ."}
{"_id":"q-en-vscode-d211de706cf7a8ef5185cd6e2300d13ed03ad47a83c5a7498c328ee98f168761","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: Code-OSS OS Version: Windows 10 Steps to Reproduce: IW from codelens IW is initially not pinned, and becomes pinned after the NotebookDocument is modified (e.g. sys info markdown cell is inserted)"}
{"_id":"q-en-vscode-15f1f091dd307b09c7941055e9f5af50df50cc953b4d0f80c5c011de60976025","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.58+ OS Version: Windows Steps to Reproduce: with and . the browser launches, the server throws. This is due to: It should just be . I have a PR to fix this, but needed an issue to reference."}
{"_id":"q-en-vscode-9953d7bb306d4192a9468e8ea1ac905232b0f3e7f1ded0b27f7ed291a6a752e0","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.58.2 OS Version: Linux x64 5.4.0-80-generic Steps to Reproduce: open a vscode window open the embedded vscode terminal any file path to a path that exists, but prepended with the uri 'file:///' prefix (example: file:////home/somedir) ctrl+click on that path error message appears: \"Unable to open 'somedir': File is a directory The expected behavior (opening the directory in vscode) works for regular paths (/home/somedir)\n+1\nduplicate of\nCan we get a little explanation of how this is a duplicate? Seems completely unrelated. This is incorrect handling of RFC 8089, RFC 3986 links while the other is failure to handle softlinks while using remote ssh. The issue here is file:// schema is treated as if it can only be a file while on the posix path it detects file vs directory.\nCan reproduce the problem, it should say folder not file: ! !\nRepros on Windows too. Same thing happens in the editor: ! Above should use 3 slashes:\nSo the link does work (tried it with multiple folders)... but I guess I expected the underline to include the bit. It's a nit and I don't feel too strongly about it, but thought I'd mention. !\nlet's wait until is back for that. I'm considering this verified for now\nProblem is back, looks a little different though: !\nI think the new problem is Windows only"}
{"_id":"q-en-vscode-c5b44040474a3f13e85aec9e4fe120b78cf45f781b20d5f1d8b19192013f1bb1","text":"Testing fyi Single click on the instruction breakpoint properly reveals that instruction breakpoint - great. Double click should pass focus to the disassembly view. Just like we do for regular breakpoints\nWill fix this."}
{"_id":"q-en-vscode-5a8c907bf3fb964b95a48a4633de6f98df827949ca23279e7654975f314c9590","text":"We only respect the setting when a is found: We should always support the separator.\n//cc\nWent ahead with a fix myself because I figured it is more involved: without formatting we cannot just add the separator so I ended up using the method that matches the config in that case on the relative path.\nThanks for the proper fix! I had couple of attempts but first failed when list of files in with stopped to work (apparently needed ) Second was but it produced different results for and . Haven't thought about using same normalization as in\nwhat are some verification steps for this issue?\nthanks for bringing this up. Steps from : different values for \"Copy relative path\" for some nested filed - separators should match selection (or system separator for ) I've checked latest version in and looks like fix doesn't work at this moment: apparently won't replace with ! Maybe it make sense to use ?\nis a valid character in filenames on POSIX so this is a bit tricky... I don't think we should ever convert it to on a POSIX system as that would represent a totally different resource.\nYep, in name is tricky but is it common case? Anyway: Alternative option will be to \"force\" path formatting which will work with and I'm assume it will always have as separator that later converted to if needed without causing problem with -that might replace from filenames This will be \"similar\" to missing formatting as would work with it as with path anyway. But then there will be inconsistency with : - relative as expected, expected separator- relative as expected, expected separator- - not expected but return full path (same as current behavior in ) as it's processed with , separator not supported at all. I'm not sure why returns full path but I guess it's incorrect setup for in web playground, as works OK for files in VS Code desktop.\nReopening with verification steps web on windows .copyRelativePathSeparator: a path from the memfs system that it still uses I can take this as I own label service from Going to push out as products are not affected.\nIt looks like the code that is typically doing this conversion when a formatting is there is this one: Maybe we can do something similar instead of and replace any of or with the separator."}
{"_id":"q-en-vscode-b34373be1648503ee64d033f74d6e142c212f9b722029bed8c8a7cbaf40ba2bc","text":"! //cc\nWorks great in most cases! Maybe this button needs manual tweaking though to align with the one we have in search view: ! !\nRemoved the border radius for the expanding chevron: <img width=\"139\" alt=\"CleanShot 2021-09-13 at 15 40 09 src=\"https://user-\""}
{"_id":"q-en-vscode-89fa0dc86ddf3d08be29750af027104b692ce66c9591f28fd5704b1c2ed9d153","text":"Issue Type: <bBug</bthe getting started page split editor split editor cursor on left hand side Bug A scroll bar is shown even though there is plenty of space. Scrolling doesn't seem to do anything <img width=\"679\" alt=\"Screen Shot 2021-08-23 at 11 19 37 PM\" src=\"https://user-\"VS Code version: Code - Insiders 1.60.0-insider (, 2021-08-23T10:42:53.526Z) OS version: Darwin x64 20.6.0 Restricted Mode: No <details<summarySystem Info</summary </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nI would like to contribute to this. Can I help you with the issue?"}
{"_id":"q-en-vscode-f40436ac8a948d027b0480a9929d9003241f561262ce79334bdafd91eed38f34","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --VSCode have add a new keybinding for , so, remenber to add ctrl+i for toggleSuggestionDetails which is usiing the same let add this !"}
{"_id":"q-en-vscode-1a8e0ee6c88486e58802cc1ffeb8ac582094f9b70676ddff011add1ab671237c","text":"Issue Type: <bFeature Request</bWhen foldind HTML tags inside JSX code, the closing tag is hidden and sometimes this make the understanding of the code a little bit confusing. Also, copying the folded code becomes a bit difficult. I was wondering if its possilble to make it work like in HTML, where when you fold a tag it is possible to see both opening and closing tags. Imagine that we have this JSX code: When we fold the tag, it is displayed like this: In HTML the code will be like this, showing the closing tag: so its easier to copy the code and understand the structure VS Code version: Code 1.59.1 (, 2021-08-19T11:54:37.513Z) OS version: Linux x64 5.11.0-31-generic Restricted Mode: No <!-- generated by issue reporter --\nRelevant code:\nHi, would love to take this on as my first contribution!"}
{"_id":"q-en-vscode-fc46fe919a87f6cc2de9a67c5b9e7482b0e14c1af5af262dc19377185788e455","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.60.0 (user setup) OS Version: Windows 10 Steps to Reproduce: through VS Code Search tab and tab through all the buttons . ( refresh , clear search results , etc. ) ( Search tab is just an example I can see the same behavior in all buttons) Expected : On the tab navigation to each of the button, the tooltip should be displayed. On escape the tooltip should be dismissed. Keyboard users will not know what is the use of the button through the icon for the button. Actual : User cannot see the tooltip when they tab into a button. They can only see the tooltip when they hover over. !\n, Hey, I search to see if there is something already open but couldn't find it. I hope it is not a duplicate.\nDup of Reasons why we are not changing this still hold which I mentioned Leaving this open for tracking purposes.\nGitHubTags:#A11yMAS;#A11yTCS;#Win32;#DesktopApp;#A11y_VSCode;#WCAG1.3.3;#A11ySev3;#VisualStudioCode-Win32-Apr2021;#Benchmark;#Visual Studio Code Client;#Keyboard;#CEACCReviewed;\nRemoving label because we got confirmation from the testing team that this can be moved to Severity 3. Reasoning why this is Severity 3 can be found\nHi , we met the same issue. In our Teams Toolkit extension, we have several buttons on tree items. When user tab to the buttons with keyboard, tooltip will not show: ! Accessibility team has raised a Sev2 bug to us for this issue. Is it possible to fix this?\nsame issue seen in bug\nIt can be done in one line due to a contribution by the Edge team to Chromium However, loads of tooltips don't show up, eg this toolbar (top right)- ! Maybe because is on an element which inherits the attribute from a parent element. This can be reproduced on as well.\nthanks for investigating. When I make that change, I don't see it working for any of the buttons. Which work for you?\nHi, iirc it worked on New Search Editor, Search (text box), Replace (text input) and the icons in the activity bar. By the way, I noticed then that is missing in the activity bar.\nwhen I test with the above change, it doesn't work for me for any of the buttons. I looked at the docs and it should be supported. Any clue why that would be?\nyou need to close and reopen after making the change. Reload Window wouldn't be enough. You can also visit in Edge and it should have the same behaviour.\nThanks, yeah, I run OSS to test, which should work, but doesn't.\nIt looks like we'd need to move all native hovers over to using custom ones to do this.\nI sent a draft PR and explained there in more detail what works and doesn't work. (Tested on Windows 10)\nI am reverting this for now because we don't want multiple tooltips showing up on hover. I'll try to investigate in December.\nRemoving these lines, IE the custom hover, fixes the issue, but I worry about making this change so close to the release: it looks to me like we are using custom hovers all over the place because we didn't have this flag enabled before. Do you think we can consider removing those in favor of the native ones?\nthe custom hovers are better for accessibility as they're exposed to the screen reader. The eventual plan is to move to custom hovers throughout the workbench. Generally if the custom hover is being used a should not be present.\nPartially blocked by\nisn't this separate to adopting it in different places? I would have expected us to need a focus handler to show the hover for that specific case\nThis can be closed when happens\nBen has adopted the custom hover in our base widgets, so this may now be easier than before?\nKai mentioned I should consider adding a delay like we have for mouse hovers here. I need to see if this is acceptable by accessibility standards.\nAFAICT, the WAI-ARIA guidelines do not specify that there cannot be a delay.\nit seems like this already works thanks to your changes\nWhen I tested on native Windows explorer, I saw a delay even when using keyboard, so it seems the same delay should apply for keyboard and mouse.\nCurrently the mouse and keyboard delays are exactly the same which is what we want. We do not have delays where we support instant hover such as navigating between action items on a toolbar/actionbar. FYI: pressing closes the hover\nawesome work!"}
{"_id":"q-en-vscode-7874fb157ca85f40a7c69380459fb3ad76de5cd2bbd54c826b73df634e861576","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: No (Needs InlineSuggestions from some extension) <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.60.0 OS Version: MacOS Catalina 10.15.7 Steps to Reproduce: GitHub Copilot. The exact version shouldn't be important, but this was verified using 1.4.2586. a blank Python document and enter the following text: There are a single space after each sign. Put the cursor on the end of the first line and trigger Copilot inline-suggestion by pressing Backspace then Space. Don't accept it, but press Down-arrow, Backspace, Space. After pressing Down-arrow the suggestion on the first line disappears (as it should), but after the Backspace+space, you may see the suggestion briefly appear again, before the new inline-suggestion appearing on the second line. This is a bit flaky, and it may not appear every time the cursor is moved. It may also depend on latency to the Copilot server, and so be less pronounced in e.g. North America than in Europe or Asia. Edit by This is the buggy behavior: !\nPlease let me know soon if we should release this with our Recovery Release tomorrow!\nThanks for your lightning fast fix, Releasing this quickly would be awesome - it's a quite distracting glitch!\ncan you verify that the issue is fixed in todays insider build?\nAs he's out today, I've had a play with today's insiders build and can't reproduce it, whereas I can with the insiders build from 6th September. So looks good to me!"}
{"_id":"q-en-vscode-870921397260b1476160bfe686734efdcc1c1c23c42c255d6112dd0122eeeaea","text":"<!-- Please search existing issues to avoid creating duplicates. --- VS Code version: 1.59.1 Extension version (available under the Extensions sidebar): v2021.8. OS and version: Linux x64 5.13.11-1-MANJARO Python version (& distribution if applicable, e.g. Anaconda): 3.9 Type of virtual environment used (N/A ...): virtualenvwrapper Relevant/affected Python-related VS Code extensions and their versions: pylance v2021.8.3, jupyter v2021.8., Visual Studio IntelliCode v1.2.14 Value of the setting: Pylance from the VS Code marketplace to see if that fixes your issue] Extension loads successfuly Extension does not load, console reports \"Unexpected token ' in JSON at position \" [NOTE: Self-contained, minimal reproducing code samples are extremely helpful and will expedite addressing your issue] the extension by clicking a .py file <!-- Note: If you think a GIF of what is happening would be helpful, consider tools like , or . --<details<summaryOutput for <codePython</codein the <codeOutput</codepanel (<codeView</code→<codeOutput</code, change the drop-down the upper-right of the <codeOutput</codepanel to <codePython</code) </summary<p</p</details\nLooks like you might have a bad update. Do the following to remove stale versions of the extension: Python extension (if you have pylance uninstall it first). all instances of VS Code. to, (on windows) or on Linux/Mac. any folder with the name starting with VS Code, and install Python extension (also pylance if you uninstalled it in step 1).\nI followed all steps, but nothing changed. should I try to install older version of python ext?\nHi Can you try installing an older version of the Python extension, and if it doesn't work can you try a different version of VS Code (either an old one or the Insiders build)?\nI've installed python extension , jupyter and no pylance. here's the log: As you said I'm going to install older version of vscode\nHi Did you try installing an older version of VS Code?\nFYI - This looks like some JSON stored via the became corrupt. The JSON is written a few lines lower, so it is unexpected that the value is corrupted. This might indicate some kind of bug in the storage layer. I don't remember writing that code, did you perhaps write that? Do you think it is so important such to fail extension activation when the JSON is corrupt, or should we catch that exception and return ?\nI would suggest to be more resilient in extension activation as you suggest and fallback to an empty object here if any error rises from but log it as error: This is the first time I hear about a problem with persisting data into SQLite. However, seeing makes me think that this is not an officially supported Linux distribution.\ncurious if you could share the following 2 files with me (via e-mail is fine: find it on ). Steps: open the VSCode window where this issue happens View Command Palette =Log Storage Database Contents Help Toggle Dev Tools Find the 2 paths for the DBs send both files to me\nI can still see the original issue in vscode-insiders v1.61.0-insider\nwhats the exact version you are on? And is it the issue that the extension fails to activate?\nVersion: 1.61.0-insider (user setup) Commit: Date: 2021-09-16T08:31:28.161Z Electron: 13.4.0 Chrome: 91.0.4472.164 14.16.0 V8: 9.1.269.39-electron.0 OS: WindowsNT x64 10.0. python extension: v2021.9. Error 2021-09-16 17:51:12: Error discovering pytest tests: [SyntaxError: Unexpected token S in JSON at position at (<anonymous) at a.runTestDiscovery (PATH.vscode-insidersextensionsms-outclient) at runMicrotasks (<anonymous) at processTicksAndRejections () at async (index 0) at async _.refreshTestData (PATH.vscode-insidersextensionsms-outclient) at async v.refreshTestDataInternal (PATH.vscode-insidersextensionsms-outclient)] { stdout: '[{\"rootid\": \".\", \"root\": \"...\nI fixed the issue as reported in that extensions fail to activate when the contents in storage fail to get parsed. What you report here seems different to me. Looks like the python extension is calling on something that is not valid JSON? Have you asked the Python extension what this could be?\nI see, I'll create a new issue. thanks"}
{"_id":"q-en-vscode-8192e3345186279377469531cd16c8a81ccb73a0a344a1e6540ed39061f95f99","text":"Currently when using , ArrayBuffers just turn into an empty object. We should support these for efficient data transfer between extensions/extension hosts. We recently support for them on a protocol level, so support here should not be terribly hard...\nI see two possible approaches to support this: this logic specifically for commands (or even just for a specific command). This is what we do for webviews for example: our implementation aware of array buffers instead of only supporting VSBuffers. Take a look at how is used in our rpc code as a starting point: The webview serialization logic I linked to above would probably be a good starting point for this"}
{"_id":"q-en-vscode-1a90f0c04eb67e8046cbc8216ec2088d876d50374d65dc6b2516429026cbac89","text":"Issue Type: <bBug</ba JS file: Fold and place cursor after the opening of the class Bug Nothing happens VS Code version: Code - Insiders 1.61.0-insider (Universal) (, 2021-09-14T05:14:54.400Z) OS version: Darwin x64 20.6.0 Restricted Mode: No <details<summarySystem Info</summary </details<details<summaryExtensions (55)</summaryExtensionVersion ------ salesforce-soql-editor1.8.0 comment-tagged-templates0.3.1 docs-view0.0.9 emojisense0.9.0 folder-source-actions0.2.1 jsdoc-markdown-highlighting0.0.1 markdown-checkbox0.3.2 markdown-emoji0.2.1 markdown-yaml-preamble0.1.0 test-ext0.0.1 vscode-eslint2.1.25 typescript-notebook2.0.3 gitlens11.6.0 tsl-problem-matcher0.4.0 EditorConfig0.16.4 codespaces1.0.5 vscode-pull-request-github-insiders2021.9. vscode-drawio1.6.2 vscode-postfix-ts1.9.4 latex-workshop8.20.2 vscode-platform-specific-sample1.15.0 vscode-platform-specific-sample1.5.0 vscode-memfs0.0.3 template-string-converter0.5.3 vscode-azurefunctions1.5.1 vscode-azureresourcegroups0.4.0 vscode-docker1.16.1 vscode-language-pack-de1.60.3 vscode-edge-devtools1.3.0 python2021.10.-dev vscode-pylance2021.9.2-pre.1 jupyter2021.9. jupyter-keymap1.0.0 remote-containers0.194.0 remote-ssh0.65.7 remote-ssh-edit0.65.7 azure-account0.9.8 hexeditor1.8.2 js-debug-nightly2021.9.417 live-server0.2.8 vscode-github-issue-notebooks0.0.106 vscode-typescript-next4.5. vsliveshare1.0.4836 debugger-for-chrome4.13.0 vetur0.34.1 java0.81.0 code-spell-checker2.0.4 rest-book4.2.0 ayu1.0.5 luna-paint0.9.0 sort-lines1.9.0 vscodeintellicode1.2.14 vscode-java-debug0.35.0 vscode-java-dependency0.18.7 vscode-java-test0.31.3 (2 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nI'm trying to document this command but find its behavior difficult to understand and rather unpredictable This could be wrong, but I think it tries looking for the next fold above the current level. In the case above, that means that works if you are inside another method: But it doesn't work if you are in the class body I'd expect this command to work more like where it just jumps to the next folded location in the file\nFYI Got to next fold looks for the next fold of the same level. In the example above there's no such fold\nThanks for tagging me, The way I have fold jumping working is that it takes the cursor position and figures out what fold you are currently in. From there, it checks if you want the next previous or parent fold. Parent is pretty easy. It just gets the fold that's at level X - 1. For next and previous fold, they are implemented as the next or previous fold at the same level. The reason for that is because it makes it easy to find associated parts of the file. For example, if you are at an , , or fold, you can easily navigate to the other conditions. Similarly, in the above example, if you are on function , you can easily jump to the other functions that are defined. I don't think it's valuable to jump to the next or previous fold by file order. For example, in the below code it would not make sense to me to have jump to next fold go to b1, then c1, then c2, then a2. Please let me know if you want me to make any changes though.\nWe need a better label then: ?\nThat seems reasonable. Want me to do that? I probably shouldn't change the actually command ID though, right? That would break key bindings.\nThanks for the explanation. I agree that this should be renamed. Even with a better name though, I still find the behavior from my example confusing. When the cursor is inside the class body, I'd expect that the folding points for and would be siblings to the cursor's location (because they all have the same folding parent). Instead though we jump out of the class entirely\nOne other note on the name: maybe the command should call these something like instead of ? Conceptually I think of a as being code you've actually folded, but my understanding is that these commands actually jump between foldable lines in the file\nsounds good to me. I see how that's confusing, so I put together this JS sample to illustrate where the \"next fold\" is for different cursor positions: Cursor 2's next is straightforward since you're on a folding point. Cursor 3's next I think is reasonable. You are in a fold, so you find the folding point and go to the next of that. Cursor 1 and 5's next start getting a little weird, but following the same logic as cursor 3, it seems logical that is the folding point, so the next sibling is at . Cursor 4 is basically your original question. I agree that it gets a little confusing. The logic is the same as 1, 3, and 5, but visually, if you're between 2 folding points, I could understand the desire for it to go to those points. I think the confusion stems from using the command when you're not on a folding point line. If you are, I think it's easy to understand, but when you aren't, not so much. What do you think of the following? If you are on a folding point line, do what it does now. not, go to the next folding point sequentially. With those updates, the above example looks like this:\nWe don't use 'folding points' as a term. I suggest 'folding range'\nOk. What do you think of changing the implementation as I mentioned in my previous comment?\nYes I like the suggestion (I thumbed up it :-) )\nI wish I could turn on reaction notifications in GitHub. I implemented this in . I update the command from \"Fold\" to \"Folding Region\", but I did not add \"Sibling\". I hit a bit of a dilemma there. With the change, it's no longer necessarily a sibling. However, I can still see people getting confused by previous or next folding region command \"not working\" when you are on the first or last sibling folding region's start line. For example, in the above example, the previous command won't move you when you are at . I think that's really valuable. It lets you move up and down the different cases of branches without going past them. I wouldn't want to run previous at and end up at , then next brings me to . Any suggestions for that? I'm pretty sure nobody would want the command label to be \"Go to Previous Folding Range, but Stop at First Sibling\", but that's essentially what this does.\n/verified Looks to follow the logic set out by I used this example: When on line 8 it jumps nowhere is a little confusing, but that's because the next folding range is the current?\nThat's because it's the last sibling. The benefit is that you can move to previous and next sibling folding ranges without stepping out of them (at which point going to the next or previous sibling won't be the original set of siblings you were on). I agree, it's confusing, but I think it's valuable to do it this way."}
{"_id":"q-en-vscode-7b2163386cd2405b1839e1d018c09f1c9490bca46a04cfda7b8bc2c351e7c6f5","text":"This new capability allows the terminal to control when the tab description shows, so we can align the description in the terminal editor with how it shows up in a terminal panel tab.\nTo verify: Run a task, move the terminal to the editor area, check that the description shows up like it did in the panel In a terminal, change the directory, verify you see it in the tab description in the panel Move that to the terminal editor, make sure you see it there as well\nIs this tooltip the task description? <img width=\"486\" alt=\"image\" src=\"\"I think I'm a bit confused on how a task would be affected by a different terminal changing directories.\ndrag the terminal into the editor area and you should still see in the label of the tab."}
{"_id":"q-en-vscode-b586eff4c4ffdef758fe6c5c822311b587ca798302ae78a50c8c40e0ec4be006","text":"I just reloaded a window with my repo open, and it attached to a terminal which is already open in my repo, so now this terminal is open in two windows: ! The right window is the window that originally owned this terminal One thing about this terminal is that its cwd is some other folder outside of both workspaces, if that matters.\nI reloaded the left window again, and it reattached to the correct terminal. !\nJust so you know, this happens pretty often. I just reloaded a window and got some random terminal from another window. Let me know if I can collect any other info\nAlso hit this, I believe it was after restarting insiders for an update. !\nI was able to repro once out of many attempts. I wonder if there's a command to reload all windows to simulate what happens on update? Maybe then I'd see it?\nexiting the application (cmd+q) and reopening it is essentially what happens on update\nSomething similar just happened with the most recent insiders update, but now the terminals aren't synced, it's just that the working directory of one window's terminal is set to the other window's root instead of it's own. <img width=\"1440\" alt=\"image\" src=\"https://user-\"\nHappened again, accidentally ran a destructive command in the wrong directory :( cc is there any logging or something I can enable to help you get to the bottom of this?\nI spent awhile investigating. I don't understand what could be going on here because we store and reconnect to terminals according to the workspaceId. said that he thinks it's a race condition. Setting log level to trace and toggling the shared process would help when it reproes\nToday I had something new happen: had 3 VS Code windows open. reloaded window 1. 1 now shares a terminal with window 3! Totally different workspaces. No other windows were reloaded. No windows were closed. Is there a way I can disable terminal reconnect for local windows? I've also run destructive commands when this happens and I don't want to risk it.\nYep, it's not working atm for remote terminals, so:\nHello everyone, there is any plan to try to fix it in the nearest milestone?\nThis code that we for transferring terminals via dnd between windows looks suspicious\nHas anyone seen this in awhile? I cannot reproduce\nI don't remember seeing it recently\nOkay pls reopen if you see it again and I'll be happy to debug with anyone that can reproduce it\nSorry, I cursed myself because this happened again today but I probably won't have consistent repro steps. I reloaded the window with my vscode workspace and got a terminal from this extension sample which is also open in another window, although the terminal panel had not been made visible in that other window's session yet. I reloaded the window several more times and always got the right terminal after that. <img width=\"1049\" alt=\"image\" src=\"https://user-\"\nwe don't create a terminal until you open the panel, so what do you mean by the terminal had not been made visible yet?\nIn case it is important to the issue, I'm just saying that the drop-on-document terminal that was wrongly attached to the vscode window hadn't yet been attached to the drop-on-document window, because it was from some earlier session.\nI also used to see this periodically but I've had terminal reattach disabled to avoid it. Let me know if it will be helpful and I'll re-enable reattach!\nSame as\nYes it would be very helpful I spent 90 mins yesterday trying to reproduce to no avail. If you could enable trace logging too in case you repro, that might give the insight we need\nWe should review the code around terminal adoption of orphan processes, I suspect the issue is that it's not transactional.\nit appears we're only doing an orphan check on terminals that are in the workspace as set in the map\nI think somewhere all unadopted processes get taken by some window? Might not be remembering right though.\nIf you see this again after lands, please enable trace logging and get the output channel logs for Pty Host and Window.\nI was seeing this previously on occasion but haven't seen it over the last week so marking as verified\n+1 on no longer seeing it since re-enabling re-attach a few weeks ago.\nJust saw this again. The incorrect re-attach occurred on 1 July and I only noticed it today. Pty logs: Window logs\nThis seemed for a while like it'd stopped happening, but just ran into it again: In my case I'd just reloaded a single window.\nJust had this happen again, but once again I didn't have trace logging on. I'll try to leave it turned on for next time. Pty Host logs: Window logs:\nCould the needed log statement to troubleshoot this on the terminal side be converted to an log. That might help track down this bug sooner.\nthey're actually warnings already so they should show up:\nJust ran into this again. I have the same terminal restored in two distinct windows. I ran into this after restarting to update Insiders on macOS (I had 6 folders opened). The Pty Host log contains the following: https://user- Maybe might know if there's a way to easily simulate a restart caused by an update. Alternatively, maybe you can try changing a setting like which shows the restart prompt. I think this also leads to the terminals attaching incorrectly.\nMight be related to (maybe not as that's remote only?)\nthis is a little confusing to me. No reconnect is meant to happen on restart, when updating node-pty gets replaced so it can't be running, all processes are torn down and the pty host shuts down with the main process. Did you see this? !\nChanging this to always return isOrphan = true: I can get it to reconnect which is meant to be prevented by one of my other changes.\nIt's a little suspicious that the entry doesn't get deleted: That determines here which can bypass the orphan check:\nI can confirm that within the last two days (both before and after the upgrade to 1.70.0) it has happened to me both on window reload and on full VSCode restart.\nWhen I see this it also happen on restart.\ndo you see the history restored message when it happens on restart?\nYes, I see this every time when restarting to install a new version (Gear Restart to Update) or when changing the setting : <img width=\"473\" alt=\"image\" src=\"https://user-\"\nI'm 99.9% certain the answer is yes, but I will make sure to notice the next time it happens.\nOk, I figured this out. What a journey! Here is the fairly reliable repro: Clear state - Given 2 folders and Open folders and and close all terminals File Exit Setup for the process revive - Window : Open 3 terminals and type \"a\", \"b\", \"c\" into them respectively, this helps identify them to ensure they're in the correct order later Window : Open 3 terminals and type \"d\", \"e\", \"f\" into them respectively File Exit Process revive - Window and should revive the processes and have all terminals in the correct order they were created (this works) Process reconnect Reload window and window at the same time Window and should revive the processes and have all terminals in the correct order they were created :bug: Note that this is a timing issue and has a chance to not reproduce the problem. This is what is meant to happen when exiting and restarting the pty host process. Terminal processes are revived (aka. re-created) and their actual pty IDs for this pty host process are mapped to, starting at 1. Notice below how the new IDs start from 1, they just represent the order they are created in which shouldn't matter when reviving. After that, when reloading windows, terminal processes do not get revived but reconnected to. So an mapping is expected: What is happening and what is causing this bug as well as a related issue[^1] is that the revived pty ID is cached, so if reloading happens to use a that is mapped to a different pty ID in , it will reconnect to that one instead: The above results in: Then on reload the pty ID mapping occurs instead of being undefined, which causes a reconnect_ to the wrong pty. So in the above they get re-mapped, shuffling the order within a window and potentially connecting to a different window's ptys: ! [^1]: The folder's terminals may not get restored at with the warning \"Couldn't get layout info, a terminal was probably disconnected\" in the pty host output channel. This started happening due to one of the previous workarounds to prevent 2 windows attaching to the same pty.\nWow. Nice detective work! Thanks, looking forward to the fix.\nReported here\nI actually saw it again once earlier this week too. It does seem to happen significantly less often than it used to.\nI just ran into this again today. After a restart due to an update, I got a terminal from a folder showing up by itself in folder . Here's the only relevant content I found in my logs:\nI hit this today. Moving this forward to January 2023, since it probably was just lost in the queries by being assigned to August 2022.\nI've hit this once per day for the past several days. It seems to be occurring more frequently again.\nThis would probably be addressed if terminals moved into utility process and there was a message port per terminal:\nUsing the new pty host trace logging, this is what happens on a reload with a single terminal that will be reconnected, with the environment parts truncated:\nI believe this is now fixed as a part of I'm still not able to reproduce but the current theory is that the async isOrphan check that was removed in the PR was not only delaying reconnection by 4 seconds, but also causing a race where the wrong window could adopt it. Instead we no longer check whether it's an orphan and trust the saved workplace state to be correct.\nThanks for following up\nI haven't seen this in a long time. Not sure whether there is anything that can be done to verify it.\nI just ran into this now: had ~4 windows open: Window 1 is vscode-pull-request-github2 : terminals cwd were all in the workspace Window 2 is vscode-pull-request-github: terminals cwd were all in the workspace Window 3 is repo1: 2 terminals, both of which had a cwd outside the workspace Window 4 is vscode: terminals cwd were all in the workspace VS Code reattach: Window 1's terminals are correct Window 2 has reattached to one of Window 3's terminals and has lost it's original terminal Window 3's terminals are correct Window 4's terminals are correct. Pty logs from Window 2:\nthanks for the info! Looking into it\nI spent a few hours looking into this, here are my findings: This warning in the logs is common across recent failures: , I tried and tried but could not reproduce this on Insiders or OSS I'm adding more verbose logs to the info log level that gives me the old/new pty id mapping as well as dumping a bunch of info when the above error happens I looked into whether should be correct and while I found it could cause issues when the pty host restarts (), it seems like it should work after a restart I could the cwd being outside the workspace would make a difference as we don't even take that into account when reviving the proc () The management of would be a lot simpler if we moved the id to being a guid instead of a number, but that would be a shame as we then need to send the guid along the wire for pretty much every request to the pty host I looked into possible race conditions where could get stale/invalid values. Both a thorough review of the code and injecting timeouts (random, differing by window) into important calls didn't cause problems Current plan is to wait for to merge and await a repro with these logs present, then analyze them. I'll also likely be continuing to fiddle with terminal reconnection in debt week to improve perf further and reduce redundancies.\nI'll keep an eye out for this and try to maintain a setup similar that one that I just saw it with, but I have to say, this has gotten so much better. I almost never see it.\nI tried to repeat my setup before updating, and it did happen again. Setup: Window 1 is vscode-pull-request-github2 : terminal cwd is outside the workspace and on a different drive (C vs D) Window 2 is vscode-pull-request-github: terminals cwd were in the workspace Window 3 is repo1: terminal cws was in the workspace Window 4 is vscode: terminals cwd were all in the workspace After update: Window 1 is vscode-pull-request-github2 : has a new terminal Window 2 is vscode-pull-request-github: correct Window 3 is repo1: has attached to Window 2's terminal and now they share the same terminal Window 4 is vscode: correct Logs:\nThanks it seems there's some inconsistency with the layout info and the revived procs, so I suspect this silently error swallowing now: Added logs for that here: , I'll merge for July so you might not be able to test for a week but when you do please also grab the Terminal output log as this is renderer side.\nCross post\nWith logs in , I was able to reproduce the problem! ! The repro is very complicated, it is essentially trying to get into a state where 2+ windows have \"old pty ids\" that conflict with one another, and then increasing the latency of the pty host in order to force the race condition to show itself. these settings: default log level to trace all windows window 1 (W1) with folder A 1: Open 2 terminals 1: Close window 1: Open window 1: Open pty host log, you should see this: window 2 (W2) with folder B 2: Open 2 terminals 1: Close window 2: Close window VS Code (W2 should open) 2: Open pty host log, you should something like this: 2: Close window an external terminal and type another external terminal and type the commands in both terminals in quick succession 1|W2: Open pty host log, you should see something like this: The terminal with end up attaching to the wrong pty. The fix for this is simple, just add workspace ID as part of the key of the old id, so becomes . This should make conflicts impossible to occur."}
{"_id":"q-en-vscode-d640502bdeccb11711a39bed6860c62002b3405fc8b8f1f4157eadcd63896b80","text":"The <kbdToggle Size to Content Width</kbdcan only be toggled a few times, before freezing completely. Resizing terminal could reset it. https://user- Does this issue occur when all extensions are disabled?: Yes System info:\nplease see my explanation on\nThank you for your detailed explanation. I did test on the insider version, the blur seems to decreased with smaller canvas required but this freezing issue still happens all the time. Moreover, the issue is only on Linux so I'm not sure the explanation is quite sufficient to describe this. Let me know if there's anything I missed, thanks again.\nwhen you do this how long as the lines you're working with? ideally, we probably need an upper cap based on the environment's max canvas size (which would need some searching).\nIt's hard to say when the doesn't work as expected this time. Do you have any idea to get this revealed? https://user-\nwhen the width of the longest line is less than the width of the terminal, as is featured in your gif, toggling the width won't do anything (by design).\nthanks for your information, I did some tests again and got some conclusions: in the menu doesn't work, so I recorded with shortcut . IIt also freezes after the second as expected (bottom scrollbar indicates that): https://user- lines hidden when press are affected (still freezes after the second toggle): https://user- number 2 above seems to happen almost every line length: ! Let me know if I can do any more tests to provide more information.\nAs compared with how this works for windows/mac, linux is broken given the same text content !\nThis actually is working by design - when you copy and paste, it includes newline line endings that will indicate to the terminal that line isn't wrapped.\nbased on , it seems like we should go with the min value since we don't want to calculate this per platform/browser, so going with 8000\nI started on Windows to try to understand how works. It didn't appear to do anything on Windows. On linux, I can see the toggle happen, but there is no scroll bar.\nit will only happen if there are wrapped lines. so increase the line width or resize the terminal to be more narrow.\nthe scrollbar is very low on linux and i'm not sure why. I'll open an issue for that - separate from this one\nWindows !\nthanks, i'll make a new issue for that\nNow the same thing is happening on linux too. I will leave this for someone else to verify.\nThank you for getting back. I've tested with the insider build but it doesn't work, I've thought it was some sort of thing related to my particular machine, which confuses me these days. So it's good to track that again on a new issue. Most of behavior I have described in I don't spend much time on Windows so didn't test carefully. Let me know if there's anything else I can do.\nBecause of the way this feature works where it stretches the canvas to the size of the content we are limited by the canvas API. As a compromise I'm going to add a cap to the number of columns the terminal can take up."}
{"_id":"q-en-vscode-2b2eebf7768ecd84d2fded79c3cc16263dc2cff5413c50ba8c039c8e6ee8bb15","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled? Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.61.0 Commit: Date: 2021-10-07T18:11:58.853Z Electron: 13.5.1 Chrome: 91.0.4472.164 14.16.0 V8: 9.1.269.39-electron.0 OS: Darwin x64 20.6.0 Steps to Reproduce: the Terminal Sample from the vscode-extension-sample repository the a task that reuses a terminal. For example, I used a user task like: On the first run of the task the 'link' text will create a hover link correctly. If you re-run the task so the terminal has 2 (or more) runs of the task in the same terminal then the link will fail to detect correctly. Adding a breakpoint to there appears to be 1 call per execution of the task in the terminal. I think when the terminal is re-used it re-registers the link provider and if two link providers have overlapping links they both seem to fail (which is a known issue but due to this bug all link providers start failing if tasks are run multiple times).\nPutting a breakpoint in the registerLinkProvider call for xterm (here ) I see multiple copies of all of the link providers: <img width=\"574\" alt=\"Screen Shot 2021-10-12 at 9 15 30 AM\" src=\"https://user-\""}
{"_id":"q-en-vscode-08750695ba99529d470aa38a34c6ab020067dc4d7672375b99cef7cd046076b5","text":"Issue Type: <bBug</bI'm using MacOS. After opening VSCode and getting into full-screen mode: If I use the integrated terminal of VSCode to create a new file, e.g.: VSCcode switches my screen to my previous workspace that causes interruption of work. Everytime I use to create a new file, I'll have to use + to move back to VSCode. VS Code version: Code 1.61.1 (, 2021-10-14T01:23:30.274Z) OS version: Darwin x64 20.6.0 Restricted Mode: No <details<summarySystem Info</summary </details<details<summaryExtensions (17)</summaryExtensionVersion ------ code-runner0.11.6 terminal0.0.10 go0.28.1 vscode-docker1.17.0 python2021.10. vscode-pylance2021.10.1 jupyter2021.9. jupyter-keymap1.0.0 jupyter-renderers1.0.3 remote-containers0.202.4 remote-ssh0.65.8 remote-ssh-edit0.65.8 remote-wsl0.58.2 vscode-remote-extensionpack0.21.0 vscode-yaml0.23.0 code-settings-sync3.4.3 open-in-browser2.0.0 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nI am not sure if this could be a regression from the work we did to use the command.\nI just built a version without the open cmd fix, but still saw the behaviour. I'm wondering if it could actually be caused by changes made during Edit: It seems reverting both commits fixes the issue.\nIt is unrelated to my changes. When I simply change this line to always go into the of : It does not reproduce anymore. Looks like when we are using , macOS does not really know whether a new window opens or an existing one reused. I feel this is a pretty bad regression actually.\nI see that has fixed this bug and merged the PR. It seems like this is in (not sure what that means) but the issue still persists. I wonder when will this be fixed? In the next VSCode update? Thanks.\nIt should be fixed in our current latest stable release ().\nThanks, Just updated to today and I confirm that it's fixed."}
{"_id":"q-en-vscode-45e7e3e7d1b48eca5c8b50f4449210f1f43157abcaef410b0fd38a713bb0ba81","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: n/a <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: latest OS Version: macOS 11.2.3 Steps to Reproduce: repo npm version < 14 or ≥ 17 to see this message message or just read the commit diff to see the problem\nfrom check is for ≥ 17 but error message says to use ≤ 17, thus exactly v17 is both valid and invalid"}
{"_id":"q-en-vscode-2edc0cb37ef6a48b076354894c97bd8edb74a7629c6824d3a46e00996d0ef335","text":"Copied from My workaround is export VSCODEGITASKPASSEXTRAARGS=\"--title=foobar\" but the proper bugfix should happen in VSCode's GIT_ASKPASS implementation.\nI experience the same problem, from the moment been released. Working through remote ssh the git processes stay open, without anything happening.\nStill facing this issue on WSL.\nWill be fixed in upcoming 1.62.2. Meanwhile a workaround is to and then manually\nI can confirm that this is fixed after updating to 1.62.2 ()."}
{"_id":"q-en-vscode-7ee8a689957c63b06c39376634bb67232c9f4ed6b6d322dd895723c543110640","text":"! It could be that I caused this, but I doubt it (injected text has nothing to do with the gutter, maybe they shared a CSS rule).\nI can reproduce but I'm not aware of any changes in that area from the debug side.\nI would like to work on this issue!\nBefore fixing anything we should do a \"git bisect\" to find the commit that broke this.\ncommit . it should\nthanks for pointing out the commit! If you like, you can provide a PR, that would be awesome!\nPR"}
{"_id":"q-en-vscode-eb91c628ed3ba450656d846a909f78ebe8edb73e5daf542f1fd40ca65830333c","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: git commit (1 Dec 2021), with openvscode-server patches until commit Browser: Firefox OS Version: Ubuntu 20.04 Steps to Reproduce: a freshly opened window, open or type some text in it. Try using Ctrl + Shift + Right, text should be selected at word boundaries. a terminal, then close it using command. using Ctrl + Shift + Right again. Expected: Text should be selected at word boundaries. Actual: No text is selected. Cursor doesn't move at all. This is the the same issue as filed at [1] and [2]. If it's reproducible in 3 server implementations (counting openvscode-server), then it's probably an upstream issue. Sadly I don't have an access to GitHub Codespaces, so I can't verify it with an \"official\" implementation [3]. You'll notice in keyboard troubleshooting log below the following 2 lines: It's been suggested in [1] that vscode might still think that is active despite the terminal is closed. Thus, a workaround I've been using is to have the following in (with default unbind): That is, wherever exists, augment it with . [3] I mean, doesn't have terminals. I've tried using Live Share; that doesn't change. See Keyboard shortcuts troubleshooting. The first keypress is before the terminal is opened. The other two are after it's closed. <details<pre[2021-12-03 01:36:00.490] [window] [info] [KeybindingService]: / Received keydown event - modifiers: [ctrl], code: ControlLeft, keyCode: 17, key: Control [2021-12-03 01:36:00.503] [window] [info] [KeybindingService]: | Converted keydown event - modifiers: [ctrl], code: ControlLeft, keyCode: 5 ('Ctrl') [2021-12-03 01:36:00.509] [window] [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. [2021-12-03 01:36:00.513] [window] [info] [KeybindingService]: / Received keydown event - modifiers: [ctrl,shift], code: ShiftLeft, keyCode: 16, key: Shift [2021-12-03 01:36:00.517] [window] [info] [KeybindingService]: | Converted keydown event - modifiers: [ctrl,shift], code: ShiftLeft, keyCode: 4 ('Shift') [2021-12-03 01:36:00.521] [window] [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. [2021-12-03 01:36:00.808] [window] [info] [KeybindingService]: / Received keydown event - modifiers: [ctrl,shift], code: ArrowLeft, keyCode: 37, key: ArrowLeft [2021-12-03 01:36:00.813] [window] [info] [KeybindingService]: | Converted keydown event - modifiers: [ctrl,shift], code: ArrowLeft, keyCode: 15 ('LeftArrow') [2021-12-03 01:36:00.817] [window] [info] [KeybindingService]: | Resolving ctrl+shift+[ArrowLeft] [2021-12-03 01:36:00.821] [window] [info] [KeybindingService]:  From 2 keybinding entries, matched cursorWordLeftSelect, when: textInputFocus, source: built-in. [2021-12-03 01:36:01.146] [window] [info] [KeybindingService]: + Ignoring single modifier shift due to it being pressed together with other keys. [2021-12-03 01:36:03.304] [window] [info] [KeybindingService]: | Resolving [KeyE] [2021-12-03 01:36:03.310] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:03.401] [window] [info] [KeybindingService]: | Resolving [KeyE] [2021-12-03 01:36:03.406] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:03.737] [window] [info] [KeybindingService]: | Resolving [KeyX] [2021-12-03 01:36:03.744] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:03.832] [window] [info] [KeybindingService]: | Resolving [KeyX] [2021-12-03 01:36:03.838] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:04.210] [window] [info] [KeybindingService]: | Resolving [KeyI] [2021-12-03 01:36:04.218] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:04.344] [window] [info] [KeybindingService]: | Resolving [KeyI] [2021-12-03 01:36:04.348] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:04.396] [window] [info] [KeybindingService]: | Resolving [KeyT] [2021-12-03 01:36:04.404] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:04.518] [window] [info] [KeybindingService]: | Resolving [KeyT] [2021-12-03 01:36:04.523] [window] [info] [KeybindingService]:  No keybinding entries. [2021-12-03 01:36:04.762] [window] [info] [KeybindingService]: | Resolving [Enter] [2021-12-03 01:36:04.769] [window] [info] [KeybindingService]:  From 20 keybinding entries, no when clauses matched the context. [2021-12-03 01:36:04.824] [window] [info] [KeybindingService]: | Resolving [Enter] [2021-12-03 01:36:04.827] [window] [info] [KeybindingService]:  From 20 keybinding entries, no when clauses matched the context. [2021-12-03 01:36:06.677] [window] [info] [KeybindingService]: / Received keydown event - modifiers: [ctrl], code: ControlLeft, keyCode: 17, key: Control [2021-12-03 01:36:06.685] [window] [info] [KeybindingService]: | Converted keydown event - modifiers: [ctrl], code: ControlLeft, keyCode: 5 ('Ctrl') [2021-12-03 01:36:06.689] [window] [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. [2021-12-03 01:36:06.692] [window] [info] [KeybindingService]: / Received keydown event - modifiers: [ctrl,shift], code: ShiftLeft, keyCode: 16, key: Shift [2021-12-03 01:36:06.696] [window] [info] [KeybindingService]: | Converted keydown event - modifiers: [ctrl,shift], code: ShiftLeft, keyCode: 4 ('Shift') [2021-12-03 01:36:06.700] [window] [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. [2021-12-03 01:36:06.766] [window] [error] Cannot refresh property when process is undefined: [2021-12-03 01:36:07.142] [window] [info] [KeybindingService]: / Received keydown event - modifiers: [ctrl,shift], code: ArrowLeft, keyCode: 37, key: ArrowLeft [2021-12-03 01:36:07.147] [window] [info] [KeybindingService]: | Converted keydown event - modifiers: [ctrl,shift], code: ArrowLeft, keyCode: 15 ('LeftArrow') [2021-12-03 01:36:07.151] [window] [info] [KeybindingService]: | Resolving ctrl+shift+[ArrowLeft] [2021-12-03 01:36:07.155] [window] [info] [KeybindingService]:  From 2 keybinding entries, matched , when: terminalFocus && terminalHasBeenCreated terminalFocus && terminalProcessSupported, source: built-in. [2021-12-03 01:36:07.913] [window] [info] [KeybindingService]: / Received keydown event - modifiers: [ctrl,shift], code: ArrowLeft, keyCode: 37, key: ArrowLeft [2021-12-03 01:36:07.920] [window] [info] [KeybindingService]: | Converted keydown event - modifiers: [ctrl,shift], code: ArrowLeft, keyCode: 15 ('LeftArrow') [2021-12-03 01:36:07.925] [window] [info] [KeybindingService]: | Resolving ctrl+shift+[ArrowLeft] [2021-12-03 01:36:07.930] [window] [info] [KeybindingService]:  From 2 keybinding entries, matched , when: terminalFocus && terminalHasBeenCreated terminalFocus && terminalProcessSupported, source: built-in. [2021-12-03 01:36:08.344] [window] [info] [KeybindingService]: + Ignoring single modifier ctrl due to it being pressed together with other keys. </pre</details\nI can't reproduce this issue in Firefox or Safari.\nyou mentioned Safari, which implied that you have macOS. Could you please re-try using Linux or Windows? macOS uses Command key for those keybindings, and so duplicated keybindings doesn't exist. Folks at cdr/code-server can't reproduce it on macOS either, but subsequently can reproduce it on Linux. Please re-open this issue.\ndoesn't reproduce for me there either !\nThat looks like Microsoft Edge, based on Chromium engine. Could you please try again on Firefox?\nI can reproduce this in Ubuntu/Firefox\nDoes not repro on Ubuntu/Edge\nCreated a PR with a workaround"}
{"_id":"q-en-vscode-15dcf26c2075b273b564ff3e85db8e295e886cb9030a31891df53738e666b732","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes In make a new keybinding: Intellisense shows many arg keys that are inaccurate or not implemented. Here is what I get for intellisense: These appear to be the correct and implemented args to this command: Version: 1.63.0-insider (user setup) Commit: Date: 2021-12-02T05:15:40.605Z Electron: 13.5.2 Chromium: 91.0.4472.164 14.16.0 V8: 9.1.269.39-electron.0 OS: Windows_NT x64 10.0.\nRelated:\nThis issue persists as of vscode I was going to open an bug issue mentioning that some params were not working... when in fact I was using the ones suggested by the intellisense e.g. instead of . A while back I opened: about how the command does not resolve variables whereas does. Also related I guess needs some love to be more consistent feature-wise with ."}
{"_id":"q-en-vscode-ef4a335dd35e6a2f2e5d9a244fe6f38cb4efe65ac13a1180d7ae1e016178d0c0","text":"Issue Type: <bBug</b(sorry, these are not good steps because I can't figure out exactly what triggers this) a language pack installed (I have the german pack) a new workspace or reload VS Code Bug Sometimes I get a prompt to switch display languages. I select no but the prompt will come back again later I see this maybe once or twice a day and haven't figured out exactly what is causing it VS Code version: Code - Insiders 1.64.0-insider (Universal) (, 2021-12-14T13:26:59.630Z) OS version: Darwin x64 21.1.0 Restricted Mode: No <details<summaryExtensions (53)</summaryExtensionVersion ------ salesforce-soql-editor1.8.0 tsl-problem-matcher0.5.0 comment-tagged-templates0.3.1 docs-view0.0.9 emojisense0.9.0 folder-source-actions0.2.1 jsdoc-markdown-highlighting0.0.1 markdown-checkbox0.3.2 markdown-emoji0.2.1 markdown-yaml-preamble0.1.0 test-ext0.0.1 vscode-eslint2.2.2 dendron0.73.0 dendron-paste-image1.1.0 typescript-notebook2.0.6 gitlens11.7.0 EditorConfig0.16.4 codespaces1.4.2 copilot1.7.4273 remotehub0.20.0 vscode-pull-request-github0.35. vscode-drawio1.6.3 vscode-postfix-ts1.9.4 latex-workshop8.22.0 vscode-platform-specific-sample1.15.0 vscode-platform-specific-sample1.5.0 vscode-memfs0.0.3 template-string-converter0.5.5 vscode-azurefunctions1.6.0 vscode-azureresourcegroups0.4.0 vscode-docker1.18.0 vscode-language-pack-de1.63.3 python2022.0.-dev vscode-pylance2021.12.2 jupyter2022.1. jupyter-keymap1.0.0 jupyter-renderers1.0.4 remote-containers0.209.6 remote-ssh0.70.0 remote-ssh-edit0.70.0 azure-account0.9.11 hexeditor1.8.2 js-debug-nightly2021.11.1217 live-server0.2.11 vscode-github-issue-notebooks0.0.113 vsliveshare1.0.5196 debugger-for-chrome4.13.0 vetur0.35.0 code-spell-checker2.0.13 rest-book5.0.1 ayu1.0.5 luna-paint0.11.0 sort-lines1.9.0 (2 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nplease ping if you need help with this.\nHere's the code that seems to trigger this: Would extension updates also trigger this?\nand I looked into this with and we were shocked to see that background updating of an extension had an InstallOperation of \"Install\" and not \"Update\" Is this correct Easy repro steps: Using OSS with the marketplace Place a breakpoint on the line linked above ^ (or just at the top of ) Install an old version of a language pack No longer \"Ingore Updates\" Reload the window Your breakpoint will hit when the language pack updates to the latest version again and you'll notice the is set to (aka 2) instead of .\nYeah, it's a bug on my side that I am not propagating the operation properly. Sorry about that. Fixed.\nI did a lang pack install and used it for half a day and didn't seen any prompts to switch display language. Is there anything other way to verify this?\nI have the German language pack installed but haven't been using it. I haven't seen any prompts to switch to German recently so :+1:"}
{"_id":"q-en-vscode-ecc8feb663b037847c3af0c8846b6fbac379b7ca1ffd1496fbd08c2d269d183f","text":"I do not see the notification toast when I select \"New Terminal\" from command palette. But I see this issue that the OP reported I think: ! ! Originally posted by in\nI see this problem: Start a new terminal - full screen Drag to bottom or right = non-functional If you split the panel right, it is fine\nReproes only when you have more than one terminal in the editor area\nCan do it with a single terminal, clean open of VSCode (mac), no other windows. !\nThanks, I specifically meant when dragging and dropping into the panel.\n/verified second terminal is no longer blank"}
{"_id":"q-en-vscode-f69e3f3483cccc800e588161a3b14bbe01a82cdc9ca476bceecf19cd0589fa94","text":"allows extensions to disable the \"Open Preview\" button in markdown files. This option should also disable the \"Open Preview\" button in the context menus. ! Originally posted by in\n/verified In stable: ! In insiders: !"}
{"_id":"q-en-vscode-08dbf1052771baf9bcf7d59e03a9062edb3a6acc92337c5955b84d56bae7288d","text":"Parent issue Update: Activate the fish shell integration by following these instructions: We're keeping this open until automatic injection is looked into.\nDeferring to focus on polishing how shell integration works in general\nFinal implementation details aside, fish's event handing will make at least some of this dead easy: Some will be more difficult, at least to deploy auto-magically, because fish's are the output of shell functions, not the contents of variables: So to transparently preserve the user's original prompt, like the bash script does, would require modifying a shell function on-the-fly, not just a variable. Yikes. That said, I put the above code, plus the following, in my own config— —and I had all of VS Code's working perfectly with VS Code v1.70.0 + fish 3.5.1. It's only a proof of concept, but hopefully this saves somebody a few web searches down the line, at least. :)\nvery handy info thanks! The bash script handles both PS1 and PRONPT_COMMAND, might be able to wrap the function automatically in a similar way?\nWell, fish doesn't have either: if you want something to run right before the prompt displays, you put it in a function with . There is, alas, no event handler that runs after displaying the prompt but before accepting input (which is where you'd want ). There are also no events specific to the right-side prompt only. FWIW, there's also no equivalent in fish—\"instead it leaves the lines indented to show that the commandline isn’t complete yet\"—but the lack of and sequences in the above didn't seem to have any noticeable impact on VS Code's ability to parse the terminal in my testing.\nah I see, I think this stuff should still work though. We actually don't currently use some of these positions and the 633 E sequence kind of removes the need for them all together. So if there are limitations like this we can just take note for the future, like when finalizing the sequences this month. The limitations imposed on us by the Windows backend are much more frustrating to deal with than this sounds If you want to put a PR out with just the script as a starting point (just the script) we could then give it a try and see if we can get it in as an experimental thing. I guess just manually installing until it's considered stable, plus the automatic injection was a handful for the other shells. You could also move the 633 A to just after 633 D, I'd expect that to work fine.\nI've been wanting to do this since last month, and comment finally motivated me to put up a fish plugin here: The trick I have seen in is that they wait for a event and then initialize the code, at which time the prompt function should be ready however they are defined.\nFYI those sequences may change as they are yet to be finalized ().\nNow that has reminded me fish can do e.g. (which I just... totally blanked on yesterday ), I think I could put together an experimental PR for this—one that could be easily tweaked if/when the escape sequences are finalized.\nAlso thanks to a PR from the fish history file is now loaded so pulls from it when fish is active !\nthat would be fantastic\nThanks everyone! The script should be available in tomorrow's Insiders, see for manual install instructions (replace with ) I'll keep this issue open until we no longer consider it experimental.\nThe new function does not work properly in my case. My fish configuration uses a right prompt and shows the exit code of the command (green check mark). The exit code is always shown as good, even if the executed program returns a non-0 code. The right prompt is also misaligned. It is one line below where it is supposed to be. ! The status code issue seems to be caused by executing first, because this will overwrite and . The misaligned right prompt seems to relate to some newline shenanigans fish does. It removes the last newline from the prompt, but since is after the prompt it doesn't work any longer. I found these related links: , I change the above snippet to this it seems to work better for me.\nthanks for the report! Do you want to submit a PR or should I try out the fix?\nI am not interested in progressing this further or submitting a PR.\nno worries, thanks for reporting\nIsn't this already done? Or is this about automatic integration?\nI was keeping it open until the kinks were worked out, plus automatic injection would be great if it's possible. Not sure anyone has looked into it yet.\nshould we change the title of this issue or create a new one to capture automatic injection?\ni think the right way to add automatic injection would be to append a directory to that contains the integration script in a subdirectory (see )\nwe wouldn't be able to stick it in XDGDATADIRS though in case a user has a file there already. Changing XDGDATADIRS also seems dangerous and I think this would only work on Linux\nit sounds like you may be confused as to what XDGDATADIRS is so i whipped up a pr for this, with links to the relevant spec and fish docs. alternatively i'm confused as to what you might mean. XDGDATADIRS is a set of directories to search for data files, ordered by preference. we'd be adding a directory to the end of it so what a user may already have there is irrelevant, and our directory would be part of the vscode distribution so the user really shouldn't have any files in there. aside from that user things go in XDGDATAHOME, XDGDATADIRS is for system components. i'm not sure what danger you see there. we're adding another directory to find data files in; we're not preventing any other data files from being found. this is documented universal fish behavior on any system it supports. i linked the literal default fish configuration above that demonstrates this. also back when i last used a mac most tools did support the xdg base directories, including fish (and nix, which makes rather heavy use of XDGDATADIRS). the only ones that didn't still don't on linux (like electron ).\noh it's a list of dirs, I thought this was setting a single dir. And you're right we don't need to worry about Windows here. Thanks, will check out the PR soon\nSince the issue is now closed, is the problem reported here something you would still want to fix? It doesn't seem to be fixed on main so far.\nthanks for the reminder, I forked it off to\nTo verify: Make sure fish is installed on your system (macOS or Linux) Launch a fish shell and verify the shell integration decorations should up to the left of the terminal\nyep, works for me!\nUnfortunately this does not work in a newly-installed fish shell via homebrew on macOS. https://user- Pty host logs:\nwhat's your ? Also what is the output of ?\nworks for me with this info: !\ndid you manually install shell integration? This issue is about automatic injection, the update in the original issue was to help users manually activate it before this was done.\nYes I did\nand disabled automatic injection (set ) as you see in the screenshot. it also works for me in a subshell, so I know that i'm testing / verifying manual injection\nit also works for me with automatic injection\nis 3.6.0, and the shell integration path is\nLooking at the code it doesn't look like we actually copy a file to :\nhm, the path does appear to exist <img width=\"1045\" alt=\"image\" src=\"https://user-\"Manually executing it with results in an exit code of 1 and no output, however\nyeah I think what happened was I already had the injection config in my fish file when I tested automatic injection - likely from testing manual injection at one pt\nArgh, I had removed the manual integration but still have other FinalTerm-compatible shell integration and probably thought this was working because of that On 26. Jan 2023, at 18:19, Megan Rogge yeah I think what happened was I already had the injection config in my fish file when I tested automatic injection - likely from testing manual injection at one pt —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID:\nTo verify make sure you don't manually activate shell integration in your , then check with and without and make sure shell integration works and doesn't work respectively.\nWhen I open a terminal in the fish shell and run some commands, then run \"Run recent command\", the commands in the fish terminal don't show up, I assume that means it's not working for me\nworks for me: <img width=\"741\" alt=\"Screenshot 2023-05-30 at 2 43 13 pm\" src=\"\"Do you see the grey/blue/red dot?\nNo, I guess shell integration is not enabled in the fish profile, I'm not sure how to troubleshoot that.\nIs there supposed to be automatic integration in stable VSCode now? It doesn't seem to be working for me, but manual integration still does. Open and make sure it contains the manual integration code:\nhe's trying to test the newly fixed automatic injection in Insiders\ncan you run ? Also enabling trace logs and getting the pty host log when launching would be helpful (be mindful of any sensitive environment variables)\nHere's my pty host log which includes\nthis isn't working for me after I removed my !\nwhat do you mean by removed your XDGDATADIRS? weird, it's there so it should be run afaik. any idea what could be going wrong here wrt the XDGDATADIR auto activation?\nI think I had manually enabled fish shell integration which is why it was working. In order to undo that, to test automatic injection, I removed that directory. Then it didn't work\nIE my expectation was after undoing manual injection, it would work automatically. That wasn't the case.\nI had misunderstood how this injection works because it is different from that of and . Deleting my in the comment above was the reason I saw it not working. I now see it working. !\nsince it's working for and I we'll close this and include it in the release notes. It would be good to get to the bottom of your issue, we can track that in another issue? We can have a call to investigate when you're free if that works.\nSure, call any time\nNot really, sorry. Maybe the snap packaging interferes somehow? From your shell session, are and present?\nWe investigated, it ended up being the fish version being too old. After the update it worked great! Thanks for all the help\nThis isn't working for me in 1.79.0 (non-insider), as far as I can see the fish config is there but XDGDATADIRS isn't getting set. And looking at the code I found -- what happened?\nwe had to disable it because it was printing an error when starting up for someone on the team: For some reason was undefined at the time, it got set later, after our applied.\nLooking at again I believe this would happen when something we need, like , is configured in either or in a file with a basename that's sorted before the one we expose. The file is intentionally read last so people can override vendor configuration, so since we need to run after user configuration maybe providing vendor configuration is just not the way to go here… sorry Maybe instead we could invoke fish with ? That would ensure the integration happens after any configuration but before* interactive input. If I have the spoons and noone beats me to it I'll try to whip up a PR for that next week.\nOne way to make sure something runs after the user's but before interactive input is to put the necessary code in a self-erasing function tied to the , e.g.: This can be placed in the appropriate vendor-config directory as simply (e.g.) with no need to worry about sorting order. The function will be defined during startup, executed just before printing the first prompt, then immediately undefined so it doesn't run twice.\nI tried something like that thanks to your suggestion. It's not working atm and not sure why... lmk if you see what could be wrong here\nI think you also need to add to this function:\nRemoving verification-needed as this was covered by a test plan item."}
{"_id":"q-en-vscode-a4469e191077243c3384bb1ea4ba932a08617c398431723d3a64ec787c9289de","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --Hi! I want to solve this issue as a good-first-issue. Please help me if you can :) In Github there was a problem that README and issues don't work for RTL languages. () Then on November 16, 2021, Github announced that we fixed this problem: () But vscode does not support it yet. For example we see ) in github like this: ! But in vscode, we see the markdown preview like this: ! Feature or improvement request: Providing support for correctly previewing RTL languages texts.\nHi! Can you help me to solve this issue? (as a good-first-issue) I opened a pull request here:\n/verified"}
{"_id":"q-en-vscode-83021b20165cc6a6c6fcc41ebb2d4d548c953544fc493f69519b9e7a01c08b75","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: stable and from sources Steps to Reproduce: server vscode in safari simple browser :bug: Observe icons don't load (icons load if disable Cross Origin Restrictions in develop menu is checked)\nHow are you running the server? I see the icons loading in but not on localhost using : ! This seems expected since Safari seems more strict about loading content from localhost in this case\nTry codespaces, it repros there too. (For running the server I use ) It works on serverless (i.e. and ) because they are served in a different way I think\nI'd also like to see it fixed. It's still reproducible on latest Safari version. <img width=\"290\" alt=\"image\" src=\"https://user-\"\nThis definitely seems like a webkit bug but I can't put together a minimal repo. The issue is the the codicon font face doesn't go through the service worker during loading despite it's parent css file being loaded by it: ! I also tried debugging the production code in codespaces but Safari's dev tools are pretty much unusable with our minified service worker Instead, I'm going to push a change that inlines the font in the css file to hopefully work around this issue\nLooks to be working in latest insiders !\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\n/verified"}
{"_id":"q-en-vscode-0ad0842faad09a4a06e94d511a030a70d7991a344f2ec89c7f08ab6a0621c713","text":"Issue Type: <bBug</bthe Extensions tab enough text in the search box to cause scrolling back to the beginning of the search box Actual: The typed text overflows the right edge of the search box Expected: the typed text doesn't overflow the edge of the search box <img width=\"210\" alt=\"Screenshot 2022-01-09 at 18 00 24\" src=\"https://user-\"VS Code version: Code 1.63.2 (, 2021-12-15T09:38:17.605Z) OS version: Darwin arm64 20.6.0 Restricted Mode: No <details<summarySystem Info</summary </details<!-- generated by issue reporter --"}
{"_id":"q-en-vscode-88fa059c970fc39869b85d520497c1f457334bc821978b4605a005ec5eda8b54","text":"The 'Last updated' and 'released on' dates on the extension page use the American format. This is confusing to many people. ! I tried to Spanish and German, but the problem remains: ! ! Consider using a unambiguous date format such as yyyy-mm-dd (ISO 8601)\nI am not locating the datetime as a resource for localization, I'm assuming it is concatenated in runtime so not a loc issue. Please help if I'm mistaken.\nI think using the unambiguous date format is a good way forward. However, we could do something where en-US is displayed one way everything else is displayed the other way... but I don't think that is 100% accurate.\nis this issue closed? or still needs help?\nIts still open and there is a PR which has to be reviewed.\nHello, is this issue still open? I am new to contributing to GitHub projects and this seems like a good first issue for me. Is there a way to claim this issue? If I am able to work on this, do you have any advice for me in contributing for the first time?\nThanks for your interest. This is already being worked on. Please look for another good first and help wanted issues. Thanks/"}
{"_id":"q-en-vscode-e36138300b3fe36a633132ce898a76b6012027220087a5d94898d0b3bd61a43a","text":"<img width=\"600\" src=\"https://user-\"<img width=\"600\" src=\"https://user-\"<img width=\"600\" src=\"https://user-\"<img width=\"600\" src=\"https://user-\"IMO the labels should be consistent and we should use the icon + label as much as possible: !\nI like your idea of showing the icon + label. But would like to show it next to the version as they are coupled. Also showing an addition line for a pre-release extension in the hover seems redundant to me. So, I just icon to the existing label and this is how it looks: ! !"}
{"_id":"q-en-vscode-157a6f42988b6361cf5d0f517e4184111d68a664c99b36f8110cd61c275ebd2a","text":"Testing On macOS Ensure the setting is not modified (default to ) Create two folders and , create file in Drag it to Undo There isn't any prompt of warning\nRenaming a file and undo doesn't warn either.\nThese are not destructive operations, correct? The original expectation was that all destructive operations would prompt, do you think that any operations at all should prompt?\nThe way I interpreter is it will warn is it prompts and tells you what's going to happen when you undo/redo as the undo/redo is focus based and it's likely to mis-trigger. I don't necessarily want to disable this feature (as it's convinient) but I want to be aware of what's going to change. Undo the dnd is one of them as I might miss where the file will be moved into (especially when I'm traversing in the undo stack).\nChanged the config around to better stratify this:"}
{"_id":"q-en-vscode-daf3c1b306cf014bc3ebb4be8fc45b9b4ea9b203cb3207c120c8aefda3b8d145","text":"<img width=\"452\" alt=\"Screen Shot 2022-01-27 at 11 21 24 am\" src=\"https://user-\""}
{"_id":"q-en-vscode-802a85067a5676c0cf65026601cd1398cb5b601a6106927c7cb2f3db6f9a522b","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.64.2 OS Version: Windows 10 Steps to Reproduce: a git repo in root of a drive (network share in my case) a change of a file to Source control a see file diff version is shown as empty command line shows diffs properly when git repo is located in some subfolder git diff works in VSCode properly\nVS Code Version: 1.67.2 OS Version: Windows 10 () I am also receiving this error, and I have pinpointed it in the Git Output, hopefully, it helps the path is missing the first character in the path for a specific file on the 'git show --textconv' lines 'ub-folder' vs 'sub-folder' If the command is run manually, with the correct path there are no issues Working (Non-root folder) [2022-06-01T06:49:42.910Z] git ls-tree -l HEAD -- fserverc$gitsharesub-folderMapper [229ms] [2022-06-01T06:49:42.991Z] git show --textconv HEAD:sub- [318ms] Not Working (Root folder) [2022-06-01T06:51:54.868Z] git ls-tree -l HEAD -- fservergitsharesub-folderMapper [226ms] [2022-06-01T06:51:54.897Z] git show --textconv HEAD:ub- [262ms]\nThe issue is in the file The following Line needs to be changed to As the current process is removing a specific length, when sometimes the variable for 'to' can contain a single slash or double slash (which the + 1 would handle the double slash) I recommend using the regex to effectively trim any leading slashes rather than a calculated length. I have a working file here I can provide further details if required as well.\nthank you very much for looking into this. I have looked at this today and while adopting a regex to remove the characters from the resulting path will work, I think that there is something else happening here that I would like to understand before making a fix. Debugging the issue I have noticed that when you open a folder from the root as described in this issue, the value of the parameter contains a trailing . This is why adding +1 to it, will actually remove an unwanted character. I will investigate this further and will get a fix in this week.\nQuick update on this issue. While the method can be improved, the root cause of the problem here is that returns a path that has a trailing character. This then results in the repository root being incorrect which can result in unexpected behaviour, diff editor not working being one example. This method already has code to handle mapped drives but it does not seem to handle this particular scenario. As we are in the process of shutting down for the next Stable release, I will be looking at fixing this issue sometimes next week. I will update the issue when the fix will be released in Insiders. Thanks for your patience.\nI believe I have tracked down the root cause of the issue to the call: Running the following tests clearly illustrate the problem. Notice that the second call contains a trailing . This particular behavior seems to be by design based on an old issue that I have found - One option that I have considered was to remove the trailing slash from the repository root but that does not seem to be a good option as there are Windows path that are only valid if they contain a trailing slash (ex: ). I am more inclined to tweak the implementation of to always use as that seems to be able to correctly handle directory paths with the trailing . what are your thoughts? Thanks!\nHi, I'm also having this issue, even on the insiders build. I disabled all installed extensions. Expected Git Output: Actual Git Output: The drive in question is a mapped network drive. Whenever I try to view the file diff, the previous file is blank. This problem applies to all files on the network drive, even ones in subdirectories. Let me know if there's any other information I should provide.\nI have file to track the problem that you have reported.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nYes\nI'll mark this issue as verified as part of . Please file a new issue if the bug continues to occur on the latest Stable and/or Insiders versions of VS Code."}
{"_id":"q-en-vscode-1f22cf7616d6a1efa5e982e948ba14a837a5a2f05f7c8261ffe6441eea2f1dce","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --Use to create if exists in directory. Not just using latest commit to create .\nAre you using an extension which adds the ability to create a PR? If so, please report the issue on the extension repo. The Report Issue dialog from VS Code's Help menu may be able to assist you. /needsMoreInfo\nYes, after recheck,it's a feature of extension,which seems to be a builtin exttension.\nCan you give me the full name of the extension? Maybe a screenshot of it in the product? Or the view where you can create a PR?\nThe full name is exactly <img src=\"https://user-\" alt=\"drawing\" width=\"200\"/Here is I create Pr !\nI'd like to fix it. Any leads are appreciated.\nSure, this would be a great starting point:\nVerification: and clone a branch a change to and commit to push/publish that branch. Code should tell you you can't push to that repo and will offer to fork the repository, accept. should then ask you whether you'd like to create a PR, accept. You should get shown exactly 9 PR templates to choose from. Reference:\nI tried but after selecting a repository to fork (which then successfully creates a fork on Github) nothing is pushed without any error message. However, the developer console and the Git extension logs show this: ! It seems the source of the problem lies when calling : Maybe no rename occurs when calling (with no exception thrown).\nVery strange, can you easily reproduce? Delete your fork and try again.\nYes. I checked it several times and the results were the same.\nI did get the list of PR templates: ! However, I had no way to get back to the picker after taking that screenshot (which dismissed the picker). At minimum I feel we should use , since the Create PR functionality is not exposed as a command palette command, I am stuck. Additionally, after I selected a template to create the PR, nothing happened. Finally, it was unexpected that I get two popups to create a PR when the GitHub Pull Requests and Issues extension is installed: !\nYou mean no PR was created, right? Nonetheless, I think the Github extension needs more visible error messages (e.g., via notifications).\nYes, no PR was created (I checked on GitHub).\nCreated two issues from the discoveries. Thanks!\nActually three. Creating a PR just doesn't work at all anymore:\nFor step 4, when I try publishing the branch, I get the error \"Git: remote: Permission to joaomoreno/test-pr- denied to rzhao271.\" There is no option to create a fork: !\nIs the Github extension enabled for you?\nJust confirmed that it's enabled for me.\nCan you file a separate issue?\nVerification of this issue fails since the user is no longer prompted to create a PR as found in :/\nWait you hit this too? I totally can't repro, that's why I asked to follow up with a separate issue. We can move this to May just for verification sake.\nYeah I hit the same thing as\nI just tried testing this: On pushing the new branch, I get ! and then: !\nReopening due to comment but feel free to do what you want with this\nI got the same issue on the latest Insiders <img width=\"1202\" alt=\"image\" src=\"https://user-\"Command output: Remote URL:\nthis is\nMoving verification to July to account for"}
{"_id":"q-en-vscode-7c83c36523a0597d3d39ddbada578e546e9260f8402984a2a9c40aef2b9b563e","text":"Issue Type: <bBug</bmicrosoft/vscode-extension-samples the helloworld-sample folder , build, verift that its contributed \"Hello World\" command activates the extension and shows a notification. the vscode version references in in two places to use 1.43 minimum. code to activate function in , for example: , run, use \"Hello World\" command to activate. breakpoint on return line above. a folder. Explorer button or context menu to create a new file. Confirm that the event fires and your breakpoint is hit. from breakpoint. Explorer context menu to copy/paste or use drag/drop with modifier to drip as a copy. :bug: The event callback does not run, despite the API doc stating: Issue also exists in 1.64.2 VS Code version: Code - Insiders 1.65.0-insider (, 2022-02-21T05:15:17.930Z) OS version: Windows_NT x64 10.0. Restricted Mode: No <!-- generated by issue reporter --\ncc\nthe extension host's event is only emitting from 's event, but the explorer is using the 's , which doesn't seem to talk to the working copy service. What do you think the fix here should be? (cc\nthe bulk edit service is using the for its operations. If you look at , there is a bunch of usages for that service. The working copy file service is a wrapper around the file service that ensures the relevant events are emitted and can be participated on. I suspect that for the copy/paste operation or drag/drop we are possibly not using the bulk edit service but the file service directly? I suggest to make a pass over all explorer code and look for raw usages of to replace with bulk edit service. I think that would even enable undo/redo properly.\nMy bad, the real bug is that this block: doesn't handle events. cc\nPragmatic fix would probably to treat the same as ."}
{"_id":"q-en-vscode-1836dfa7bb6b6ae89ddcf4de90013de7446b6e1d478e32ae4f1c555b357fed34","text":"It seems something odd is appearing in our builds, specifically for our \"video\" svgs that has a link. The SVG doesn't contain anything crazy: Which throws the following error in the console: The strangest thing is in OSS everything looks fine but in the builds, the image is small and not centered. <img width=\"1089\" alt=\"CleanShot 2022-02-23 at 09 01 27 src=\"https://user-\"<img width=\"1024\" alt=\"CleanShot 2022-02-23 at 09 01 14 src=\"https://user-\"cc\nIs this new? cc any recent changes to CSP?\nI see this in Stable as well, so we likely never caught it\nThis is happening because the is being stripped out of this svg at some point during build."}
{"_id":"q-en-vscode-2c23da1b1f437c645fef2cd60880ff1da8dc136eedcbabff8ca3e86eb87af192","text":"Just broke again:\nAdded a timeout to the initialize request which will tell me if that's the part that's timing out, but it's probably not. Any guess\nDuplicates"}
{"_id":"q-en-vscode-9f9065c3e2676035d5fda9209406cbeac10568aee06b4ead89ec900984e33304","text":"This video shows text in another cell being undone even though focus is in a different cell. Doesn't seem to matter whether \"Notebook: Undo Redo Per Cell\" is enabled or not ! This video shows that after a cell as been and text has been to it, Ctrl-Z can be used to remove it, but it cannot be redone. You've lost anything typed in that cell permanently. !\nMarking as important and a regression as I can't seem to figure out whether the \"per-cell\" setting is working the way it's intended or not.\nActually it appears that I may have disabled the \"per-cell\" setting a long time ago and didn't realize it. This may explain both behaviors above. I didn't realize that after changing that setting VS Code must be reloaded for the setting to take effect.\nSo removing \"important\" and \"regression\" for now. Let's just review what the behavior is that we want. It does seem problematic that if you have the setting off, that it doesn't seem possible to restore a cell you may have accidentally deleted. When the setting is on, Ctrl+z seems to work as expected.\nAfter internal discussion, re-marking as important. It's a data loss scenario if you have cell-level undo off.\nPossibly related to\nWith Cell Undo / Redo on (this is the default): Undoing and redoing text per cell works as expected Yes, with my testing Undoing and redoing add / remove of cell works as expected No, see bug 3 Undoing remove of cell via delete button No, see bug 4 With Cell Undo / Redo off: Undoing and redoing text per document works as expected No, see bug 1 Undoing and redoing add / remove of cell works as expected No, see bug 2 Undoing remove of cell via delete button No, see bug 4 Bugs: 1 Two cells in document. Changed from per cell to document mode. Typed in Cell A. Moved to the second cell, hit Undo. Instead of undoing the text, it blew away the second cell. Typing Redo did bring it back. I expected the typing in Cell A to be undone. Note, this bug I don't see after reloading the window. Might be that the setting change mid flow got the stack into a funny state? 2 Two cells in document. Shift-enter to run and add a new cell. Type in new cell. Ctrl-Z makes the typing go away and eventually the cell. When the cell is fully removed, redo will not restore it, so the typing is lost. Note, I can only replicate this will per cell turned off. With per cell turned on I first need to not be on the current cell to make it go away via Undo. But in that case I can make it go away with Undo and the text is restored with Redo. 3 However with per Cell on the cell + text does come back, but when it comes back it comes back without the per cell undo redo stack with per cell turned on: 4 When removing cell with the delete button, it cannot be restored with undo. This is true with either setting on. This is what I saw in my first pass. My personal ranking in importance would be like so 4 - Clear data loss. In both modes a cell is deleted and I can't get it back and the info in it. Easy to do and damaging. Effects both modes. 2 - Also clear data loss. I find this as a bit less important as I can only repro it with perCell off and I need to Ctrl-Z through the cell to get to the cell loss case so I've already removed what I typed in to the cell to get here. 3 () - Data loss here, but the data loss is not the actual code just the undo redo stack. So I'd vote for less important. 1 - I only reproed this when changing the perCell setting without a reload so an issue, but lesser importance.\nSome related issues:\nNote I'm going to use this issue specifically to resolve the issue where with perCell mode turned off you cannot restore a cell that was deleted with undo. The only specific scenarios in this fix are as follows: a notebook document a new cell (type something in it if you want) the cell using the delete button in the toolbar +Z (Run Undo) from a previous cell in the file the cell would not be restored, now it will be Scenario 2: the last cell in a notebook document shift-enter to run the cell and create a new cell some word in the new cell using Undo in that cell, it should walk back the words typed and eventually delete the cell start pressing Redo, it should restore the cell to the document then restore the words typed the cell would not be restored.\nthanks for the detailed investigation, they are insightful! For bug 1, you are correct that flipping the setting messed up with the undo/redo state. We register based on the setting which is not updated when we flip the setting. We might want to dispose it and register a new one when the setting changes. For bug 2, with your fix in, I can't reproduce it with the per cell setting off. Not sure if it's similar to bug 1. For bug 3, your comment is accurate. It's like a document and we don't persist them usually when a resource is gone. I wonder if we should fetch and \"save\" the undo/redo stack for the cell on the cell deletion element in the undo stack.\nReopening to cover the issues with flipping the setting. I'll close after that fix and will cover the bigger issue on persisting undo / redo per cell after delete.\n/verified? Still not sure this is working. Is per cell undo supposed to allow undo to affect other cells? Repro: to per cell undo in first cell in second cell +Z Expected it to only affect the current cell with focus. Actually it removes the typing in the first cell.\nOtherwise seems to behave more like I'd expect."}
{"_id":"q-en-vscode-6caa0a8d7e73b07f06d6268a756e3d9de5b424453f93b7d9660ea3a7e3122655","text":"Repro: in a terminal and select something ! We should hit ctrl+c beforehand () which should work properly on all shells\nWe should also do this when is clicked in the decoration's context menu\nI attempted this in but it was more complicated than I originally anticipated. Moving to backlog."}
{"_id":"q-en-vscode-42db0d76263239affccd24e6c9d9efea58be9033d629675f1cad9be8e05e7fcf","text":"Open following url Install dialog is shown asking user to install and open extension <img width=\"275\" alt=\"image\" src=\"https://user-\"- After clicking on install, following prompt is shown <img width=\"454\" alt=\"image\" src=\"https://user-\"Second prompt seems to be unnecessary. It is expected to open the URL after installing"}
{"_id":"q-en-vscode-b447de76605ddd7c5f9f1dfc77b4163eb872d873838eb61b525d4bcc51121332","text":"System info: macOs, Zsh, shell integration on. to vscode folder and hit tab it autocompletes to Hit tab again to autocomplete the names, it instead creates a big space and adds again. !\nNot sure if we're tracking this one yet but I've seen it too in pwsh/Windows.\nWorks fine in pwsh.\nIt only occurs for me when shell integration is on. I've had to disable it in the meantime\nMaybe I didn't see it on Windows actually. Happens on zsh/mac when pasting followed by enter/left: <img width=\"418\" alt=\"Screen Shot 2022-03-17 at 10 57 44 am\" src=\"https://user-\"\nOh I bet I know what's causing this, we need to wrap \"invisible\" parts of the prompt in and\nSimpler repro: tab <img width=\"242\" alt=\"Screen Shot 2022-03-17 at 11 08 21 am\" src=\"https://user-\"l"}
{"_id":"q-en-vscode-d009257ae05865f01ab5015522d5fdd097a226b9f5448df12c96c545c01cfd0a","text":"The original issue comes from Comment from Looks like there is some problem with the editor focus when the test message view is peeked out?\nI found a way to view test results and work with tests. When the tests are executed, I need to press CTRL+m for tab key to move between elements, then tab until the close button and after that panel with buttons to go to previous and next failure results there will be a tree with test results. That's accessible with NVDA, I am sure it's also accessible with Orca. Thank you for help all!\nThanks for the issue! at the moment we reveal the failed position and then focus the editor. Instead is it preferable to focus within the peek view like the references peeks do?\nyeah, it would be better if you focus inside the peek automatically for accessibility reasons. If you want, you can also differentiate if the command is triggered via keybinding, if yes then move focus, otherwise do not. This is what the editor hover does. Though I think it might be a bit confusing in this case, so I suggest to just pass focus inside the peek.\nThank you for such an active feedback! I also think that it would be more convenient if the focus moves directly to the peek instead of moving it to the editor with the selected line that can be easily deleted by accedent. Thank you!\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nhow about some verification steps unless resurfaces?\nSteps in the original post should still be accurate. Opening test peek output (for a failed test) via should focus and screen-read the peek view, rather than staying in the editor."}
{"_id":"q-en-vscode-690b5f793cb158ebd4860c099dc54fe404e7ef922cc6e3f92c53439f469e5382","text":"The fix for at only replaces the first instance of in the string, not all of them. I suspect the regex needs on the end. <img width=\"1028\" alt=\"Screenshot 2022-03-17 at 14 46 58\" src=\"https://user-\" FYI :)\nheh, I found that and was going to change it, but GitHub fell over. I can easily add the , but it's not clear to me where/how to write a test against this class (which is actually the part that I get stuck on any time I want to contribute something). I can't find any similar tests in this area to use as a template.\n/verified"}
{"_id":"q-en-vscode-27e000a7b793d7a6612df0d13caae90af2df96559af99f0a2bb60ae2212e72ce","text":"Steps to Reproduce: command a language Settings file is opened. Settings Editor with language filter shall be opened given that the editor supports it."}
{"_id":"q-en-vscode-c762cfeda913fee3590284fced772f67355268f4ee7e677742a14c91c592dd0f","text":"Forgot to respond and got closed due to inactivity. - We don't necessarily need them to read each item, but at the minimum screen reader users need to know that these buttons exist. Even a small announcement along the lines of \"tree items actions available on tab\" would suffice. Repost of:\nDoes this issue occur when all extensions are disabled?: Yes/No VS Code Version: 1.63.2 OS Version: Win11 Steps to Reproduce: a debug session in VSCode on Narrator keyboard and navigate focus to the debug target ! Expected Result: Narrator should announce all context, including action items associated with the treeitem Actual Result: Narrator announces just the tree item text and associated action buttons\nI am not convinced this should be announced. As Rob already mentioned on that issue the user can just press tab and focus onto the toolbar. We do not do this in other places that the parent element mentions all the children element inside it. Or the element that is about to get navigated to next. Let's leave this open for now to see if there is some feedback from users regarding this. For now I suggest no action.\nNot announcing anything does violate accessibility standard guidelines as visually impaired users who rely on screen readers wouldn't know that there are buttons associated with each tree item. I am filing this issue on behalf of the project where our products are required to meet accessibility standards and this component has been flagged as an accessibility violation. I understand that the prioritization might differ, but I wanted to give some context behind the issue.\nAgreed that \"they could just hit tab and find them\" is neither valid to fulfil our accessibility needs nor is it something we can expect from users who rely on navigating without seeing. As a quick fix, is there a chance to add an aria label to elements like these that announces \"more functionality available\" to explain to people that hitting tab will get them there?\n- Instead of adding extra announcements, we could also change the type of the component to be a tree grid/tree grid item. Screen reader users should expect more content on tab if they are on a tree grid item. Any thoughts on this solution?\nthanks for your creative suggestion, but a is something that is like a table and might not fit here. Check out the example here does it make sense that VS Code trees start using a role because each element might have extra content inside. How do Screen Readers treat vs ?\nThe suggestion came from the Edge Accessibility team. The core idea is that the existence of interactive buttons/content should be inherent based on the role of the element which is true for a tree grid item but not a tree item. As for screen reader behavior, treegrids are announced like tables, but they also state item position and level similar to a tree item. For instance, when you enter the grid, it will say: \"Entering table, 5x2... etc {contents of the row} expanded 1 of 3 level 1\" From this, the user will understand that there are additional cells with content to the right (maybe just one extra cell containing the toolbar with buttons) and also that they are in a tree structure.\nThanks for clarifying. Also are you suggesting that all trees in vscode have the role? Since every tree could have additional actions inside. What's the downside of using the role? Let's see what thinks when she finds time to reply. She is the author of Orca screen reader.\ntl;dr I think Tree Grid might be appropriate here. Long form: I definitely wouldn't expect interactive elements inside a normal tree view item. The two interaction models I've seen on those are: a) Put these items in the context menu. b) Have the user press Enter on the item to enter into a secondary layer where one would then tab through interactive elements. I first saw this in the old FireBug developer tools extension for Firefox, and later, while I was working at Mozilla, we also took this approach in the built-in developer tools in the HTML tree inspector. A tool bar with buttons etc. inside a tree item is certainly not Windows-Standard, and Windows standard is what most ARIA specs and authoring guidelines primarily derive from. Tree grids come more from the Linux side, where this is at least I saw them first in GNOME Nautilus. So yeah, I would lean towards a tree grid model where one would right arrow into the buttons for an item, if available, or tab to the toolbar, etc. We'd have to play with the aria-readonly property on this tree table, since some screen readers sometimes think they're editable when they're actually not, or vice versa.\nthanks for chiming in, I really appreciate your expertise. In VS Code it is pressing to get into the toolbar that is inside the tree element. Some follow up questions for everybody on this thread so we decide how to proceed further: about the Breakpoints view? This has the role and also has interactive elements inside each ? there some unintended consequences of going with the instead of the in all our trees? Or should we use the only if there are interactive elements inside? Example: Explorer does not have interactive elements, but Search does. We need to have a consistent experience across VS Code, thus these questions. fyi\nis a good question. I'll consult with our Accessibility team, but my initial thoughts are that we would need to change the role from a listitem. I'm not entirely sure if the best fit is a tree grid element or another role. wouldn't change all trees to treegrids as each role has its own set of implications and ignoring that would confuse screen reader users. If all these trees are derived from the same base class, I would add a constructor argument to differentiate treegrids and trees.\nLooping back for the first question - I believe a grid/grid cell role would be best for that kind of component.\nLet's assign to May as I would like to experiment with these changes at the start of a milestone\nI have changed the role for the whole CALLSTACK to be a for this use case and for each item to be a . So now as you say users should expect more content when they press tab. If users get delighted by this change I would be open to do this in more views.\nHey - I appreciate the fix! Unfortunately, the component that we're using in our extension is the . Would it be possible to port the same fix over to this component considering this is the actual component our accessibility team is flagging up?\nevery tree item has an field which your extension can use to manipulate the role or the label of each tree property. I suggest you try using that. For example you could try different roles, or to make the label more verbose as Christian Heilmann suggested here Allowing extensions to change the role of the whole custom tree is currently not possible, and if that is what you would need then I would have to discuss with Alex Ross who is currently on vacation and this would require API changes."}
{"_id":"q-en-vscode-19350a4955672e97c53887f5ff72423265704fbd2a35131cf3205a00ba0a6312","text":"Hi ! I am running into an issue issue when using vscode 1.66 : when running pymc models there is a progress bar output which was working fine before the update and now the progress bar simply prints down :( See pic below ! This is how it prints in Jupyter notebooks (anaconda) and it was looking so in vscode 1.65.2 ! Here is the info about my session\nUpdate It actually works when I am not looking ie, I am not in that specific cell of the notebook ¯()/¯ !\nHere is a reproducible example: Code: Output: !\nI'm having a similar issue using enlighten progressbar. This likely started after a recent vscode update. A new progressbar is printed everytime it updates/redraws. I'm guessing this is the same issue. (This works/prints fine when run using jupyter-notebook on a web browser) !\nThank you for working so fast on this one! Looking forward for the next update/release !\nI couldn't verify this fix, using : !\nthanks, it's not being picked to the release branch yet.\nWorks for me too. !"}
{"_id":"q-en-vscode-cde009137b0e09615d9f8c3522cdd2eb9025f65af4707e139ac320aa0b4216c4","text":"System: Red Hat Enterprise Linux Workstation release 7.9 (Maipo) OS: Linux myhost 3.10.0-1160.36.2.el7.x8664 SMP Thu Jul 8 02:53:40 UTC 2021 x8664 x8664 x8664 GNU/Linux This began as an occasional annoyance and has escalated into a 100% reproducible annoyance. Using the latest snap install, during normal editing/debugging/etc. on local system. With increasing frequency, Ctrl-O/File-Open would completely freeze the UI with just a blinking cursor. Sometimes switching away and back would repaint part of the screen, increasingly not. Tried switching the output window to various things before attempting the file open (extension manager, main window, etc), but no output was observed before the freeze. Once the UI is frozen, only kill/killall seems to be able to get rid of it. Not sure if it means anything, but de-maximize and maximize still resize the window frame, although the contents aren't redrawn. Can't remember if that requires app cooperation or it that is a pure WM thing. Using snap installed code-insiders made no difference. Uninstalling snap versions and downloading 1.65.0 from the MS repo so far seems to have restored functionality (1.66.0 requires a glibc version unavailable for RHEL7). Haven't tried with (all) extensions disabled, but did disable everything I don't need constantly. One change in workflow recently seems suspiciously coordinated with the start of the issues - began using VS Code on another machine (nfs-mounted home folder, so shared settings/etc.) to attach to running containers (as container-specific user, to specific folder inside container), debug processes inside container, etc. Perhaps something got saved in the settings that is redirecting local File-Open to the remote-containers extension? Probably not, but the timing is mighty suspicious.\nInteresting. With 1.65.0 still running, snap installed the latest (1.66.0) again and ran it with: Ctrl-O still hung the 1.66.0 app. Apparently it isn't an extension.\nme is the same vscode is really disgusting\nyes, I ran into the same problem, vscode hangs when execute any open command, open folder or open a file\nSame for me on my Ubuntu 18.04 :-( Installed via tail of :\nAlso happening here, is there any known workaround at the moment to still open a folder? (Also seems to be related)\npoints to mismatched gtk library version loaded in the app process or others facing the issue, can you please follow the steps below and attach the generated profile\nThanks! Attached as requested above.\nIs the log cut off ? I don't see most the library loads that are required for startup.\nhmm, sorry, would be sufficient ? Because that is an only single output which I've got after execution. p.s. I am using managed PC and don't have permissions for :-(\nIn the past when I tried strace directly with snap packages and it would fail, hence I rely on the flag. Not sure if it works better now. No worries, I hope others hitting this issue might be able to provide the required trace. In the meantime, can you check if the issue exists with our debian package ?\nDebian package works well (with some strange behavior). Thanks for a temporary solution. Strange behavior -maybe because I've disconnected all my monitors, but Open Folder dialog opens in background, thus I have to bring it manually to the foreground (but it is better than hanging program anyway ;-))\nHere is the log I have with the instruction\nI am also having this problem with the latest update (though I never had the application hanging problem) on Ubuntu 18.04\nThanks for the log, it confirms the root issue. Snap bundle ships with a version of libgtk3 that gets loaded by the app for all gtk operations, can be seen with With Electron 14 there was a refactor in the file dialog implementation to support XDG portal specification and the gtk library that was trying to be loaded as part of this is not bundled within the snap and in which case will search the system library paths, if the library cannot be found then the old code path will be utilized that will use the alreaedy loaded gtk library in the process but incases where it is found from the system and gets loaded, if the version mismatches with the already loaded library in the process, it causes the above app freeze. This should be fixed in Electron to use the same calls as chromium ;l=79-82? to load the gtk library.\nso how should we solver it ?\nI've got the exact same problem. So, if this has been fixed in Electron 18 when can we expect an update to VSCode? Because I'm on 1.67.1 and the error still persists.\nThis issue is very annoying: ! Hope upgrade to electron18 as soon as possible\nThis issue started for me after installing . Not sure if that is helpful? After removing this package (and also running afterwards), the issue was gone.\nGot the newest version; but same error... Commit: Date: 2022-05-17T18:23:40.286Z Electron: 17.4.1 Chromium: 98.0.4758.141 16.13.0 V8: 9.8.177.13-electron.0 OS: Linux x64 4.15.0-177-generic snap\nUpdate to Electron 18 is gonna take a bit longer, I will try to get the fix into Electron 17 as well. The issue is not yet addressed in our insiders and stable.\nLatest version (from 2 days ago) still has the same bug (Ubuntu 18.04). Does anyone have a workaround? Any way to install an older version? EDIT: Downloading the Linux 64 .deb from solved the issue for me.\nif you're just looking for a workaround, try the terminal command: or . It work's for me at least. But for a more solution maybe 's helps (I haven't tried it myself).\nSame issue. Ubuntu 18.04. Channel: latest/stable: 2022-05-18 (97). thank you for the workaround. Removed new snap version and installed 1.65.0. Solved for me.\nI was having a similar issue with vscode 1.67.2, non-snap on Ubuntu 20.04. After I noticed the issue I tried an apt upgrade, which didn't fix it. Downgrading to 1.65.2 fixed it for me. I am using i3 for window management.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nVerification Steps: Install Ubuntu 18.04, images available from Install gtk3 <= 3.22, Install stable vscode from snap, Launch code and open the file explorer Verify application freezes Install insiders from snap, Launch code-insiders and open the file explorer Verify file explorer launches and application does not hang\ncan you help in verifying this since you have Ubuntu 18.04?\nHey , I can confirm that I get a dialog with the same setup and . Thanks! p.s. nevertheless, the strange behavior of dialog appearing in background remains the same as for previous solution using - but that is fine for now."}
{"_id":"q-en-vscode-2b5fc75b29804529424fbb521708f13d785c0b8fb756f4503b0d4b6264ac8d55","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.66.0 (x64) OS Version: 10.0. Build (Window 10) Steps to Reproduce: VS Code (or use the shortcut ) the first miliseconds after opening a new instance of VS Code, a \"Window 7\"-like window frame layout appears, before the usual VS Code look shows up (see the attached GIF). The GIF below first shows the expected behavior, opening version 1.65.0, followed by the current behavior of version 1.66.0. ! The two executables were downloaded from: and\nRefs\nthat's the reason I went back to 1.65, frames are so annoying.\nHi just a small update: I've downloaded the latest VS Code Insiders and the problem now only appears when launching VS Code for the first time, but no longer when opening new windows from a running instance of it.\nThis issue persists in the new latest version release, .\nThe issue persists in today's release, .\nSince the related bug in the Electron repo has a label, and today's VSCode uses , I believe we will be experiencing this bug for months.\nThe upstream Electron issue is at The issue was introduced in Electron v14.0.0-beta.5 and the commit that introduced the issue is The PR fixes the issue where frameless resizable windows show Windows 7 frames during startup, but there's a larger issue that is still unresolved even with that PR. I also searched through the thousands of to see whether they had anything to do with the issue, but I didn't find anything relevant even after reverting several suspicious commits. Running with Electron v14.0.0-beta.5 or newer on Windows 10 reproduces the issue. The PR fixes the issue for frameless resizable windows by calling at an earlier stage of program initialization. Assume the window is frameless and resizable. Then, the program sets to earlier in , and the following occurs: is not run, meaning is not removed from the style set in Line 62. This style parameter is passed in to a call at . is not run. Instead, the block above is run, which adds a style to the style passed into the call. , because of these styles being passed in, after the window is shown, we get a single message, and no messages after. NC in this case refers to non-client area, which is the frame. DWM stands for desktop windows manager. After adding a long of 30 seconds to the call above, I was able to use Spy++ to log messages to that window and see the following: I have also attached full logs to the end of this comment. For frameless windows that cannot resize, even if we call earlier, because the program would pass in , for the and calls above, the style would not be in the styles sent to the calls. As a result, what I see in Spy++ is being toggled three times, and several calls that result in messages being emitted. Using Spy++, I confirmed that the calls have the flag set. I have searched the Chromium and Electron codebases for and (as well as ) to see why the frame might be rendering more often when is not passed in. However, I haven't found anything within the codebase itself. Searching Windows documentation, I found the following: causes to be sent. Reference: However, I'm not sure whether this could explain the difference in behaviour between resizable and non-resizable windows. can pass in or to send to the window, which is why I searched for those two constants within the codebase. Reference: Also, Chromium handles by returning at , which I assume causes the message to occur. Also, Electron calls at . I noticed that if I don't call there, the Windows 7 frame doesn't show up, but then the Electron window doesn't render with the proper style (see ). I also noticed that when the Windows 7 frame shows up, it has the style of whatever was passed into that specific call. For example, if I add to the call, then for a second I see a Windows 7 frame with applied. I also tried creating a minimal project in VS to reproduce the issue but wasn't successful. The frame that rendered seemed to have a Windows 10 style rather than a Windows 7 one. I also tried registering a listener in for the messages, and it did seem to get hit, though I couldn't do anything with it. There's also the mysterious , but I wasn't able to apply it to resolve the issue.\nThe PR that linked above (partially based on ) should resolve the issue for VS Code Insiders and hence close this issue, but there's still more I'd like to investigate, especially for frameless non-resizable windows. I some lines to log the hwnd and style passed into each relevant and call. One example of the styles that show up: There's also another window being created really early, but I don't think it influences the issue. Good scenario (meaning no frame showing) with set to and staying Bad scenario with set to then Bad scenario with set to and staying The call with style causes the Windows 7 frame to render, and its style corresponds to is actually the same style just with , once again showing how mysterious the constant is. While working on , I also noticed that changing the order of the calls seemed to fix the issue for frameless non-resizable windows. In particular, on the Electron side, in , if I first called to pretend that the window is resizable, and then called (where ) after the call around line 365, then the issue is resolved even for frameless non-resizable windows. However, this workaround delays the call to hundreds of lines later, and I wouldn't be surprised if it would cause a regression elsewhere. Just like in the log file of the good case of the previous comment, only shows up once, rather than three times. For completeness, here are the logs with the styles for this scenario, excluding the logs: In this case, the style never comes up! I also tried delaying the call and calling instead of , but that doesn't work because it fails to pass to the call as well as the first call, both of which need for the frame not to render: can be used to change the non-client (NC) rendering policy and hence result in a call. However, I did not find any references to within the Electron and Chromium codebases that were actually called when running my minimal repro.\nlets continue the investigation in\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\n/verified"}
{"_id":"q-en-vscode-f8f4d4fadcf77ec7771ecbe82e4bd47f3cabc428bc5247ef865f43bd3bf9d48b","text":"Issue Type: <bBug</bSteps to Reproduce:  in a Markdown cell in Jupyter notebook. The monospace font used in Jupyter notebook () Markdown code blocks is not consistent with in the settings. ! This does not occur in a seperate Markdown () file. ! VS Code version: Code 1.66.0 (, 2022-03-30T05:50:14.623Z) OS version: Windows_NT x64 10.0. Restricted Mode: No <!-- generated by issue reporter --"}
{"_id":"q-en-vscode-1c40a0cf3bdc81282a1dcb9225acbbb5a49c262a89a85cc00736106739af6ed2","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.66.0 OS: MacOS Monterey 12.3, Darwin x64 21.4.0 Steps to Reproduce: new new markdown cell, name it for instance or will only show and have deleted the 1) or 1. written manually\ncc -- It looks like this is being sanitized out of md strings. A markdown cell containing will be sanitized to render as . If the cell contains it will render as expected, only removing the html tags. It seems to be happening at the following:"}
{"_id":"q-en-vscode-d87b1f134a1457a56104f60977bbf10d0624d85c64e1944a4860597c093b50ca","text":"<img width=\"776\" alt=\"image\" src=\"https://user-\"I don't have stable reproduce, but this happens pretty often, especially I'm moving lines around.\nI don't think this is a bug of injected text: When there are not decorations, there cannot be any injected text. This looks like there are some orphaned decorations.\nDo you use the setting? Do you still see this with latest & greatest insiders?\nI'm using the setting but haven't seen it for a while since I reported the issue.\nLeaving this open, lets find steps\nIs it possible that this is caused by the recursive issue?\nCould be the case! The fix is to use it like this:\nOh, so inside the runner?\nYes, because change events are only send out after returns. Thus, is not reentrant anymore and will be assigned correctly."}
{"_id":"q-en-vscode-3323a8e37cea518fbd1faa54bc9d73b17468d0d8546e20217ee031ef7a477737","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --In response to , I think this error should only be raised when a request actually fails. I have a network setup where my host operating system reports it doesn't have network connectivity, but it can actually reach the internet just fine through a configured proxy. I believe moving the online check to the would result in a more accurate error flow: Obligatory xkcd for my unique situation:\nAs mentioned by I expect shall provide right info whether VS Code can connect to internet or not. Is it possible to get this info from the error instead of ?"}
{"_id":"q-en-vscode-4e54b4ef7b2508ae61299ebd0fd6ca2f0260fc1296885eb3c6be3631b8d7bce8","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes VS Code Version: 1.67.0-insiders OS Version: macOS 12.3.1 Steps to Reproduce: the terminal scrollback to 30'000 for With version 1.65.0 this was not a problem, but there was no highlighting. It would be enough if I can disable the highlighting or disable the highlighting automatically if the history is longer than x lines. https://user-\nThat doesn't look or feel slow to me - can you elaborate on what the issue you're dealing with is?\nIs it Upstream or VSCode Core.\nHappens for me too, when you search then type abs quickly and it takes ~5-10 seconds for the letters to appear, same goes with erasing letters.\nSearch is fast now but matches info is not being shown -"}
{"_id":"q-en-vscode-42c718140a5afd77c70ebf3ba6693021fae671eb3282267e93419cf005940e62","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.66.2 (Universal) Commit: Date: 2022-04-11T07:49:20.994Z Electron: 17.2.0 Chromium: 98.0.4758.109 16.13.0 V8: 9.8.177.11-electron.0 OS: Darwin x64 20.5.0 Steps to Reproduce: push a commit to a repo that user don't have permission message modal will show to ask user if they want to create a fork repo button create a fork for you user don't want to creat a fork, that's why they close it. <img width=\"504\" alt=\"Screen Shot 2022-04-29 at 11 18 18 PM\" src=\"https://user-\"I check the code, looks like github extention only check  to return. What I suggest If you think this were resonable, I will make a PR for it"}
{"_id":"q-en-vscode-21f953b0304d2c4e027d873d43177cf90ac056d9ec62a1aad991a57a228bc95f","text":"Issue Type: <bBug</bI had a file with unusual line endings in it, and this happened: the dialog kept reappearing after pressing 'ignore' until i finally relented. When I tried to create a new file with weird line endings it seemed to work, maybe since I had the file opened while the file was being generated, every time the file changed the dialog was re-enqueued even though it was already open? VS Code version: Code 1.66.2 (Universal) (, 2022-04-11T07:49:20.994Z) OS version: Darwin arm64 21.1.0 Restricted Mode: No <details<summarySystem Info</summary </details<details<summaryExtensions (4)</summaryExtensionVersion ------ astro-vscode0.16.0 vscode-graphql0.4.4 vscode-docker1.22.0 svelte-vscode105.16.0 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\ncould you post some steps to verify?\nI could not reproduce this myself, I did the change by realizing that the code showing the dialog is re-entrant. Could you please verify the fix using our\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nI tried adding to the end of files and got the dialog, but I couldn't get the dialog to keep showing up on Stable.\nVerified by code review\nunfortunately it was a one-off conversion script from an old system that caused the issue which I don't have any longer, so I'm not sure about the exact steps to reproduce it! edit: oops wrong account, i'm threecreepio. :D"}
{"_id":"q-en-vscode-650bbc4c8e9ada8d78bf189114f4a66385d515c3edfd92bd9d50852dbcb8372e","text":"Issue Type: <bBug</bWhen closing the Python interactive window the Terminal panel shows itself. ! VS Code version: Code 1.67.0 (, 2022-05-04T12:06:02.889Z) OS version: Windows_NT x64 10.0. Restricted Mode: No <details<summaryExtensions (8)</summaryExtensionVersion ------ remotehub0.32.0 python2022.6.0 vscode-pylance2022.4.3 jupyter2022.4. jupyter-keymap1.0.0 jupyter-renderers1.0.6 remote-repositories0.6.0 markdown-all-in-one3.4.3 </details<!-- generated by issue reporter --\nwhat's the recommended approach to debug this? I'm not sure who is opening the terminal (but I know it's not notebook or iw as it's the editor tab close button)\nI debugged this, and determined that when the interactive window is closed, it updates a context key which hide a view container in the panel. this incorrectly triggers the panel to show."}
{"_id":"q-en-vscode-5b3e6aa093bee289b961b50fa528c6abdfd41e5785d6c94381dc5f507d5b48e5","text":"Open in two different tabs Make sure settings sync is OFF Change the theme in one tab Theme is not updated in other tab"}
{"_id":"q-en-vscode-39a4725a0c341d8ddc71b304e8fcde4286286d236c54a060ae2386aa2e4ecbb8","text":"The layout controls icons can be a bit of an eye sore when in the default state. One way to remedy this and to give a stronger indication of the state, we could swap icons based on their state. Would look something like this: https://user-\nthese icon are now in: - -"}
{"_id":"q-en-vscode-5aa9783b069eb44359fd10e4cde8f2d02541ab9a59347957e5972f51697ae279","text":"Many of these onDispose listeners get created:\nwe fixed this at one pt. !\nThey were in and\nhere's when we lost it\nwant to fix it since you're working on that PR?\nLet's hold off until it's done, this PR is complicated enough without merge problems\nit's a simple fix - literally adding to the conditional on line 218\nDid another review and haven't seen this issue since"}
{"_id":"q-en-vscode-b6cc1a8f1b83d4f46ea3c38401002ad8715f32f88cf4a12f6e1df31351ddbb79","text":"Our presubmit tests started to fail consistently since our test setup picked up code 1.67.0 yesterday. Our test is basically doing first, run written with from after the extension tests succeed, run a check on the consistency between our file and documentation/settings info. After running the extension tests with 1.67.0, it looks like the second step that involves reading sees a differently formatted Actually, it's not only a differently formatted , but it includes a new property (, , ...) at the end. The test passes if I make the extension tests to use 1.66.2. We can work around by arranging our presubmit test order or rewriting our test logic to handle this. But, we want to understand what is going on behind the scene and why this modification is necessary. Thanks!\nthis looks suspicious...\nThis is designed and expected. It is the metadata VS Code stores per extension while installing that can be used later for other features like updating etc., So I would recommend your extension to handle this.\nthis has started breaking our extension's tests without warning, and moreover, the way it writes to it's modifying the whole file (looks like whitespace changes).\nhow have you gone about working around this breaking change? I'm trying to get a release out myself and am suddenly blocked\nWhelp, something else went on that broke my tests, but definitely noticed this too and will need to be careful about not committing the reformatted .\nSorry about the caused breakage. I am not sure what is your test doing, but I think it's not a good idea to depend on the content of , because VS Code uses it to write metadata about the extension. We do not remove any existing data but add more data.\nWhat's the recommended way of dealing with the property when testing an extension with set to a version-controlled source directory? Should the changes be committed or do extension authors have to take care of reverting it (which probably means parsing and clearing the property).\nI'm currently reverting it. Would love to know if we should do otherwise.\nReopening this to investigate if there is something changed on VS Code while running extension tests\nI wanted to re-report this issue, because it was previously misunderstood and closed. Of course I won’t now this is open, but I already typed out reproduction steps. Perhaps they help. Steps to Reproduce: a git repository (this makes the issue more clear) a minimal VSCode extension with a test. The issue becomes more apparent if an existing extension name / publisher is used. .gitignore and commit we actually reproduce the bug. Run the extension tests Note that running tests has modified . Formatting is changed to use tabs, the final newline is stripped, and publisher metadata is to the file.\nTo verify: Create some hello world extension and open it's and change the and properties to an extension that already exists in Marketplace. Eg:Quit Code Insiders and Start from CLI with following arguments Make sure hello world extension's is not updated with property\nThis looks related to which didn't have anything to do with testing the extension, but only debugging it, and getting the same metadata written to an \"unsuspecting\" Will this fix handle this SO situation as well? The problem occurred very rarely so I couldn't give steps to reproduce, but I was only debugging and not running tests on the extension.\nYes\nI have learned more about my situation of getting this extra data when debugging an extension. It happens when the extension writes to the . I have a couple of extensions which allow the user to make settings that will be read and converted to commands. I files a separate issue:"}
{"_id":"q-en-vscode-7cd1717f0adb959e353b65a36de95c3e195be26971f33c61434b4ab72bb4e51d","text":"Request refactorings, and extract to a function at the top-level. Expected: no crash Actual: I'm not sure on the specific version, but in the last 9 days or so we started seeing the following crash:\nActually I found this while trying to repro a different bug. The one I was looking to repro had the following stack trace: <details<summaryOriginal Stack Trace </summary</details I think this actually only repros when working in an anonymous buffer in VS Code. When you ask for a refactoring on an anonymous buffer, we come back with the same buffer name However, VS Code pops up a notification that \"You cannot rename this element\" and creates a new buffer with the contents. This confuses TSServer a lot, and makes it fail over and over again.\nI initially read the last part of the issue title as “thanks, language service” and thought the sarcasm was appropriate.\nI may steal that in the future :D\ncan you transfer this issue to VSCode if appropriate?\nDid this just start happening with the latest insiders?\nPossible introduced with\nYeah, I believe it didn't repro on a 2 week old nightly, but I can't 100% confirm.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!"}
{"_id":"q-en-vscode-31ed2146dfe9e005e698767c65bc077a11108b4ce96372315ea4a4db037314fd","text":"<img width=\"628\" alt=\"image\" src=\"https://user-\"I often by accident click on the wrong button when adding a comment, imho this action is way too prominent there because it seems like the primary button to add the comment. //cc also because both buttons have the same color <img width=\"306\" alt=\"image\" src=\"https://user-\"\nThe \"Go to Review\" button was to address confusion on how to complete a review: We could add some convention where the left-most button gets more prominent styling. what do you think?\nCatching up after a week off.. I believe we should follow the same patterns used for dialogs and notifications where a single primary button is used alongside one or more secondary buttons. From what I understand, our the order the buttons in a dialog depends on the OS. Windows gets left-aligned primary buttons, vice versa for macOS. In this case, since the buttons are always right-aligned to the comment input on any platform, I think the primary button should be on the far right to align with the edge of the container. This looks to already be the case elsewhere in the extension: <img width=\"423\" alt=\"CleanShot 2022-04-18 at 09 09 03 src=\"https://user-\"\nThanks I want to update the style of this button to be a primary button, but there may be extensions that flip the order of importance for their buttons. I'll poll some other folks and see if it seems ok.\nVerified. I assume that Go to Review button is renamed to Start Review !\nThe Start Review button is a Go to Review button once you have already started a review. But since you verified the button colors this is still verified."}
{"_id":"q-en-vscode-a85f47f55ad566a74deefde2d74d96054a86722f3f488c54bba2b6176f68dc3e","text":"Issue Type: <bBug</bJust updated VSC and I can no longer execute jupyter notebook cells This happens when working both in local and connected to a SSH host, both with python and markdown cells I get the following error: Checked the output every extension and log, there's no errors or no clue as to what is happening. VS Code version: Code - Insiders 1.68.0-insider (, 2022-05-11T05:17:12.474Z) OS version: WindowsNT x64 10.0. Restricted Mode: No Remote OS version: Linux x64 5.4.0-1072-aws <details<summarySystem Info</summary </details<details<summaryExtensions (11)</summaryExtensionVersion ------ jupyter-keymap1.0.0 remote-containers0.235.0 remote-ssh0.80.0 remote-ssh-edit0.80.0 remote-wsl0.66.2 vscode-remote-extensionpack0.21.0 python2022.6.1 vscode-pylance2022.4.3 jupyter2022.5. jupyter-keymap1.0.0 jupyter-renderers1.0.6 (2 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nSnapshot of message: ! I would post snapshots of the outputs of extensions and logs, but there is no output whatsoever generating when attempting to run a cell\nI experienced the same after todays update to the same version. Using WSL2 Debian remote from Windows 10.\nSame here!\nSame.\nI was able to workaround this issue by using the command \"Notebook: Execute Cell\".\nsame\nVia the command palette? (Cmd/Ctrl+shift+p)?\nYep, that's how I first found it. You can also invoke the command directly from a cell with ctrl+alt+enter.\nSame problem here too, thanks in advance for addressing!\nI fixed it by determining the keyboard shortcut again. (ctrl + enter) However, the run icon still does not work. But the shortcut works.\nrun icon or regular shortcuts (shift+enter, ctrl+enter) did not work form me, only ctrl+alt+enter (which I wasn't even aware of until mentioned it) did anyways, seems to be fixed with PR, thanks for the quick response\nHow can I activate the PR in Insiders?\nfix was already released for the VSC insiders build, update your installation\nwhen I run jupyter notebooki type three words in it,the window can't response occur,then the vscode crushed,what should i dot to slove this problem"}
{"_id":"q-en-vscode-891709ebe6911a08cbd0113d977de7cafcb6b0c08f757110089c3849e75912df","text":"Started seeing this yesterday: !\nThis seems to be related to my theme, probably same root cause as\nRepro for this is to set the following settings and try find in terminal:"}
{"_id":"q-en-vscode-8d8b074b9955e3a9d23386d38c4ec0dd14914bea01585fefe1561ea3eb22ebc9","text":"insiders, new user data dir choose Light+ theme extensions viewlet, click on an extension Notice color contrast is off. The text can not be readable when an extension has selection and focus fyi !\nIt seems description color was introduced by Hence assigning to him.\nYeah looks like the new color isn't getting overridden in a focused/selected state. Thought I fixed this but I'll take a look"}
{"_id":"q-en-vscode-12292fc211916cd321219bbc6f256940d22c15810d3e151e13fd7518dacf8217","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.67.0 OS Version:mac os monterey 12.3.1 Steps to Reproduce: to settings (command+,) for markdown drop <img width=\"719\" alt=\"Screen Shot 2022-05-20 at 11 55 58 AM\" src=\"https://user-\"\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.67.2. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nConfirmed on latest insiders Version: 1.68.0-insider Commit: Date: 2022-05-16T05:30:27.563Z Electron: 17.4.3 Chromium: 98.0.4758.141 16.13.0 V8: 9.8.177.13-electron.0 OS: Darwin arm64 21.4.0 <img width=\"636\" alt=\"Screen Shot 2022-05-20 at 1 11 00 PM\" src=\"https://user-\"\nFixed as of\n<img width=\"967\" alt=\"image\" src=\"https://user-\"\nHey I see with the same fixes, already merged.\nYeah but it accidentally got reverted via This just brings your changes back in"}
{"_id":"q-en-vscode-f6b7db2b91f75d881ca41460d2ea6429809ecc06fc2926c4c3c0e80bf9c89b79","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --Love the new title control: ! My only problem is that I would often drag the window around by clicking and dragging the title bar. Since this takes up a large chunk of the title bar, I have less real estate to drag... but there's no reason the drag experience can't also apply to the area covered by the title control. Can the drag of the title control bubble up to the title bar so that dragging the window works?\nThere is a technical reason... Marking things as drag-region () swallows all events and you won't be able to click onto things anymore... I can try adding the style while dragging but it might not be possible <img width=\"113\" alt=\"Screenshot 2022-05-24 at 09 15 34\" src=\"https://user-\" that's more concerning here...\nYeah, I have tried to set the style during mouse down (with a delay) but that doesn't seem to have an effect Plan was to mark the element as app region drag when holding for 200ms\nrelated issue (which caused this)\nand more context - seems like a hard problem\nCan the top 1-2 pixels not be part of the drag element/target?\nyeah, it already should be like that, there is another PR in the works that increases the title bar height. with that merged, dragging be simpler again\nThe title bar height is now aligned with the height of editor tabs. This allows to drag the window when clicking on the space above/below the CC. https://user- That's the best I can do (and aligns with Slack etc). Set/unsetting the property while dragging isn't supported...\nTo verify: enable CC via make sure you can drag the app window with the regions above/below the CC (see )"}
{"_id":"q-en-vscode-1b5449b48164c461f9a59076b45138be4a7caf26057c3d68ff35f3907a1cf51d","text":"I'm building a custom Insiders to test one of the changes. The Linux localization smoke test seems to be failing consistently Edit: This result is likely a false positive. I think I forgot to update the branch."}
{"_id":"q-en-vscode-ccecce1279b85eeabc523270163c39ae7e199d7f3d202dbef86486254533bf0d","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --! Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.67.2 OS Version: Windows 11 10.0. Build Steps to Reproduce: the new folder button in the explorer panel; the new folder \"**\", without the quotation marks; text from red dialog box under the folder name is now bold, except for \"The name \".\nProbably because of markdown\nThe text box doesn't turn red or bold, so considering that a success\nThe text box should still be red saying it's an invalid file name on windows. Just the text should not become bold\nI think the box is supposed to be red\nthe box did not turn red for me with a folder name of\nWas this on windows?\nno it's working for me on windows. it wasn't working on mac\n** just happens to be a valid file name on macOS , thanks for verifying!"}
{"_id":"q-en-vscode-001fdf302ea99a38534c34c4593090a8e25eeddf7823e50a020fc5ed5f8413ae","text":"Testing This file has several headers which contain square brackets, and the contents that are delimited by square brackets get the link validation treatment ! One escape hatch for users is probably to set , but then you end up losing all reference link validation. Wondering if there is a less aggressive in-between option here, e.g. not validating link syntax in headers?\nI'm taking this.\nI've just submitted the PR.\nLinks in headers are valid. You can escape them if you don't want them to be considered link: I'm updating this issue to focus on links incorrectly be detected inside of other links, which your screenshots also shows\nI've just submitted another PR. Could you please check it?\nWith Insiders on , it still seems like nested links are getting detected: !\nLooks like the current fix doesn't handle cases where there are no spaces in the brackets. Will push a PR with a fix\nI'm still seeing on the latest Insiders on macOS."}
{"_id":"q-en-vscode-c3100f3c8ed39e8ff3216ef4ff13d5d441566d38fb2b310e40b7897a08b11b11","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: OS Version: Steps to Reproduce:\nOur C/C++ team has verified that our extension is correctly published to the marketplace with the correct binary -- our guess is that there is some VS Code or marketplace bug that is causing the wrong vsix to sometimes be installed.\nUsers are able to download/install our vsix manually from or so the bug appears to be with VS Code and/or the Remote WSL extension.\ncan you reproduce this problem? If yes can you file a new issue with reproducable steps? The issue is not 100% clear to us from this description. Thank you\nOur C/C++ team has not been able to reproduce the problem yet. I'm guessing there's some random or unusual requirement, maybe involving a previously cached installation when an upgrade from version 1.9.8 to 1.10.5 of our extension occurs. I'm not familiar with what sort of install logging would be important for users who repro it to get, which is why I was thinking the issue should be tracked by VS Code (or the Remote extension or marketplace) and not our team. I'm aware of 4 user reports -- there could be more reported against VS Code that I'm not aware of. We don't believe our extension's warning about the incorrect binary is the root issue, and just a symptom of some other system downloading the incorrect vsix for the user's OS (we ship 10 different vsix's).\nIf you can let us know what the remote OS is and what the TargetPlatform property is set to in the file that could help us identify the problem.\nyes getting more details here would be helpful. If you get a chance to repro the problem do let us know. Appreciate it!\nA user reports that our .vsixmanifest file shows TargetPlatform=\"win32-x64\" on a Linux-x64 remote machine (with Win-x64 host), which indicates that VS Code or the Remote extension or marketplace is somehow causing a download/install of the host platform's vsix instead of the remote one. I'm not familiar with what the next steps are in terms of diagnostics, but it seems like at least there should be an open VS Code (or Remote extension) bug to track this, since currently there are just 4 or so closed issues. See\nI'm the reporter of This is not the case for me as I have installed the C/C++ extension for the first time. I don't usually use VS Code but just tried it to reproduce another problem reported by a user. Also, as far as I remember, this is the first time for me to connect to a remote workspace using VS Code. Something that might be related is the version of VS Code. I think I have originally installed VS Code back in October 2021. After that, I hadn't updated the VS Code for a long time because I'm not using it. This time, I think I first tried to install the C/C++ extension with the old version of VS Code and failed. After that, I updated VS Code to the latest version from the main menu of the VS Code, which didn't solve the problem. I'm not sure if this is really related, but is it unsupported to update VS Code from an old version (8-month old) in a single update (i.e., without any intermediate versions)? For example, some migration steps to update the cache or database format would be skipped? We might be able to ask other reporters about their situations? (), (), (this issue): How's the situation? Did you experience the problem in the remote workspace such as in WSL? Have you recently updated VS Code? When have you updated VS Code previously?\nI am using the VS code(version 1.67) in Ubuntu 16.04 LTS (64 bit OS) which is hosted VM in Oracle VM virtual box. While trying to install C/C++ extension in the VS code, it was not proceeding and VS code suggested to download it from the website and install it via vsix. So, I have done the same. The installation of extension was successful but later during debugging, it has mentioned that wrong/incompatible binaries. please let me know if anymore information is required.\nI was not using remote workspace WSL. I am just trying to use VS code for client side code at my local VM.\ncc\nYour issue might be different? You installed the Linux x64 vsix, right? What is the TargetPlatform property set to in the .vscode[-insiders]/extensions/ms-vscode.cpptools-[version][os-description]/.vsixmanifest file? It should say \"linux-x64\". What version of the C/C++ extension are you installing? Our team is able to debug with the Linux x64 vsix so we don't know what could be going wrong yet.\nSorry that I am late to the party. It would be great if I can get summary of the issue? Does it happen to be reproducible?\nThis issue has been closed automatically because it needs more information and has not had recent activity. See also our guidelines. Happy Coding!\nI have been thinking that you have checked the linked issues, but let me comment in case you haven't noticed them. I have already described the details about my situation in my initial report If needed, I can just copy and paste the comment to here. Nevertheless, I can answer additional questions about my situation if any. We may also find some information by the OP of this issue in\nMay I know if you have downloaded the VSIX that is compatible to the platform you are using? !\nIt's not me. The above comment is from In my case, I initially tried to install the extension from a VS Code client in Win 10 opened for the remote workspace (Ubuntu 20.04 / WSL2), which fails with the message of incompatible binaries. After the suggestion in , I have downloaded the VSIX from and installed it in the remote, which was successful unlike the case of I'm not sure which version has downloaded. !\nThe problem is, as I have reported in , VS Code seems to wrongly download the VSIX of in Ubuntu 20.04/WSL2 instead of the VSIX of . I have tried several times VS Code's install button with different configurations for both VS Code and VS Code Insiders, but all failed in the same way.\nGot it. Thanks for update. Can you please share us with a gif/video of what you are doing so that we can understand more about the issue? Can you please share us with the following logs after reproducing the issue? F1 Open View... Log (Window) F1 Open View... Log (Remote)\nI tried to upload it to here but couldn't do it due to the size limit. I have instead uploaded it to YouTube: Log (Window) I cannot find such a view. Instead, I see and : Log (Remote Server) Log (Remote Extension Host)\nWhich version of VS Code are you using? Instead of downloading from the installed section, can you please try searching for the extension, say by typing c++ in the search input and install? Does it work if you uninstall the extension and install again?\nAs I have reported in , I have confirmed the behavior in and . Further details are found in As it does not seem to reproduce in every environment, I guess some other factors in my environment causes the problem.\nI now noticed these additional comments. I'll try them when I get back home.\nThis installs the correct version of the extension. What is the correct way to uninstall the extension from the remote? (a) As explained in , I have already tried many times removing the remote extension---by deleting the directory as suggested in ---and installing it again. But they didn't change anything.(b) When I select the C/C++ extension from the installed section and press the [Unintall] button (see below), the remote extension is still located under , but the local extension instead seems to be uninstalled. ! (c) Also, when I type in the search input, select the C/C++ extension, and press the [Uninstall] button (see below), the remote extension is still there, and the local extension is instead uninstalled. !\nSo it means you are getting incompatible C++ extension when you use the button ? In both above screenshots, pressing Uninstall will uninstall locally as the extension seems to be installed locally. So, please try pressing uninstall button and install the extension again?\nYes. I had already tried that many times. The result doesn't change; It again installs an incompatible version as far as I use the button .\nCan you please provide following logs while trying this? Enable log level to trace - F1 Set Log Level Share F1 Open View... Log (Remote) Share F1 Open View... Log (Window)\nHere, I attach the logs: This time, (1) I first attempted to install the extension by pressing . (2) Then reopen the extension page, press and reloaded the VS Code. (3) Then, I again attempted to install the extension from the button. In the zip archive, the following files are included: --- This includes the log of in step (1) described above. --- This includes the log of in step (1). --- This includes the log of in step (3). --- This includes the log of in step (3). --- This includes the log of in step (3).\nThanks for the logs and I think I know the root cause of this issue. It seems in your case the extension is being copied from local client (Windows) to remote (WSL) and hence the bits are being incompatible. This happens if VS Code cannot find a matching installed extension in the Marketplace. In this case, VS Code falls back to copy the extension. Fixed it as follows: Before using the fallback to copy Try again to get the matching extension from the Marketplace In the fallback check if the local extension is compatible with the remote platform. This fix does not covers if you have manually installed an extension that does not exist in the Marketplace and is platform specific.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!"}
{"_id":"q-en-vscode-38103ad1ea0dbd54e281500cd8cff9c5b69e071f6fc1a434c494c51155640f91","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.68.0 OS Version: Win10 21H1 .1706 Steps to Reproduce: ! This is caused by When L122 , the element may have been created in the tree before. Thus, may be more than once and causes icon style conflict. ! A PR will be linked. Idea from After fix: !"}
{"_id":"q-en-vscode-307c4f65c291f0f0e695b35a3272e7b3199f95276205950a41909753dcb618e7","text":"!\nworks in OSS !\nalso works for me in insider's\nworks for that exact task for me !\nIt's because of this: Here's the I was using:\nThis works but it adds an extra line because of initialText so it doesn't look quite right: Before: ! After: !"}
{"_id":"q-en-vscode-79021f7d920cfc0c4f26c624f0ce2f6c1ebf3d03ae7503a974e4b3fd30f631ab","text":"Issue Type: <bBug</bWhen I drill down into the git menu and choose the option to , I really don't need the same confirmation dialog as when I just hit the top-level check mark in the source control pane. The dialog tells me ; this is an incorrect warning at this stage, since I clearly invoked it by choosing . I certainly don't want to choose the option in response to this warning, since it's a good warning to have when I accidentally hit the checkmark without staging any changes. VS Code version: Code 1.67.0 (, 2022-05-04T12:14:47.100Z) OS version: Darwin x64 19.6.0 Restricted Mode: No <!-- generated by issue reporter --\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.68.1. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!"}
{"_id":"q-en-vscode-136c2424f906be6ab94f8e1c9e61719be3c7684ae7b5f8f21f3f07b8169c35c3","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.68.1 OS Version: Windows 10 Steps to Reproduce: When I open a terminal in VSCode the following warning is logged within a few seconds: . I take note of that this issue is not a duplicate of.\nA warning as in on the window output channel or devtools console? Do you use powerlevel10k or another complex prompt by chance?\nIt's in the devtools console. Not using any complex prompt. Just plain CMD.\nOh OK, that's unexpected if it's cmd, thanks for the info\nwe should probably do something like here"}
{"_id":"q-en-vscode-e45feea09e96ace341e3bc0ed905ee2f2e58ea75aa0d8bd1556078e4e42b056f","text":"Issue Type: <bBug</bAfter upgrading to 1.68, text selection in the integrated terminal does not show selected text. This is purely a visual glitch, as I can still copy what I had selected on the terminal. My colleague has the same version but does not have this issue, so I'm wondering if it's an issue introduced through my update. VS Code version: Code 1.68.1 (, 2022-06-14T12:48:58.283Z) OS version: WindowsNT x64 10.0. Restricted Mode: No <details<summarySystem Info</summary </details<details<summaryExtensions (42)</summaryExtensionVersion ------ project-manager12.6.0 github-markdown-preview0.3.0 markdown-checkbox0.3.2 markdown-emoji0.2.1 markdown-footnotes0.0.7 markdown-mermaid1.14.0 markdown-preview-github-styles1.0.1 markdown-yaml-preamble0.1.0 keymapceditor-vsc0.7.7 gitignore0.7.0 doxdocgen1.4.0 transformer1.12.1 git-extension-pack0.1.3 githistory0.6.19 gitlens12.1.1 EditorConfig0.16.4 align-mode0.0.21 better-cpp-syntax1.15.18 cmake-language-support-vscode0.0.4 theme-karyfoundation-themes26.2.1 vscode-docker1.22.0 vscode-dotnet-runtime1.5.0 python2022.8.0 vscode-pylance2022.6.20 jupyter2022.5. jupyter-keymap1.0.0 jupyter-renderers1.0.8 remote-containers0.238.2 remote-ssh0.82.1 remote-ssh-edit0.80.0 remote-wsl0.66.3 vscode-remote-extensionpack0.21.0 cmake-tools1.11.26 cpptools1.10.7 cpptools-extension-pack1.2.0 makefile-tools0.5.0 indent-rainbow8.3.1 vscode-pbtxt0.0.4 cmake0.0.17 vim1.22.2 clang-format1.9.0 vscode-open-in-github1.3.6 (5 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nSame issue, I'm using theme. When I switch to a different theme the selection is visible. Looks like a theme issue.\nCan you reproduce the issue in Insider's? I see selection in Insider's when using solarized light, but it does seem pretty light !\nI believe we're implementing solarized light correctly here and that's how it's meant to work. If you're not a fan of the low contrast selection you can change it with this setting: Otherwise change to a theme that you prefer. Let me know if I misunderstood and the selection is actually not visible (not just subtle).\nI was using the theme for a while and I was used to the low contrast on selection. It was quite ok on my screen TBH. Now there's no visible selection. Here I've selected \"Counting objects\". I've actually pasted it from the terminal instead of typing to be sure. ! I don't think this issue is ready for closing.\nDo you have any settings set and does changing help?\nI strive to work on defaults. No settings in Setting to true causes selected letters to become slightly lighter, but no background. \"si\" selected ! Not sure if it has anything to do with it, but the issue seems to have started after the recent update (via PPA on an ubuntu flavor). After updating I didn't restart immediately, but attempted to open an image and everything crashed. Restarted after that and noticed selection not working a bit later.\nSome more things to try: Does closing all windows and running fix it? If not, can you share your full file? Running the Generate color theme from current settings command will let us see if there's something weird going on with the theme\nConfirming this is only happening in Solarized Light\nDisabling GPU doesn't help <details</summary{ \"window.menuBarVisibility\": \"toggle\", \"editor.minimap.enabled\": false, \"[javascript]\": { \"editor.defaultFormatter\": \"vscode.typescript-language-features\" }, \"javascript.updateImportsOnFileMove.enabled\": \"always\", \"[json]\": { \"editor.defaultFormatter\": \"\" }, \"[typescript]\": { \"editor.defaultFormatter\": \"vscode.typescript-language-features\" }, \"\": \"stuff=\", \"explorer.confirmDelete\": false, \"explorer.confirmDragAndDrop\": false, \"[html]\": { \"editor.defaultFormatter\": \"\" }, \"[jsonc]\": { \"editor.defaultFormatter\": \"\" }, \"workbench.colorTheme\": \"Solarized Light\", \"editor.inlineSuggest.enabled\": true, \"editor.renderWhitespace\": \"none\", \"[css]\": { \"editor.defaultFormatter\": \"esbenp.prettier-vscode\" } } </details<details<summarycolor theme from current settings</summary</details\nI can't reproduce based on those settings so I'm a bit stumped on what to do with this one\nSo I checked the terminal background on the screenshot and it's which is exactly the same color as the default selection color. the dump I pasted above had in it, which seems odd. But since you didn't raise it, I'm assuming it's supposed to mean the actual background is the one of the panel inside which the terminal is working. What I also noticed about the terminal panel is it's no longer giving me the option to move it to the bottom of the window, which is odd, but most of all when I right-click, it no longer says \"move terminal right\" as an option, but \"move sidebar\". I checked, and means I will no longer see my selection. Conclusion is what ends up being the color behind terminal text. The theme should probably set terminal background or change the selection color as a fix. I wonder what else breaks because terminal is now a sidebar not something else.\nAh ok, thanks for looking into it. Because of this, setting the terminal background explicitly is a good solution for now. I doubt it happens in many other themes since contrast is usually pretty high but I'll check Solarized Dark as well when fixing."}
{"_id":"q-en-vscode-a53e0158841d8e6006a582f1cd47dcc570150faa1e7ec27d8a48d1e06762522d","text":"Version: 1.69.0-insider (user setup) Commit: Date: 2022-06-22T05:17:09.661Z Electron: 18.3.4 Chromium: 100.0.4896.160 16.13.2 V8: 10.0.139.17-electron.0 OS: Windows_NT x64 10.0. This is a follow up of Steps to Reproduce: the dnd sample: to the of some logs in : the extension multiple files to the dnd sample view, the uris in the data transfer are not split by line separators. // cc\nAre you dragging from the VS Code file explorer view, or your operating system's file explorer?\nSorry forgot to mention, it's from VS Code file explorer\nCan you check the actual value, not just the value printed by I debugged this and oddly while just prints: The actual value of is: In this case each file name is separated by a pair\nTested in: Version: 1.69.0-insider (user setup) Commit: Date: 2022-06-23T05:16:43.681Z Electron: 18.3.4 Chromium: 100.0.4896.160 16.13.2 V8: 10.0.139.17-electron.0 OS: Windows_NT x64 10.0. It's still stored in one line: https://user- Is there anything I missed?\nThanks. Only happens on windows for me. Taking a look now\nI don't see us setting anywhere and suspect chrome is generating it automatically. See in for how this may happen Also we are possibly hitting\nNice find. I couldn't find us setting anywhere either.\nVerified on Windows and Linux in 1.69.0-insider (user setup) Commit: , 022-07-01T05:17:04.343Z"}
{"_id":"q-en-vscode-cab2e0ef1e1bf46cc1e3b8532da5219401213473fc68bf8dcc701b90332e0dd9","text":"I think we should understand why execution time seems to get slower and slower the more cells there are. Check out this video, it looks as if things get slower and slower: ! Originally posted by in\nThis test is running the \"Run All\" command, not the \"run one cell\" command..."}
{"_id":"q-en-vscode-3f21a9150845d74478db2abca52a0d1eaf3ce0a392e655e2fc18e02849775351","text":"! Version: 1.69.0-insider (user setup) Commit: Date: 2022-06-23T05:16:43.681Z Electron: 18.3.4 Chromium: 100.0.4896.160 16.13.2 V8: 10.0.139.17-electron.0 OS: Windows_NT x64 10.0.\nIt's actually allowed to have no scope (for the default fore/background), but the schema is too strict"}
{"_id":"q-en-vscode-1df0e4569c894f4c77d6f955a5e474cfe3e80e866af92c2e2c2b5cf0cf179556","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: No (the issue is produced using CMake Tools extension) <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: Version: 1.68.1 (user setup) Commit: Date: 2022-06-14T12:48:58.283Z Electron: 17.4.7 Chromium: 98.0.4758.141 16.13.0 V8: 9.8.177.13-electron.0 OS: WindowsNT x64 6.3.9600 OS Version: Windows 8.1 When using the CMake Tools extension, the build output is shown in the \"Output\" tab. And when the build is going fast (when there are many build log lines per second), the Output doesn't update until the end of the build. See this video from https://user- The build output is not realtime. CMake Tools extension a project that builds fast (less than 200ms per build log line) main.c: Note: Build this CMake project using MSVC is faster and is more likely to reproduce the issue than building with clang/mingw. After severals tests, I've found that: The VSCode log file (inside is constantly updated as expected, this can be checked with bash using The code that prevent the actual log to be printed in the Output pane is here: This code uses from here (with a default delay of 300ms): Which uses from here: And if I read this correctly, this means that: Due to IPC between CMake Tools extension and the main Code thread, for every log line, is called, and every call to that function will reset the Delayer timer of 200ms. This means that as long as there are log lines every at most 200ms, the output pane won't be updated. Which cause the \"hang\" as seen in the above video. I guess that replacing the from to a would solve that issue ?\nHere are videos to compare the modification of using instead of for the variable in . Before modification (VSCode 1.68.1 unmodified): https://user- After modification (replaced the declaration of to , replaced call from to and removed call to ): https://user- The CPU usage is a bit higher, but the log is shown in realtime (the CPU is a intel i5 3230m from 2013). Closing the Output pane while building takes less CPU (around 48% instead of 53% (when not recording), when not building I'm at 6%).\nThanks for debugging it and coming up with the root cause. Your analysis is correct. But changing this to Throttler is not the right. We would still want to delay/debounce the updates if they do not need to be real time like log outputs. But for the output channels created by outputs we shall update the content immediately. Created a PR to fix it when immediate updates are needed like in this case."}
{"_id":"q-en-vscode-d461dd3a89cd01e17e97b93985fedfd5e8dcc0e5d85517d78c011a3b174dd6c8","text":"Testing If there are no changes, we should not be able to create a new edit session. In this particular case, I would expect that an information message is being shown, and that the existing edit session is not being overwritten."}
{"_id":"q-en-vscode-f554f35bf02f3dcdc6cae604424465893bced95c4796f6a11efa8bc8d3aa3ea4","text":"Testing Have one profile with GitLens installed and other without. Notice how walkthrough shows up in Welcome page on profile without GitLens <img width=\"1653\" alt=\"image\" src=\"https://user-\"\nIt seems welcome page is not updating when extensions are removed. This has to be fixed in walk throughs / welcome page component. Not sure who owns this now. Can you please assign it to the corresponding owner?\nIt appears we are handling it correctly here\nI see is being invoked for gitlens extension when switching profiles, but I do not know why the walk through page is still showing it. I do not know the implementation of it. Can you please investigate from walk through component perspective first to see why it is still being shown?\nI disabled gitlens and used the button to reload the window. The walkthrough is no longer there. is that expected?\n!\nI changed profiles to the one with Gitlens still enabled and there's still no walkthrough\nHmmm I cannot repro this. Can you provide a gif? If I have the walkthrough open and reload my window the walkthrough page doesn't dissapear\nhttps://user-"}
{"_id":"q-en-vscode-4e74a7ca116f858aae6fa919a010473fa1bfc15bd5a1c4138497f655a4acb7da","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: =1.66.0 OS Version: iOS 14.3 Steps to Reproduce: Safari on iOS 14.3 to any vscode server (e.g. , or ) The Safari developer tools console shows the following error:\n. I see references to in the following commits: (soft reference) Thoughts?\nAs this has been supported since iOS 15, do we want to change this?\nThe changes introduced in completely break vs code for all devices below iOS/iPadOS 15, the other occurrences of don't seem to have any negative impact according to my testing (anything below 1.66.0 works). The fix I proposed in is a more elegant solution that supports all browsers and does not break anything.\nWe are generally only aiming for latest and greatest browser version. The use of seems small enough to justify a special case but we shouldn't target a lower version nor is there any guarantee that other things won't break, e.g \"fixing\" this might just reveal another bunch of blocker etc pp.\nSounds good, forwarding to to evaluate the PR."}
{"_id":"q-en-vscode-7e4d9c2d2cf7847cb935c99d0847c930a653b082b40d16a0de9e81eac543eecd","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: N/A <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.69.0-insider OS Version: Windows 10 Steps to Reproduce: and install from VSIX the test-provider-sample from - this just sets instead of adding test cases. Testing panel and open outcome: shows \"expanded arrow\" but error message doesn't appear. Error appears in Developer Tools ! ! bootstrap- [uncaught exception]: TypeError: Cannot read properties of undefined (reading 'expand') onUnexpectedError @ bootstrap- (anonymous) @ bootstrap- emit @ node:events:390 window.onerror @ node:electron/js2c/rendererinit:73 platform1.globals.onerror @ bootstrap- TypeError: Cannot read properties of undefined (reading 'expand') at Object.getId (vscode-file://vscode-) at vscode-file://vscode- at (<anonymous) at IndexTreeModel.spliceSimple (vscode-file://vscode-) at IndexTreeModel.spliceSmart (vscode-file://vscode-) at (vscode-file://vscode-) at ObjectTreeModel.setChildren (vscode-file://vscode-) at ObjectTreeModel.setChildren (vscode-file://vscode-) at WorkbenchObjectTree.setChildren (vscode-file://vscode-) at NodeChangeList.applyTo (vscode-file://vscode-) at HierarchicalByLocationProjection.applyTo (vscode-file://vscode-) at TestingExplorerViewModel.applyProjectionChanges (vscode-file://vscode-) at (vscode-file://vscode-) at RunOnceScheduler.doRun (vscode-file://vscode-) at RunOnceScheduler.onTimeout (vscode-file://vscode-) onUnexpectedError @ bootstrap- (anonymous) @ bootstrap- emit @ node:events:390 window.onerror @ node:electron/js2c/rendererinit:73 platform1.globals.onerror @ Uncaught TypeError: Cannot read properties of undefined (reading 'expand') at Object.getId () at at (<anonymous) at IndexTreeModel.spliceSimple () at IndexTreeModel.spliceSmart () at () at ObjectTreeModel.setChildren () at ObjectTreeModel.setChildren () at WorkbenchObjectTree.setChildren () at NodeChangeList.applyTo () at HierarchicalByLocationProjection.applyTo () at TestingExplorerViewModel.applyProjectionChanges () at () at RunOnceScheduler.doRun () at RunOnceScheduler.onTimeout ()\nI submitted a PR\n/assign"}
{"_id":"q-en-vscode-edcdc425ddc4f2dbb17c58a947958b2506c05e5bbbb6a13268340b19d51eea6d","text":"<img width=\"368\" alt=\"CleanShot 2022-07-18 at 08 28 16 src=\"https://user-\"\"Configure\" and \"Manage\" are really similar to each other and I had a hard time knowing which one opened what, which caused me to open the wrong one. \"Configure\" opens the settings UI with a query, and \"Manage\" opens a custom editor. My proposal would be to change \"Configure Workspace Trust\" to \"Workspace Trust Settings\" to make it explicit.\nI'd actually like to tweak your proposal to \"Manage Trusted Workspaces\" and \"Configure Workspace Trust Settings\""}
{"_id":"q-en-vscode-c7e16f6e5c2bd51e334bf37702a1fae28e3bc2f713cd406dc3eaa03b4886c6ed","text":"Repro: Using Open a Remote - WSL folder !\nThis would probably not be limited to the title, what about the hover over the tab?\nThe separators seem to be flipped: !\nI wonder if could have regressed this, is this a recent regression?\nI cannot reproduce, so this is some setting you have or some configuration most likely.\nIt might be remnants of the formatted caching bug?\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!"}
{"_id":"q-en-vscode-dccf5681783e4347602d9ddbbd53edfeb961647c2b5924b1e3a466986499d9a8","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes Steps to Reproduce: The documentation has a spelling mistake LINK: -Value of an environment variable on the host machine (in this case, called VARIABLENAME). Unset variables are left blank. To for example, this would set a variable to your local home folder on Linux / macOS or the user folder on Windows:\n/ , can you please point me to the GitHub page that corresponds to this page.\nThanks for pointing this out. It looks like we fixed the typo () but have not yet deployed the latest changes.\nYou are welcome, ! I was trying to fix the issue for Hacktoberfest 2022, but I was unable to figure out in which repo this file was located. Just out of curiosity I would like to know as to where the source file for the page with the typo reside in GitHub?\nThanks for looping me in! Opened a PR in is the repo for"}
{"_id":"q-en-vscode-1be7c11f61fe265e1bbb7a1b9915355cf7badcc4ad0e188eb02f5c38f065f2b8","text":"The focus on a search result shouldn't change when removing an element that isn't affected by the removal. ]() Should be like this: ! Likely due to changes in"}
{"_id":"q-en-vscode-faac39c896e32727c630ce119fcc94bb07a48d496c145cc231f36e3be41ef7d5","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.68.1-dev OS Version: web Steps to Reproduce: a in to get editor gutter decorations for and changed lines. a file with a single line change e.g. a word in line 5. The line will show an indicator in the editor gutter for the changed line, which when clicked will open a Quick Diff peek view. a on the same line. The editor gutter will display the comment gutter icon (diamond inside rectangle), which when clicked will expand or collapse the comment on that line, and hide the indicator for the changed content in the same line, preventing the user from noticing that there is a change or opening Quick Diff to inspect it.\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.69.2. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nWith multi-line comments () the impact is likely higher. As the user is likely to make changes in close proximity to comments to address issues raised, this may disproportionately limit the utility of Quick Diff. cc\nI'm not seeing them overlap: ! Am I looking at the correct thing?\nThat's the correct thing and in your screenshot it appears to be fine. Not sure why we are seeing the comment indicators draw over the change indicators, but we are also on a slightly older version and I am not sure if there are any other local adjustments in this space that I was not aware of. I will check this again on plain VS Code, sorry for the noise.\nIn 1.70.0-insider, I see this: for a multi-line change across the comment range: <img width=\"688\" alt=\"Screen Shot 2022-07-26 at 16 06 34\" src=\"https://user-\"- for a single-line change on the same line as the comment range: <img width=\"715\" alt=\"Screen Shot 2022-07-26 at 16 07 07\" src=\"https://user-\"This is for a simple Git repository, with a small test extension populating the comments. Caveat might be that it's \"read-only\" in the sense that I don't set , not sure if that causes problems? I see and draw in the same location and the comment thread wins because it has .\nI suspect that not having the range provider is causing this. Will try to fix for July, but will likely move to August.\nI tried setting a range provider, here are my observations: The docs for are wrong: When not provided, users cannot leave any comments—I think this is actually fine—and the gutter indicators overlap, as described. When setting a that returns for all documents, the behavior is the same as above. When setting a that returns , i.e. an empty range, the gutter indicators seem to draw correctly. However this also unexpectedly allows the user to start a discussion on line 1. For our extension, we don't want to support adding new comments on any lines, but otherwise show and allow replying to comments populated by the extension via the API. The reason is that our extension focuses primarily on supporting the author of a change / pull request in addressing the comments left by code reviewers in an external code review tool, i.e. they will almost never start a comment thread from their side. Allowing commenting on all lines of the file currently appears to draw an extremely distracting indicator through the entire gutter () which we strongly wish to avoid without requiring usage of Zen Mode. If the visuals of this were more subtle () we would not mind allowing users to comment on the entire file, even though it remains a rather rare use case for our user base.\nThe problem described in the original issue is a bug, which I will fix, You shouldn't need to have a commenting range provider at all. It sounds like there is also a doc bug, and an empty range bug, which I'll also look at. Regarding Zen mode: since the last update commenting is off by default in Zen Mode, so you shouldn't see any comments there. You can always toggle it back on with the command Comments: Toggle Editor Commenting.\nThe linked PR will fix the hidden decoration bug and the doc bug. I'm hesitant to make any changes for the range, as this may be intentional behavior or behavior that some extensions rely on.\nI checked in Insiders today and this is now looking good for us, without setting . Thank you so much for the speedy fix!\nmarking as verified by author."}
{"_id":"q-en-vscode-ab8f8a01304c1e789eadff48c6c398f5da96d90877b714987774e9df98990e4f","text":"Remove Manual Folding Range doesn't work Originally posted by in select a range in the editor invoke . Selection turns into a folding range put the cursor at the beginning of the collapsed line invoke . Should remove the manual folding range\nIt works when the cursor is at the second column or to the right of that, but not when it's right at the beginning of the line, which is exactly where it is placed after running the create manual folding range command."}
{"_id":"q-en-vscode-b416a3a4e8687848c3a4be1900c209ff2cbf372c8a05042fa782816f3209f9bf","text":"Testing open the emoji picker (ctrl+cmd+space) drag an emoji into the code editor observe that nothing happens from adding a breakpoint, I can observe that  is never invoked I'm not sure what's wrong, the same steps work fine when loading the editor by itself in a browser using . https://user-\nSeems to require holding down shift at the moment even though there is no overlay saying this\nI cannot drop an emoji into the monaco playground and also not into the TS playground. It also does not work in However it works on stable and stable VS Code when holding Shift as Matt suggested. I also do not see the drop feedback in insiders, so this looks like a regression to me.\nSorry, I think I got confused because in my insiders I have configured , so I do not see the drop indicator. I see no difference in behaviour between VSCode stable and insiders, so it is not a regression as far as I can tell. I can confirm that pressing makes this work\nOn our playground it does not work either but I am not even seeing the drop indicator: !\nThe editor at is around 6 months old, so there it doesn't work at all, I acknowledge that. In my original screen recording I was running the editor from sources. I am quite certain this used to work at some time (when Matt first did the PR at ). I'll try bisecting to see if that points to anything. The initial bisect points to\npoints to\nIf I comment out the following lines, then dropping text from TextEdit works, otherwise it doesn't: https://user- https://user-\nThe safest fix seems to be reverting the fix for by commenting out the block of code. Will push that for July and then investigate a proper solution that also in August\nCan't seem to get this to work with the Windows emoji picker and the drag/drop extension sample. Does this only work on mac?\nI only tested on Mac and don't know if this is supported on windows. Can you drag emoji into other apps, such as notepad?\nDoesn't look like it. Will close and leave for someone on mac to verify then."}
{"_id":"q-en-vscode-c6d14084c993895fbf5ee18922a5cf198b5fad9ad4e6101471f1e710643c17b2","text":"Testing In Remote-WSL + bash + subshell, if I run and then use the command decoration to rerun the command, it instead runs . !\nThis is also the reason may be replaced by ; we're sending the alias resolved command line instead of the raw command line as input.\nWe should also verify this doesn't happen with aliases in zsh\nfyi verifier, reproduces in normal in a non-subshell\nto reproduce: in a terminal to see the ones you've defined the part to the left of the command decoration menu, the command's alias (part to the right of ) is the input\nit does not\nthe alias value is also displayed in the quickpick atm\nI thought of an even better idea to what we talked about, we could use the old history method to supplement our current better method; if there are no problematic histcontrol settings and the latest history command changed between precious and now, assume it was an alias and use that?\ncould you verify this one when you get a chance?\nstill seems to repro in Remote-WSL ! Version: 1.71.0-insider (user setup) Commit: Date: 2022-08-23T10:56:12.685Z Electron: 19.0.12 Chromium: 102.0.5005.167 16.14.2 V8: 10.2.154.15-electron.0 OS: Windows_NT x64 10.0. Sandboxed: Yes\nDoes show anything? My Ubuntu wsl shows in which case we cannot use that history method which is a known edge case that I don't think we can handle currently and is not worth tracking imo.\nYeah the output is . Is there a value I can change that to which would allow me to verify the fix?\nit can be set to the empty string or - set this in your which should run before shell integration\n<img width=\"278\" alt=\"image\" src=\"https://user-\""}
{"_id":"q-en-vscode-70835ae054def748002807da15840af3c0f017b88d7c9a9e525d0801687491d1","text":"Type: <bBug</bTesting down in a TypeScript function ! Note that, In the screenshot, is actually at line 1344, the gutter doesn't have enough space for 4 numbers. Also, the visual separation between sticky lines and normal lines looks too subtle for me. VS Code version: Code - Insiders 1.70.0-insider (, 2022-07-29T10:47:14.408Z) OS version: WindowsNT x64 10.0. Modes: Remote OS version: Linux x64 5.10.102.1-microsoft-standard-WSL2 <details<summarySystem Info</summary </details<details<summaryExtensions (18)</summaryExtensionVersion ------ Bookmarks13.3.1 remotehub0.39. remote-containers0.243.0 remote-ssh0.85. remote-ssh-edit0.80.0 remote-wsl0.66.3 remote-repositories0.16.0 material-icon-theme4.19.0 material-theme3.15.2 vscode-eslint2.2.6 gitlens12.1.2 prettier-vscode9.5.0 copilot-nightly1.36.6347 js-debug-nightly2022.7.2717 abracadabra6.14.4 vscode-jest4.6.0 prisma4.1.0 code-spell-checker2.3.4 </details<!-- generated by issue reporter --"}
{"_id":"q-en-vscode-ffcc3d0bc7628edb583a43939fe66354490ff9fc9aca8f64353551569fef7ed3","text":"Does this issue occur when all extensions are disabled?: No VS Code Version: 1.70.0 OS Version: Debian 10.8 Steps to Reproduce: a zsh shell with a framework like zim installed shell integration zsh Observe: This breaks all my integrations and I lose history, etc... Usually because default settings do things like\nIn WSL: Ubuntu-22.04, z4h (zsh for humans) is telling me my ZDOTDIR has been set to /tmp/vscode-zsh . I updated to VSCode 1.70.0 and noticed this. It's a shame because the shell integration worked so well in 1.69.x.\nI'll check this out tomorrow but it seems like a candidate for the recovery release. In the meantime you can disable shell integration in your settings and install it manually with these instructions:\nThis appears to be the only change around ZDOTDIR since 1.69:\nI will echo this has also broken my shell configuration to load an SDK for a certain VSCode workspace because it hijacks the early on and doesn't provide any hooks to work around that.\nI can reproduce this when using a non-login shell. That's because we reset the in the login rc file which only runs for login shells. Unrelated to the history change\nThanks.\nwe're planning on releasing the fix in the next couple of days\nit's already available in the\nto verify, set your to something in your . remove the login args from the profile. and verify that it's whatever you set.\nAs of 1.70.1, macOS 12.5, this is still broken for me:\nlooks like we should try to detect if this is used and disable shell integration\nFYI, I've tested on macOS, WSL/Win 11, and ChromeOS. On WSL and ChromeOS, ZDOTDIR is set to /tmp/vscode-zsh, but otherwise it's the same as on macOS. When I turn off Shell Integration in the settings and open a new Terminal instance, shell integration appears to work. So maybe has the right idea. And as I said earlier in the thread, this worked without issue on 1.69.x.\nsetting to worked for now\nI think zsh4humans uses powerlevel10k which disables shell integration (or will disable it in the next version?). I think to fix this we actually need to add or something similar similar inside this if statement: Alternatively, should this line be moved above the check? To test this without plugins, unsetting inside your rc should repro.\nI don't think that line can be moved above or will override the temporary we have set for vscode but think as you said, we need to also add it before returning when it's been disabled\nOr can we remove this: Either there or at the end of whatever injected zsh script that runs last.\nwe could remove it there for non-login shells. for login shells, it gets reset here and probably can remove line 6\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nI'm trying to set in my .zshrc, but it always ends up reset to my home directory, same behavior in stable/insiders. But I'm not sure whether there is something in my zsh config that does that.\nthe gets set in not your\nTried that, and it seems like ZDOTDIR ends up set to a tmp folder <img width=\"636\" alt=\"image\" src=\"https://user-\"\nnot sure then - it works for me - it's the only thing in my file though\nI encountered the same problem as I used the following version of Insider, which seems to be fixed in commit , but the problem is not fixed. `\nhey all, I've also run into this, not 100% sure though if it's exactly the same or \"just\" related. I use remote container extension with a Debian based container and ZDOTDIR env isn't respected as long as shellIntegration is enabled. Actually I was using regular code but tried insiders after reading the comments here, without any success. let me know if I should provide more details - I tried so many combinations I can't really exactly remember right now, but of course I'm happy to try something specific if it helps.\nIs this fix supposed to be live to the stable release now? If so, I am still having this issue when using zsh4humans.\nIt's not yet in stable. it should be fixed in insider's now though by\nIs this fix scheduled for the next stable version? and when would that be?\nyes, it will be next week"}
{"_id":"q-en-vscode-eb38c5540de33ba66f351eeb068e2ac5ad3de6f4a176358feb3564d2e10790bf","text":"Repro: folder test task, wait to finish another terminal test task again <img width=\"1344\" alt=\"Screen Shot 2022-08-07 at 11 40 14 am\" src=\"https://user-\"\nIn repo, ctrl+shift+b is focusing the terminal: !\nReproduces on webgl only\nGuessing it's related to canvas recreation:\nYep: !"}
{"_id":"q-en-vscode-d492bf2b1f9b3e2eb9d93bbd24d50ffd453e825c1424dfdfa420c18606ee5486","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --vscode- a directory to clone your repo into cloned repo :bug: branch context is not preserved This impacts Continue Working On--branch context is preserved for Clone in Container and Create New Codespace. I propose adding a new optional query parameter to the uri handler here: and passing that to the command as an optional argument here Then after cloning here: and before opening the repo in a new window, we would perform the ref checkout:\nwill you create a test plan item/verification steps for this feature?\nyes I will!"}
{"_id":"q-en-vscode-97e398fc34a214628b76625de3e5ff265a3c817b7a72d7d5e7e5ff6d44f7243b","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. -- introduces which controls when we attempt to restore an edit session for the current workspace. The user still has to remember to store their edit session or use the explicit Continue Working On trigger. In order to make it harder to lose edit sessions across , , , and RemoteHub on desktop, we should introduce a setting with initial setting values and . When set to , we'll attempt to store the current edit session (if any) before a workbench unload. This will only really reliably work in desktop because in web we cannot run long running tasks in the onbeforeunload handler (, tracked in ). To mitigate this we should also explore introducing an option to continuously sync edit sessions. That first requires tolerance of edit sessions originating from multiple VS Code clients concurrently.\nOn , we do have a service worker which is used primarily for caching. Service workers can outlive their associated pages, and can be synchronously 'd to like other workers. Perhaps we can use that service worker (or a new one) to allow the workbench to send information to save before it unloads."}
{"_id":"q-en-vscode-1b23c7a024d9ad11836fa97458605797c7b8b8dd22bb52717ded766e16066b71","text":"Looking into adding additional separators that are more common, so that all the additional refactorings are not just clumped under From comment: I was wondering if you would consider adding 2 code action classes that are I think quite common: (Inline) (Move) This would be super valuable for the extension that I'm developing (P42 JavaScript Assistant), but I think also for other extensions. Afaik the TS language server also has actions for example. Originally posted by in\nHey , the new code action widget looks great, and I really like the new code action grouping, I'm wondering if we can define custom code action groups besides the pre-defined groups. ! you can see the java extension contributes a code action (in the light bulb) named \"add Javadoc\" and it's now under \"More Actions...\" menu. If we can define a new group like \"Sources\" via extension, it would be great and more clear that we can put this action in that group.\nWe have an extension that publishes some code actions. With the latest VSCode, our actions are grouped under 'More Actions...' which in our context is not ideal. We would prefer to be able to define our own grouping for our actions (in extension parlance, I think this means to be able to create our own ). Thanks!"}
{"_id":"q-en-vscode-1d7d626f08641d49130e53d471e8db79973401ce0a81e89ac9132f96ac78a05a","text":"! Only seems to appear if I resize the terminal such that a is on its own line.\nCannot repro: !\nI can still reproduce this? For example cat this file to your terminal Then drag until a quote is the only thing on one of the lines. This causes the link detection to detect everything below that quote.\nCouldn't repro on Windows but I do see some offset problem after the resizes: ! Probably an issue with reflow (mac/Linux)\nIn tried on my mac and I still can't reproduce: !\nStill able to reproduce this, have passed repo steps regarding the private repo offline.\nStill struggling to repro, I did notice sometimes it's off by 1 or 2 cells in either direction like this though: <img width=\"852\" alt=\"Screenshot 2022-12-22 at 5 08 03 pm\" src=\"https://user-\"\nThis happens because of line caching, hovering a wrapped line and then resizing it a little such that the line is still there, the line range will be off: <img width=\"866\" alt=\"Screenshot 2022-12-22 at 5 12 44 pm\" src=\"https://user-\"\nThat one repro's upstream: <img width=\"793\" alt=\"Screenshot 2022-12-27 at 6 24 10 am\" src=\"https://user-\"Since I can't repro your actual issue I'll fix this and see if that's the underlying issue.\nCan still repro this building OSS. Note you do not need to resize the terminal for this to occur if it's already the correct size when you run the task.\nUpdate cannot repro when debugging. So odd\nUpdate again, I can repro this again in insiders. Not sure why debugging is different.\nStill can't reproduce so closing as unactionable until there's a solid repro. Insiders: <img width=\"632\" alt=\"Screenshot 2023-05-08 at 6 43 07 am\" src=\"https://user-\"OSS: <img width=\"637\" alt=\"Screenshot 2023-05-08 at 6 44 55 am\" src=\"https://user-\"OSS debugging: <img width=\"633\" alt=\"Screenshot 2023-05-08 at 6 46 16 am\" src=\"https://user-\"cat this file:"}
{"_id":"q-en-vscode-8a0be80bd127cae7bcbd16f9df95ec4052115f8542e50920149321bc2b4fa98d","text":"In the following dialog that appears: <img width=\"696\" alt=\"image\" src=\"https://user-\"in and opening a protocol url, the 'Install VS Code Insiders' button should route to a new link for installing insiders specifically rather than just Install VS Code Insiders should not dismiss the dialog, so that the user can return and try again cc\nlikely a separate issue: Was not sure how this modal was arrived at, so I clicked the remote indicator and I only saw an option to download VS Code - not Insider's as I'd expect. !\nI then did the flow and saw the dialogue. Clicking on the button took me to the Insider's download correctly. !"}
{"_id":"q-en-vscode-0a1d03df981986debdbc681c23289ef8fb72b396102c2520bd690b5ac0a35867","text":"Type: <bBug</bVSCode, create a folder named, for instance, \"XXXX\". that folder create another called .vscode .vscode create a file named with this content: VSCode sure the setting \"task.allowAutomaticTasks\" is not set or is set to its defautl value \"auto\". \"XXXX\" with VSCode. VSCode will run \"echo Hello\" without asking you whether you want to run tasks configured with \"runOn\": \"folderOpen\" or not. VS Code version: Code 1.70.2 (, 2022-08-16T05:35:13.448Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nThis happens because you have presumably opened a trusted workspace - which is now when we allow these to be run if the setting is \"auto\" for allow automatic task\nThe described behavior is how it is expected to work. If you disagree, please explain what is expected and what is not in more detail. See also our . Happy Coding!\n! If I understood it correctly, the popup that supposedly asks you whether you want to run these tasks is not actually meant to appear ever because if the workspace is trusted it will be executed without asking you and if it is not trusted then it will just not exeute them instead of asking you. Doesn't it look kinda pointless to have in the description of the option of the configuration \"Prompt for permission for each folder\"? You are never prompted anyways.\nI this configuration and I am still not prompted anyways:\nNo I did not. Please next time ask the user if he did X before assuming it and closing the issue. I tried reproducing the problem with and without a trusted workspace and the problem persists anyways.\ni'll investigate this that shouldn't happen\nTasks will only run in trusted workspaces, so I will update the setting description to indicate that. Also need to fix the prompt not happening per folder."}
{"_id":"q-en-vscode-451d99d9571387e9fa1af1388e2823bc1d3769998365025cc70850de4e58af40","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --Noticed that SourcePawn extension has this feature, it should be enabled for built-in C/C++ preprocessor #if and #endif, just like for brackets [] () {} SourcePawn: ! C/C++: ! Could probably be done for other languages that have keywords rather than braces for opening and closing a block, e.g. Lua's if-then-end: !\nNot sure if this is editor or C/C++ language config / text mate.\n<!-- -96bd-47a8-a0e8- --This feature request is now a candidate for our backlog. The community has 60 days to the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our . Happy Coding!\nHello there,is this issue open?If so,could i work on this?\nShould make request to Just add as brackets to !\nThis change should be reverted if it can't be made work with all the other preprocessor statements that match the endif: #if, #ifdef, #ifndef, #else, #elif, #elifdef, #elifndef, see matching isn't required, but it shouldn't show the #endif as unmatched.\nit'll work fine with the rest of the #if's just #else etc won't work as vscode doesn't support inside/middle brackets but it won't cause any issues\nah..... one thing I forgot about ! bit of a deal breaker vscode would have to support 'transcendent' brackets similar problem to\nOh -- that seems like a different bug? ! -- The red }, not sure how you got the red {. It doesn't repro with VS Code (non-Insiders). Repro code is Should someone file a new bug on that or is it the same root issue?\nYeah, looks like the PR at fixes it though.\nThanks for reporting this! We will revert the PR as it causes too many unexpected issues. If you like to have this, you can configure the brackets to colorize in the user settings.\nOkay, I figured it out via The directives remain the unmatched color.\nWe closed this issue because we in the foreseeable future. If you disagree and feel that this issue is crucial: we are happy to listen and to reconsider. If you wonder what we are up to, please see our and . Thanks for your understanding, and happy coding!"}
{"_id":"q-en-vscode-9631f8492b6048a8f350d19b5b6ac991b790708e654b9c34f9a95674c1dfed87","text":"<img width=\"584\" alt=\"image\" src=\"https://user-\"//cc for views\nI assume this is nothing to do with the views instead its about how entires are rendered in list/tree. CC\nTree row contents are outside of the tree's jurisdiction, but are in fact responsibility of the widget user.\nTimeline doesn't use the extension tree view and instead does it's own thing. I think works on timeline.\ncurious, do we have an understanding what caused this? it seems relatively recent?\nMight be related to . any ideas here?\nIt was caused by , can you check? I see a visual difference not only with the alignment but also the spacing: Before <img width=\"254\" alt=\"image\" src=\"https://user-\"After <img width=\"269\" alt=\"image\" src=\"https://user-\"\nit looks like timeline relies on the custom tree view css. I made a change in the nesting/ordering of elements in the tree view recently and I updated the css to reflect that. Because timeline relies on the css but doesn't actually use the custom tree view, this broke the timeline. I've opened a PR with a timeline specific css fix.\n<!-- UNABLETOLOCATECOMMITMESSAGE issue --Issue marked as unreleased but unable to locate closing commit in issue timeline. You can manually reference a commit by commenting , or directly add the label if you know this has already been releaased"}
{"_id":"q-en-vscode-44cc400fb2d4e4825bebe52b7a47d8bef9b2ae1775c76f69bb9b0eb6b2387401","text":"ox1D426 all char in a-z without m has unicode highlight. ! ps. Can I get a map table for unicode char and its common type?"}
{"_id":"q-en-vscode-5a0bdfba0ddaadae457c347f6ad435a7ae2b279f5eb00dff6d4714e80ef21451","text":"<!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.72.0-insider () With the C++ TestMate extension, the hierarchy of test executables and contained tests is no longer displayed correctly. This is a regression that was introduced with . Here's an example. Create a cpp file with the following content: and compile it using taken from . Then, set the config to point to the compiled executable. With VS Code v1.71.0 you get the expected browser content: <img width=\"581\" alt=\"image\" src=\"https://user-\"With the latest Insiders build, or any build from on, you get this instead: <img width=\"581\" alt=\"image\" src=\"https://user-\"Note how T1 is now a sibling of the executable, not a child. Also, the executable no longer has a play button. Cc: to since I'm not entirely sure if this is VS Code's fault, or if there's maybe something wrong with how the extension sets the URIs.\nI'm surprised would have anything to do with it, since that only touched decorations in the editor and not the core heirarchy...\nWell, I did bisect it to that commit...\nSome more information: test rediscovery is completely broken as well since that commit. I'm talking about the situation where you change the config and TestMate reevaluates which executables should show in the browser (I ). In that case, both the old and the new executable are displayed now, both with all their tests. This is reproducible and also started to happen with This is a very serious regression and makes C++ testing a lot harder. At the moment it only affects the Insiders build, but I would hate to see it become a regression in 1.72 if it's not addressed soon. again, I can't tell which of these problems are caused by VS Code or by the extension, so it would help a lot if you could have a look, too.\nI'm doing: where : results in: around: I hope it helped\nHuh, I'm very confused why that wasn't a compilation error. Thanks for the hint on the error message...\nWhen will it be this part of the downloadable insider to test it?\n<=24 hours after it's merged The bot will ping on this issue when it's ready\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\n/verified"}
{"_id":"q-en-vscode-bef2b9e14d760e0c1c7bf08f377861be42921b5affd2a287518732931e83373f","text":"Chinese (Simplified) : <img width=\"321\" alt=\"image\" src=\"https://user-\"English : <img width=\"272\" alt=\"image\" src=\"https://user-\"\"Create Manual Folding Ranges from Selection\" is not in Command Palette when use Chinese (Simplified) VSCode Version: 1.71.0 (both macOS & Win10) Chinese (Simplified) Version: 1.71.\nThank you for your feedback. We are currently reviewing it (reference ).\nCould you please assist here? Thank you.\nThe problem is this label: should accept an  but it only accepts a string. deeper: and deeper:\nThe workaround is to use  to find this item."}
{"_id":"q-en-vscode-7626d29388379293b7c84333d9f70151633147833415a93d614395b50f987580","text":"GitHub Tags:#A11yMAS;#A11yTCS;#A11ySev2;#DesktopApp;#Win32;#Visual Studio Code Client;#BM-VisualStudioCodeClient-Win32-Aug2022;#WCAG1.3.1;#NVDA; #ScreenReader;#JAWS;#AILimited;#Element:Button; Application Name: Visual Studio Live Share Application version: 1.56.2 (user setup) Windows Version: Win10 VS Code a project (e.g., vsls-guestbook) in VSCode. till “sign in” button. till \"Live share\" under Extensions and hit enter on the NVDA and tab to drop down 'Visual Studio live share work' screen. The label of the 'Visual Studio live share work….' is being read as 'folder section button' when the screen reader focus moves to it using tab key. Similar issue is observed with JAWS. The screen reader should read it as 'Visual Studio live share work button' when screen reader focus moves to it. People who uses screen reader may face difficulty while activating an object when the accessible label is different than that of the visible labels. https://user-\nlogged for primary bug\nI think this issue makes sense. In short - aria label of the explorer view should contain the name of the top folder. Currently it is just \"Folders\". So it should be dynamic. Assigning to\nActually not too sure how to fix this one. The section being read out is the static section name of the view pane We do have an aria label within that section header with the name of the folder which dynamically updates but that never gets focused. The pane registration is static and we don't know the folder name at that time\nthere's the  method in the . Update header should set the aria label of the header. And the explorer should call updateHeader every time the content of the header changes (which it might do already).\ncould you please the bug in open state the bug can be closed only by trusted tester after verification right now the fix is not present in the app Version: 1.72.0-insider (user setup) Commit: Date: 2022-09-23T05:17:37.024Z Electron: 19.0.17 Chromium: 102.0.5005.167 16.14.2 V8: 10.2.154.15-electron.0 OS: Windows_NT x64 10.0. Sandboxed: Yes !\nplease look at the label that the VS Code bot has to the issue. It says \"unreleased\". Once the bot adds the label \"released\" that means the fix is present. We can not leave the issue open for the tester to close this since this does not align with the rest of our VS Code processes. In case the bug is not fixed after it is released, just comment on the issue and we will reopen it. Thanks!\nGitHubTags:#Closed;"}
{"_id":"q-en-vscode-66692775fba94e9b061da2a96eb1379766c0115542a648e2472cb0e2f6847c69","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.71.0 OS Version: MacOS 12.5.1 Steps to Reproduce: I'm not entirely sure how to repro this as the error doesn't really indicate the cause. What happens is, when I click \"Sync\" on Git, I get this modal warning: <img width=\"276\" alt=\"image\" src=\"https://user-\"And the Git log itself doesn't immediately have any signs of failure, no results for searching the word \"error\". It seems that some Git command failed internally but I'm not sure why the failure reason isn't part of this modal window. If the reason is in the Git log (which currently is quite large as I've had vscode open all day) then it's not immediately obvious which command caused this error. Shouldn't the error message be part of the alert dialogue that appeared?\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.71.1. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nthat does seem like a bug. If you are still able to reproduce the problem, could you please click on the \"Show Command Output\" button and share that? Thank you very much in advance!\nyeah it turns out I had some difference in tags between my local and the remote - I got an error about \"clobbering\" tags (?) overwriting I guess? when git-pulling with --tags via the terminal. I think what happened is myself and another team member ran our release script and generated the same tag on two different commits.\nthanks for the update. If you are able to come up with a set of repro steps that yowl be very helpful. Thanks!\nManaged to reproduce. This was done in a repository where I was a few commits behind the remote copy. I tagged the current commit which was with then pushed --tags pulled down the new commits from remote, my local is now I deleted the tag called which pointed to I a new tag, also called on now there's a tag called where the remote points to a different commit to my local, git cannot fix this itself I run and I see the error message: Now when you click ! in vscode, it runs in the background, and internally it hits this exact same error but the alert box that opens does not contain the actual error message: ! One observation: on Windows I've got this \"Show command output\" which when I click, it opens an editor with the full log so that's useful at least! On mac this didn't work, so maybe that's the actual bug here. Full log for reference:\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nVerification steps: Use the repo steps from In case of a conflict in tags, we now display a dialog listing the tags that are conflicted <img width=\"273\" alt=\"image\" src=\"https://user-\"* Click on \"Replace Local Tag(s)\" and confirm that you are able to successfully sync\n/verified Amazing, awesome work!"}
{"_id":"q-en-vscode-01c6ebd9d1711f2882750b7d4fedbaaa2444a639174fa0228882ad88c4dc3ede","text":"Type: <bBug</bI saw the schema prompt for 'types' under terminal contribute, but setting it does not take effect. After checking the vscode source code and issue, I found that this contribution point has been removed. VS Code version: Code 1.71.0 (Universal) (, 2022-09-01T07:25:38.437Z) OS version: Darwin x64 21.4.0 Modes: Sandboxed: No Remote OS version: Linux x64 5.10.3+ <details<summarySystem Info</summary </detailsExtensions: none<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\n!\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.71.1. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nRelated Links:\nAfter upgrading to , it still remain."}
{"_id":"q-en-vscode-2af7595938df1edd965064c330b62302fbade74b1ee550cb4d3cefc0a7ebcec3","text":"Here is my code snippet Trying to run this creates:\nI can't seem to reproduce this on code-web or\nI was using The only changes were these in\nI can still reproduce this in my code. I ran out of source to see what might cause the problem and here is a detailed stack trace: Might be some timing issue\nMore digging. Here are my findings: I do open the terminal from a debug adapter which brings the debugger viewlet to front right after I created the terminal if I wrap my code into a everything works fine\nany tip or fix for this. I improved the performance on my end and a 2 sec delay is now very bad. My Python execution in the Web is now < 500 ms.\nI need to reopen this. I tested it on todays insider and I still get\nBuild I used is: <img width=\"500\" alt=\"image\" src=\"https://user-\"\nI will compile steps today on how to reproduce this.\nSteps to reproduce: clone cd vscode-python-web-wasm git co dbaeumer/noTimeout npm install npm run compile npm run esbuild open VS Code run as a task or in the terminal run as a task or in the terminal click on the link in the tunnel terminal and press continue open a browser on Run command Use the link from the tunnel open developer tools open in the browser window From the play menu in the editor run Observe: you get the above exception and the terminal is not showing the output Run the action again Observe: the terminal reveals correctly and shows the output\nCan reproduce with those steps, the only thing missing was enabling workspace trust and reloading after installing the web extension\nDebugging this section: It's visible before so it doesn't do the extra open view, and not visible after it because the debug console steals the opened view.\nEven if this issue was fixed such that did not throw (eg. make an optionally synchronous version of or just try ), it would still set visible and then immediately show the debug console. So if you wanted to show the terminal, you would need to call it after the debug console shows (eg. via timeout as you have). I don't think this is an issue that needs to be fixed in the terminal, but rather you should either prevent whatever is showing the debug console as it happens some time after the terminal is created, or move the call to after the debug console shows. let me know if you disagree.\nIMO there are two different things: the fact that the terminal is not in front because the debug terminal comes to front. I do agree that this is something I should address the fact that the terminal seems broken (e.g. it doesn't render the output for the first created terminal). This I agrue is a bug in the terminal. IMO the terminal should function correctly after a create even if I don't call show. So if a user later on clicks on the terminal the output is shown correctly.\nthe second issue should be fixed by , it won't fix your problem though\nto understand it correctly. This will ensure that the terminal is functional after . I can then ensure that I bring the terminal to front () after the debugger is initialized.\nyeah, removing that await should ensure the view is visible so the element will be there to attach which means no exception that ruins the state."}
{"_id":"q-en-vscode-113b94828e758e4759fcd8aec852b804af8800bba09ffb69e440916e6c064b25","text":"Locally I have an empty cell at the bottom of my notebook because I executed the last cell When I try to merge in a change that has new cells, I get a merge conflict that requires I open the ipynb in a text editor to resolve. !\nWe can explore how to handle this gracefully in the serializer.\nnot sure what needs to be done in the serializer here. can you provide more information, i would have thought this was purely an issue with merge conflicts.\nCan this please be solved? It's hard to work collaboratively within a team using Jupyter notebooks and VS Code if we can't sort out merge conflicts.\nI suggest there be an option to disable the notebook renderer temporarily in all editors. git operates at the text level and when merge conflict happens, it will likely produce a malformed JSON file. We now have an error message to notify users to fallback to text editor, but \"resolve in merge editor\" button still don't work. !\nNote: When there are merge conflicts, one will always have to open a plain text (json) editor to resolve the merge confligs I.e. its not possible to use Notebook UI to resolve merge conflicts (at least not yet) Steps to test Ensure you have merge conflicts in a Notebook document Open the SCM view, and click on the IPYNB file, and verify you can see the merge conflicts in plain text editor view Click on the button and verify you get the merge editor (again plain text), see here When you have merge conflicts, try opening the ipynb file from the explorer view, if this fails verify you are displayed with an error message with the option to open the file in plain text view"}
{"_id":"q-en-vscode-6a70eab42509901c265c452297c694512c9cea70cf8b200746cf6d02bf654e1a","text":"Fyi there is no long running shutdown support on web, so this warning gets printed on unload: <img width=\"955\" alt=\"image\" src=\"https://user-\"\nIn web we probably cannot support the value.\nThis is still on\nI cannot reproduce this, just to be sure this is on the latest\nYes, reproduces even with caches disabled: !"}
{"_id":"q-en-vscode-196df7edebace878fea0c66587bb689df44b90c7ca6a237874288107c1449f29","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.71.2 OS Version: macOS Monterey 12.5.1 (21G83) Steps to Reproduce: file named the file to the title of the file's editor tab <img width=\"199\" alt=\"image\" src=\"https://user-\"Is still , not the new name . until reopen the file. But rename to is ok. Also reanme to is ok. Just rename to is not ok.\nCan you try to reproduce with our nightly insider builds? You can give our preview releases a try from:\nrename to is not ok. rename to is ok. It seems that some intersting things work wrong here.\nCan you try to run where is an empty folder? This will ensure Code is starting with a fresh data directory, e.g. no specific settings and without any extensions running.\nAs you see, nothing changed.\nHm yeah I think I see it too, the first rename works fine but then renaming back does not work, is that what you see?\nthis is a funny one and indeed a regression from our work to always go through the editor resolver when resolving editors and not go through the as a shortcut. Steps open an empty folder with 1 file open the file rename it from the explorer to verify the tab label updates rename it from the explorer to the label does not update Here is how it used to be the rename operation will eventually hit since this is a default text editor we would directly use we figure out that we have an existing editor input so we do not create a new one however we update its preferred resource this triggers a label change event that will update the tab label =this works every time you rename, irrespective of going back and forward of the name Here is how it is today we also end up in however we always now go to the editor resolver (which imho is good) the editor resolver however does not consider the input to be a new input after the first rename because the canonical resource is still how it was in the begining, it will do an early return as such, the inputs preferred resource is never updated Now I am actually not so sure how to fix this. I am a bit worried that updating the method to respect resource path casing may break other behaviour that relies on the current behaviour. Maybe the resolver should not be so clever to do an early return? Why do we do this?\nYeah maybe, I think it was to prevent creating too many inputs and calling unnecessary factories when things are not actually changing at all. I guess we could always remove the early return and see what breaks.\nAt least for text file editors, there is caching so that we do not end up with an editor input each time, but I think the same issue already exists today: if you do not have a tab as active and the editor resolver is used to bring up that editor, we would delegate the decision whether to create an editor input or not to the provider and so its up to the provider imho to do the caching.\nSorry, got caught up in other work and missed this. Will attempt a fix early in debt week for the November release as it is too risky to go in this late"}
{"_id":"q-en-vscode-6fff1a019633c53127e2c4d13b6b6124bc396e22f3d352fe401b86c5dd7f2088","text":"Open terminal and run command After it finishes, there is a freeze. Some times I see window is not responding dialog Now type command Freezes again Here is the profile that is captured while running command Here is the profile that is captured while running command I have set to\ngetOutputMatch can be too expensive: ! Is some quick fix not working on just a small line range?\nThe second one: !\nOh none of the quick fixes use offset or length currently. We should force that, otherwise the regex will run on every line."}
{"_id":"q-en-vscode-0f304df3e012dfddf2ec11d395c1c6fe615e6c731a4595deaa67b405e03c8a61","text":"Testing clone a repo from within VS Code when done a notification is shown IMO a modal is the better choice of UI because this something I am awaiting and wanting to take action on <img width=\"1527\" alt=\"Screenshot 2022-09-27 at 11 15 02\" src=\"https://user-\"\nit makes sense to me. please let me know if you have any concerns. Thanks!\n, as having notifications snoozed always, I was missing this notification but I think it should be modal as it is part of a flow that I triggered.\nNo concerns with using a modal"}
{"_id":"q-en-vscode-42949ae023af4ebdd70106c0934d31b7f23abe318b6a4edec46a74a11c436100","text":"Testing The feature is called . But it does more than fetch, it also pulls. Fetch typically only 'downloads' commits, but does not update any branches. Maybe the proper name would be:"}
{"_id":"q-en-vscode-17810cff045cbfd93adfb40cd6fa1dd10ffaa4d38f9080f64446e04e32c47230","text":"Testing I wonder if we should run the command to save all files (not untitled files) before triggering the \"Continue on\" fully? I think ensuring that dirty files are saved but also that the data was synced up before opening the workspace elsewhere would be best?\nI am not presented with any options after running the command - perhaps we shouldn't show that in the command palette unless there are some or add an action that helps me get some options? !\nShould we say here? !"}
{"_id":"q-en-vscode-bdfa0c95fd34fa24c5f7a5d09a6e519b3eeca4a39454cae5f6d3f676272c629e","text":"Testing I think the merge editor CodeLens actions are the only examples I've seen of icons being used—not sure they are obvious/big enough to really be of much value here. Is there any prior art for using them? <img width=\"269\" alt=\"CleanShot 2022-09-27 at 11 42 19 src=\"https://user-\"<img width=\"244\" alt=\"CleanShot 2022-09-27 at 11 43 33 src=\"https://user-\"\nI can remove them. Personally I like them, as they give a non-textual clue about what is going to happen.\nWhy don't we start with the minimal text and then we can always choose to add them (or something else) back. Harder to remove things once they are in."}
{"_id":"q-en-vscode-b69c0690c2519f95aab030c47bf64a8d481c1199112ca95b1da195b8ed08c5c0","text":"Testing ! I would have expected it to be centered.... though this could just be this codicon.\nIs this happening with all codicons or just this one?\nI can see it for codicons. Fixing it"}
{"_id":"q-en-vscode-380ae46906e1f1854614e7d2d73ec22af60fd8b13a46d94af8892aad31338243","text":"<img width=\"364\" alt=\"image\" src=\"https://user-\"I think this dialog could get some UX love. If we want to keep the very long primary button label, I would use title casing to be consistent with other dialog buttons. But I also think just having \"Accept Merge\" is fine, given the message in the dialog points out about the conflicts. //cc\nNow that I see this dialog: <img width=\"291\" alt=\"image\" src=\"https://user-\"I think we can align both How about: \"Accept with Conflicts\" using in the dialog \"The file contains unhandled conflicts\"\nPR:"}
{"_id":"q-en-vscode-50cd005a8d9a954035057ba52a0af3720bbc4fb57bebdb8f9f47012b10bbaec3","text":"Just wondering whether the \"Extensions\" output channel is actually used for anything? I have checked it when troubleshooting extensions many times and don't think I've ever actually seen any output in there\nThis is being used by debug adapter Feel free to remove it if you are not using.\nOh, wow No clue about that\nI guess it was introduced in , previously showed errors from all extensions, now I think those errors go to other general logs, and probably just this one usage got left behind. any idea if there was some discussion about this in the past? Or I will remove it\nnot aware of anyone using it. I think we can go ahead and remove it. fyi\nbit off-topic here but I was looking around for an issue around a \"combined\" log (showing the channel inline) - is there such a thing and I'm just missing it? that way I wouldn't have to click around into various logs when I'm trying to figure out what's going on with vscode - I realize I could tail the various files and write a script to make that more convenient I was also thinking maybe an extension could fill that need but then I saw\nI think there is an issue to open multiple output channels at once. Please search into our issues.\nthere are only 10 open issues tagged output - altho this issue is about output and got labeled debug instead - I looked at the debug tagged issues too and didn't see it - should I create one?\nHere is the similar request -"}
{"_id":"q-en-vscode-40e26b0d6ceec9c7908e3ed799220b72e3364c904cb2b13b760a29009e4acc0a","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: n/a <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: (currently at ) OS Version: n/a Steps to Reproduce: the existing code in and turn it into a function that accepts a parameter: that function in , e.g. using . should error, but don't: - According to the error on , requires =1.10.1, but is only enforcing = 1 or = x.10. Granted, there is no yarn v3.10 yet, but there may be someday. And if it actually requires =1.10.1 (and not just =1.10.0) as the error states, it should enforce that as well.\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.71.2. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nRelated docs change:"}
{"_id":"q-en-vscode-d87ee6f2e3fca592a7ab77b5da2be2eca2ad99db20718285a58a04f6354c2881","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.72.0 OS Version: Windows 10 Pro, 21H2, build: .2075 <brCommand links inside notifications were implemented as described here, or here , i.e. This is now possible using the following syntax: Originally posted by in The feature works great when a single command link is to a notification but fails when two (or more) are present, see the sample code below (2). <brSteps to Reproduce: a new blank extension (JS/TS) the following code to your extension's main entrypoint/command: Expected behavior is to see the notification with 2 command links and their titles but the following is rendered instead: <p align=\"center\"<img src=\"https://user-\" alt=\"notification-command\"</pNote: the provided commands and should be valid and registered for the commands links to work when actually being clicked but that is irrelevant and out of scope for the issue.\nThis looks like a parsing bug to me when converting markdown to HTML.\nBug actually seems to be in since notifications don't use full markdown:\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\n/verified\n<p align=\"center\"<img src=\"https://user-\"</pIt now works as expected, great job, thanks!"}
{"_id":"q-en-vscode-929c7e48871ff728190f22909017464dd0e97fbc1973b59fd6ef7060d5b5f4e1","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --Hello, I searched for an existing ticket with , but didn't find anything related. <!-- Describe the feature you'd like. --Prior to v1.72.0, the extension sidebar showed which version an extension will be updated to. And when hovering over the extension in the sidebar, or opening the extension in a tab, you could see the currently installed one. That way you could see for example if an update is a major, minor, or patch update, and depending on that dig into the release notes or not. Further in the past, IIRC, the extension sidebar even showed both, installed and target version, which was even more convenient. Now with v1.72.0, I don't see anywhere the target version of an extension update. Many extensions don't keep a proper changelog, so that's not a reliable place to check. Here's an example for vscode-icons, a very popular extension with 12 million installs: ! Only when going to the Marketplace webpage () we see the target version: ! =Feature request: Please make the target version for extension updates visible inside of VS Code again.\nLatest version is now shown in the hover. Update button in the extension editor shows the version."}
{"_id":"q-en-vscode-c5ce083f6b31876e779ec8d0951ba90255a923682d8b14e20e6b440bb1e7ec96","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: Version: 1.72.1 (Universal) Commit: OS Version: macOS 12.6. Steps to Reproduce: this file: sure it has a extension. all current instances of VS Code it with VS Code with all extensions disabled: around. In a few seconds the editor will lock up. This file has some extremely long lines, around ~65,000. I was under the impression though that VS Code stopped rendering at some cutoff, however. It appears the issue is the syntax highlighting for C#. If I open it as plain text, then I can no longer reproduce the behavior.\nHere is a reduced repro: By using the command, I could find that the oniguruma regex engine gets stuck trying to match the regex: which comes from As a workaround, you can reduce the limit for C# files from 20k to something smaller e.g.: We also have the option of reducing C#'s default. We already did that for javascript to 2500."}
{"_id":"q-en-vscode-b799770cc89c4d83dda5dbad899f906699c6866f6ae03aa60cfebed7ef34b54c","text":"Repo explorer, collapse the timeline view a breakpoint in  window Bug See the breakpoint being hit even through the contents of the timeline are not visible This is caused by this code: It seems to me we should try to defer this work until the view actually becomes visible\nwould you be willing to verify for completeness?\nWorks great!"}
{"_id":"q-en-vscode-c031b28303fffbe622cd54dc1be02a95cc469127d2905bbbe4384232c0d16f29","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.72.2 OS Version: macOS Montery 12.6 Steps to Reproduce: merge editor while rebasing on a branch with a lot of branch pointers or ones with a long name The text from the \"Current\" column overlaps into the \"Incoming\" column. Confirmed that it happens with the default theme as well. !\nPlease try out the latest insiders, this should be fixed.\nStill appears to be an issue in the latest insiders ! !"}
{"_id":"q-en-vscode-23b95517cda23ccb26fb7baf74a5932155b983228ab6bda7048fbb17f4ddbfbf","text":"These are a few more unused exports which git blame says have been around since Hello Code, based on I removed a few false positives which are indirectly referenced through the public api exported from these files and which therefore should probably remain public for consumers"}
{"_id":"q-en-vscode-8b1b4d2a8a6ca7efc70a35068c3c0e8536ce2790351acb19b59749b48ee156f0","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.73.0-insider OS Version: KDE Neon 5.26.2 Steps to Reproduce: to settings (UI). for \"duplicate\". \"Explorer: Incremental Naming\" to . the file explorer, hold down the key (to copy a file instead of moving) and drag a file to the same directory. Actual result: VSCode hangs until I am presented the \"This window is not responding\" dialg. Clicking has the same effect as . Expected result: I am telling VSCode to duplicate a file, while I have taken away its algorithm to automatically create new file names. Therefore I might expect VSCode to show a dialog with a text input. The text input would contain the file's current name, which I can alter. Confirming the dialog would copy the file using the new name. Or something to this effect. But maybe that solution is trying to counter a symptom rather than solving the root of the problem. Because how would should VSCode respond when drag-copying multiple files? Maybe the root of the problem is the ability to set \"Explorer: Incremental Naming\" to in the first place. Maybe VSCode must always have a 'duplicate file naming algorithm' in order to function? Even if it is a one. Removing and using the algorithm as a bare minimum would surely fix the problem. Addendum: The description of states the user \"will be prompted to override the existing file\". So this might be the expected / intended result."}
{"_id":"q-en-vscode-953073402dcf159d8212ce059bcae23824c1137fe80c111f2a9632e3b0bd31b6","text":"I noticed that there are a couple colors here that are available in : ! but the rest of the colors are available in the monaco workbench : ! these probably should be with the others. possible code pointer:"}
{"_id":"q-en-vscode-3163dc249ee3ddf956edba92a2a8dca38d6191f8e114a8ebea190c85e64d7ef3","text":"Type: <bBug</bThe experimental \"user profiles\" feature uses absolute paths which break when using a (which is done by creating a \"data/\" folder in VSC's root directory). These paths seem to be stored in \"data/user-\", at least in part. This makes VSC forget any extensions installed on the default profile as well as any other profiles. It also prevents VSC from finding the settings for any non-default profiles: When trying to edit the settings for a non-default profile, a new \"\" file is created where VSC expects the file to be. It displays an error message instead if the expected path is invalid (ex. portable installation's drive letter changed). Steps to Reproduce: a portable (.zip) version of VSCode and extract to a folder. an empty \"data/\" folder inside that folder. VSC and install any extension with the default profile. a second profile. Change some of its settings and install any extension. VSC and move the VSC folder to another location. VSC again. VS Code version: Code - Insiders 1.74.0-insider (, 2022-11-03T18:43:35.409Z) OS version: WindowsNT x64 10.0. Modes: Sandboxed: Yes <details<summarySystem Info</summary </detailsExtensions: none<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --"}
{"_id":"q-en-vscode-d1f86c446cfbc5eb3989a54b125dafcec2da3fb31d8c506fd404cbe767d113c9","text":"This is a follow-up from and describes the problem that save participants causing a save operation\nreproduced when undoing some fixes (applying the fix below)\nfyi"}
{"_id":"q-en-vscode-7dae061ba3d1638005fd09904c6784556e4ea65fa963baba0a94e2bcd5f9023e","text":"Currently when there is one suggested kernel and that's the one users are using in current document, its presentation in kernel picker is weird <img width=\"601\" alt=\"image\" src=\"https://user-\"This can be simplified as a single \"Selected\" group at the very top <img width=\"601\" alt=\"image\" src=\"https://user-\""}
{"_id":"q-en-vscode-f77e558da8ced409736b73f8864b15364c02f867caa01ce911b58da192c1611e","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --I love that vscode had pylance for python toolchain, this is frustrating when you have a language support but an editor default that does not show syntax error. Please reconsider these samples of scenarios, Python uses indentation to delimit code blocks, which is 2 or 4 leading whitespaces. YAML uses indentation to delimit nesting level, which is also 2 or 4 leading whitespaces. Python is afaik the majority in ML community, see e.g. has 98.0% of python code, both pytorch and tensorflow has a python frontend. YAML has grown popularity comparable to json as configuration tool, e.g. a majority of linter that supports also supports . Ignoring whitespace diff actually leads to syntax error and thus broken code, it is not just a visual difference. <!-- Describe the feature you'd like. --Make this default:\n<!-- -96bd-47a8-a0e8- --This feature request is now a candidate for our backlog. The community has 60 days to the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our . Happy Coding!"}
{"_id":"q-en-vscode-e3bca3f89572df5e7d694e4050f429735c6f8bb501bac471e368639e242ff01a","text":"Steps to repro: a new instance of VS Code (no extension installed) an ipynb file running a cell out of extension installation suggestion Cell is still \"running\" and cannot stop execution (Expected: Cell execution is stopped or can be stopped by clicking )"}
{"_id":"q-en-vscode-f8b01bda7a395f2a0409efc2f6014a2ca5cce38f65d321bde45b3d20cba2901c","text":"With I went over most editor contributions (50+) and adopted idle value instantiation where I could reason that they would still work correctly. But some editor contributions interact in more complex ways, sometimes with each other, and I wasn't sure. Hence, this issue. I would like to ask that all calls explicitly use . If a contribution uses eager, a reason should be written as a comment next to it. There are multiple options available: Some things to keep in mind when changing an editor contribution to something else than : the contribution will not be created necessarily before a text model is attached. Normally, a contribution is created so early, at a point when the editor has no models. So it is safe for a contribution to simply use to check if the contribution needs to do anything. When being instantiated later, it is possible that the contribution will be created at a time when the editor already has a model, so the contribution should be prepared to deal with that in the ctor. if a contribution needs to be activated via editor events (pasting, typing, mouse events), then most likely you can use . if a contribution is mostly driven by actions or commands, then the contribution will be instantiated when the first action that needs it is executed, so it is safe to use . if a contribution is only driven by actions or commands, it is safe to use . Anyways, here's the remaining list of contributions for which I wasn't sure what would be correct:\nReplacing me with who took over quick pick."}
{"_id":"q-en-vscode-fc8dd226da21c64a0a3cb8c0bbeb77839e471e6e1d494933f19e5c80f112768e","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: Version: 1.73.1 (system setup) Commit: Date: 2022-11-09T04:27:29.066Z Electron: 19.0.17 Chromium: 102.0.5005.167 16.14.2 V8: 10.2.154.15-electron.0 OS: Windows_NT x64 10.0. Sandboxed: No OS Version: Windows 11 Steps to Reproduce: a volume in Window that contains a git repo directly in the root of the volume (in my case it is an NFS share hosted on my NAS, but this also happens if it is simply an empty disk drive too) the root directory in VS Code and go to the Source Control window that it shows the diffs, but all the folders contain the volume letter at the beginning of the folders/files and the diff previews don't work !\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nI do see this message even though I created the volumes: !\nI think you can capture that in a new issue, but you should be able to open files from the diff if you trust the repo\nYes. You can open files. But since this is I removed my\nauthor-verification-requested's can still be verified by team members (and often are if we don't get responses in time for endgame). It's just a helpful label so the bot can post a message and start the process asking the author to do the verification.\nLooks good! It works for me! Thanks so much for fixing this! :) /verified"}
{"_id":"q-en-vscode-12bfae9c4d4eec713a1af163345aa0db82529ad30805cc072ec8862832ed9906","text":"I just now noticed how the progress bar was a the very end of the quick pick but not below input box, I think it should be below input box because the progress is for what I typed. Steps: have some history entries in Cmd+P picker be on a slow OS (Windows) start to type and notice progress bar is below the history entries //cc\nSeems like this might have been a regression at some point. for relocating it back near the input.\nI like where it is now. To me it says \"the results are still populating\".\nThe reason I do not like it: the progress bar is at a non-deterministic location depending on the number of items the progress bar indicator is away from the location that triggers it (the input box)"}
{"_id":"q-en-vscode-a2f991c13850fa839e51c8fbb7f9e068c6d748d2fb5ac1c4455ead84ccba22f8","text":"Testing In the export flow, there's the step to \"Export as...\". The options are \"File\" and \"Github\" <img width=\"615\" alt=\"image\" src=\"https://user-\"I can make the educated guess that I'm not exporting the whole of Github from my VS Code window, but it'd be better to say \"Github Gist\" instead, or \"Export to\" instead of \"Export as\". I prefer the former since it's clear where it's going--otherwise people might assume it creates a repo on their account or something.\nIt was GitHub Gist before, after feedback I changed to GitHub. I will bring this to UX and see what other think too.\nGood feedback. We will also do a user study next week so we will see how users react here. Alternative rendering: \"GitHub\" and then \"gist\" as the description a bit to the right - QuickPick should support this.\nThere was some confusion in the user study when exporting as GitHub. They were expecting that it will be part of a repository. So calling it GitHub Gist might be the way to go.\nI think adding as description would be good\n!\nLovely, tried it out. Verified."}
{"_id":"q-en-vscode-4fbb86983cf5a3247a1f2a66aa87d04290ba0a96fd76d7a192175fae85a86d76","text":"My impression of the slowness that mentioned In stable and insiders, I think the run button shows up at about the same time. Around 7-9s from \"reload window\" to the run button appearing, hard to tell whether stable is slightly faster. But after it appears, and the preferred kernel has been discovered, I see \"Discovering Kernels\" for another 5s or more, which could lead to an impression of slowness if the user doesn't realize they can go ahead and run a cell at this point. Originally posted by in\n<img width=\"183\" alt=\"image\" src=\"https://user-\"When there is already a selected kernel, we should show the kernel name and a spinner if there are pending detection tasks."}
{"_id":"q-en-vscode-b3fa76f3edcfe637cbd34af81529848b8ee6199e63125b3166b0f0d50cc5afd2","text":"Testing microsoft/vscode-jupyter Open a new window where no kernels have been used Open the kernel picker, get the kernel source picker Click the back button The quick pick flashes closed/opened and shows the source picker again\nWas trying the MRU out with my new machine and ran into the same thing. Just adding a few more comments: The progress bar makes it look like the process to find kernels just hangs I'd expect to see like the old experience feels unfinished--it's my first time picking a kernel for this project so I expect it to tell me to just !"}
{"_id":"q-en-vscode-b7fc99e331aa661d93a1cd429f915f23bbab78bed337add2356ec31afd5970ac","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: N/A, extensions are required. <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.74.0 (user setup) Commit: Date: 2022-12-05T16:38:16.075Z Electron: 19.1.8 Chromium: 102.0.5005.167 16.14.2 V8: 10.2.154.15-electron.0 OS: Windows_NT x64 10.0. Sandboxed: No Steps to Reproduce: a network UNC based workspace with a GIT repository that is not currently marked safe. the process to mark the repository safe via the new feature. The repository still will not work. Looking at the .gitconfig file will show that the UNC path was to the safe directory list without the needed prefix.\ndo you mind sharing the exact output when running inside the unsafe repository as well as the exact path that is being to the file? I am asking as we are currently extracting the path from the output of the command so the issue could be upstream. Thank you!\nHere is an example ... I've removed details such as the SID's for security: Note the suggested command to add the path to the safe.directory list includes the needed prefix, but its possible that this is something specific to 'Git for Windows'? This is my personal documents folder via Folder Redirection, its stored on a network path, and since I am the systems administrator, my files are all owned by the administrator, thus ALL my repositories no matter where they are, are considered unsafe. :(\nthank you very much! I was extracting the path from the first line of the output which is not correct. Will work on a fix.\nYou might also note, GIT seems to also flip back and forth as to whether the server in the UNC path should be all capitalized or all lowercase. Don't know if it is something VSCode is doing that causes that, but it can change its mind between two sessions even though the case of the path doesn't change from what I can see in VSCode. This causes an already marked safe repository to be marked unsafe, which is exactly what happened here.\ncould you please expand on that? Are you seeing this casing issue in the git output window or VS Code UI?\nDepends on where you look. So far today, VSCode's Explorer is indicating the path's server portion as uppercase and the source control is also reporting the server portion as uppercase (hover over a repository), so they are agreeing, but next week or next month, it will switch and they won't agree, and it usually is GIT that has changed, not what VSCode shows in the Explorer view. (Explorer view) ! (source control view) ! I've been trying to track this down for a long while, and now the safe-repository mechanism also gets tripped up in this.\nthanks for the example. I would suggest that you keep an eye out for any discrepancies on the casing of the path and file a separate issue when it occurs again. As far as this particular issue goes, the fix will be available with the Insiders release next week."}
{"_id":"q-en-vscode-b5fef431cd83e674fb85525ab4ea866ef7d18599ee4c7e4de2bdd58debf18267","text":"Type: <bBug</bthe set to the cursor around with arrows. Animation is there. something. Animation is not there. VS Code version: Code - Insiders 1.75.0-insider (, 2022-12-12T05:36:38.306Z) OS version: Darwin arm64 22.1.0 Modes: Sandboxed: Yes <details<summarySystem Info</summary </details<details<summaryExtensions (42)</summaryExtensionVersion ------ better-comments3.0.2 vscode-nginx-conf0.1.3 ng-template15.0.2 vscode-tailwindcss0.9.1 better-toml0.3.2 npm-intellisense1.4.3 vscode-eslint2.3.0 vscode-quick-select0.2.9 xml2.5.1 gitlens13.1.1 scroll-faster0.2.2 prettier-vscode9.10.3 vscode-jest-runner0.4.59 auto-rename-tag0.1.10 shell-format7.2.2 macros1.2.1 vscode-pull-request-github0.57. vscode-postfix-ts1.11.3 Angular213.0.0 center-editor-window2.3.0 graphql-for-vscode1.15.3 vscode-clangd0.1.23 ts-error-translator0.5.1 dotenv1.0.1 vscode-docker1.23.1 black-formatter2022.7. isort2022.9. python2022.21. vscode-pylance2022.12.21 vscode-typescript-next5.0. vsliveshare1.0.5788 prisma4.7.1 vscode-yaml1.10.1 better-pageup-pagedown0.0.1 rust-analyzer0.4.1316 vs-code-prettier-eslint5.0.4 justfile2.1.0 vscode-scss-formatter2.4.4 shellcheck0.29.1 vscode-input-sequence0.2.0 errorlens3.6.0 vscode-lldb1.8.1 (1 theme extensions excluded) </details<!-- generated by issue reporter --\nMaybe related to /assign\nThis was an intentional change to disable the animation when typing/deleting. See\nokay, that makes perfect sense. closing.\nAlso having the same issue. Has there been a fix?\nWe have intentionally removed animating the cursor when typing for which explicitly asks us to do that. But I think some folks actually prefered the old behavior of animating also when typing?\nYes, I love the animating when typing.\nA good common ground could be that it becomes a setting to also have the caret animated while typing\nI've pushed a change where there will be 3 possible values for : off: Smooth caret animation is disabled explicit: Smooth caret animation is enabled only when the user moves the cursor with an explicit gesture. This should cover users who don't want it like in issue on: Smooth caret animation is always enabled\nthank you for the changes. I noticed now cursorSmoothCaretAnimation is turned on by default and can't be turned off at all. Looks like a bug.\nsee which is already fixed."}
{"_id":"q-en-vscode-0c1f822479559b7df7ccfd7f3c06f1a7b6be18df9bfa92424aa9dfb14e2e32a2","text":"We have noticed in the user study that users interact with profiles in different ways and they do not find the indication of active profile super useful. We should also be aware of all the Menus that profiles are present in right now in addition to the Profile activity bar item are confusing. Main menu Account activity bar item Gear activity bar item\nAfter discussing in today's stand up we agreed upon having its own space for profiles and therefore we wanted to show profile indicator and drop the actions in accounts and gear menus. We shall investigate on how to make this profile indicator appealing to users -\nLooks great -verified"}
{"_id":"q-en-vscode-e28ed666496e39ffbc4b31969f62cfaa9f0f541da5286ea00c808d8a2cbac672","text":"Example block quote in the extensions page. <img width=\"375\" alt=\"image\" src=\"https://user-\""}
{"_id":"q-en-vscode-ca197bcb65a18fcb1c2d44a19ac3d9c307fae31c5c8c0881303f8cab3829fecc","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.75.0-insider OS Version: Linux Hello, IMHO, the block cursor could use a different background. Default Dark+ Experimental (v2): ! ! Default Dark+: ! ! Default Dark+ Experimental (v2): with some cursor background color. For ex: ! ! ! ! Edit (one more option) (with color set to 's value, giving inverted cursor effect) ! ! Thanks"}
{"_id":"q-en-vscode-843bf7a1f1cdd297eaeab34aeeee63ea76f44b994bbe1e363163f7277df74030","text":"I was resolving a conflict for with 500 file paths showing up in the terminal as a result and noticed how typing became very slow. Kudos to auto profiling, we have a trace:\nTerminal output: !\nThis was caused by , it may fixable on the VS Code side though.\nRepro: vscode repo : over a link and keep it there into the terminal ! ! ! !\nSomething is not getting disposed, I verified that the number of times each link detector is called increases\n!\nCan repro upstream:"}
{"_id":"q-en-vscode-b6e4fee69a290bffdb841f012de480a211a127e700be60a5d289710122250459","text":"Light Theme the Extensions view the Extension item in the list -all letters turn black, the contrast is not good !\nLooks like these buttons are inheriting the codicon styling, and as a result, the color. I removed it from Light+ V2 to reset it back to null. Ideally this button wouldn't use codicon styling cc"}
{"_id":"q-en-vscode-1dc6594cccf9d207b81311ef2b399b7133107228ba73e7232babd11be6526bd9","text":"<img width=\"1754\" alt=\"iShot2022-12-2811 40 36\" src=\"https://user-\"\nEasy repro: ! ! Likely Windows only.\nIt's an encoding issue in pwsh's shell integration cwd detection\nNo, my MacBook has this problem too.\nyep the fix was for pwsh on all platforms\nUpdated 1.74.3 today, but found that this bug still exists, do you want to wait for the next version to fix it?\nthe fix is in 1.75.0 (insiders only atm)\nThanks."}
{"_id":"q-en-vscode-39be89cc3025cad97b328842caf895a0f99b01e53512e276f93a4085e04f50d4","text":"When opening as quickpick with , the latest Insiders builds only handle click events on the checkbox and not on the row item. In Stable, clicking on the row works as expected. This sounds very similar to and might have the same cause. <!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Unsure, probably yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.75.0-insider, Commit: OS Version: Windows11(.963), macOS Ventura 13.1; using the VS Code in the web Steps to Reproduce: a quick pick UI with multiple selection options Try to click a row and observe that the selection doesn't change on the checkbox and observe the selection does change This bug was observed with the Codespaces extension, but should happen for any extension/UI using multi-select quick picks.\ncan you update to the latest Insiders from today and try again?\nSure! It's still broken with the latest insiders: Version: 1.75.0-insider (Universal) Commit: Date: 2023-01-09T05:30:04.983Z Electron: 19.1.8 Chromium: 102.0.5005.167 16.14.2 V8: 10.2.154.15-electron.0 OS: Darwin arm64 22.2.0 Sandboxed: Yes https://user- Here's Stable for comparison: https://user-\nYes, more fallout from , apologies"}
{"_id":"q-en-vscode-f87b764e56af7fda90e2f1ab491c9cb168af223c177a4684c98ce8fb4b68f10b","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- Version: 1.75.0-insider (Universal) Commit: Date: 2023-01-10T05:21:34.436Z Electron: 19.1.9 Chromium: 102.0.5005.194 16.14.2 V8: 10.2.154.23-electron.0 OS: Darwin arm64 22.2.0 Sandboxed: Yes Steps to Reproduce: an existing workspace folder with git a few branches with commits to switch branches list of branches does not display the recent branches on top the same workspace folder in vscode stable and the order is as expected"}
{"_id":"q-en-vscode-a806dab862282045fea2961170f5824d084f06d4e07f53f462aec5184843bb27","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.75.0 insider OS Version: Windows 11 Version 10.0. Build Steps to Reproduce: perfectly clean VSCode Insider version any folder and create/open .md file like this: Preview to the side ! on the second title (L5) then vertical white line appears on the left ! I click on the secound line, the top title is highlighted normally ! when I directly click on the first line, the vertical line doesn't appear !"}
{"_id":"q-en-vscode-1187b902fa9f4fe0d81d6c857764a2e78453fd6697429ce68535e4b13d6811b6","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: Version: 1.74.3 (on ) OS Version: Linux, Gnome, Chrome Steps to Reproduce: an image in the media-preview extension click on the image and select copy to paste the image somewhere Nothing happens. Also note that Cut and Paste make no sense here as menu items. I assume that these are just the default context menu options from - perhaps these should be removed entirely from the preview extension and we could manually implement copy?\nHey I would like to work on this issue.\nassign this to me\nAre these the correct steps to reproduce the issue? I'm unable to verify it on my mac. Can you provide more information? I previewed the image and then copied and pasted it. <img width=\"950\" alt=\"image\" src=\"https://user-\"\nI would like to wokr on this issue.\nI would like to try solving this issue.\nYou right-click an image and select \"copy\", then you click a folder in VSCode's explorer and select \"paste\". The image should be pasted into that folder but it doesn't.\nI would like to work on this issuse. here, I tried to copy and paste (Ctrl+v) it. But it doesn't work. https://user-\nAfter a day of researching, I think I can explain why this issue happens. Basically in the core webview module (see ), three default context menu actions (cut, copy and paste) are to the menu, which are accessible throughout all webviews, and happens to be one of them, so after you right-click anywhere in a image previewer, you will be able to see these three actions. After you clicks one of these actions, the editor will post an message to the image previewer, telling it to cut, copy or paste. If there is a correspondent listener for this perpose, the image will be actually cut, copied, pasted, but unfortunately there is no such listeners defined. This problem has existed for quite a bit of time. The original modification that cause this problem is from here . It is origianlly meant to allow users to copy selected text within an extension's README page. There is also a closed issue that describe such issue in . Personally I think providing action cut, copy and paste for all webviews is a little bit hasty, since there are not universally required for each webview, and even some of them need, the subject to be copied may vary from different webviews, or even different places in a same webview. I can solve it by disabling the default actions first (there is a option to achieve this), then let image previewer to implement its own copy and cut functionality, but I think this is rather a quick and sloppy solution. There should be another better way to solve this issue.\nWhat about adding a handler :\nHere's what I tried VSC insiders without any extensions Open a folder with a .png file Open the png file in vscode Right click and select Paste into folder in vscode = does not work Paste into markdown cell in vscode = does not work What am i doing wrong?\nPasting image data into folders isn't supported by VS Code. Pasting into a notebook cells works fine for me though. To verify though you can also paste into an external image editor"}
{"_id":"q-en-vscode-96dcf093e3cdef8d6dc308f53733fe0e52f73534987c26c442df46e820eab5d7","text":"Type: <bBug</ba TS file at the language status item in the status bar Bug The bracket icon is instead rendered as a pie chart icon <img width=\"222\" alt=\"Screenshot 2023-01-18 at 1 28 37 PM\" src=\"https://user-\"VS Code version: Code - Insiders 1.75.0-insider (Universal) (, 2023-01-18T08:53:10.224Z) OS version: Darwin x64 22.2.0 Modes: Sandboxed: Yes <details<summarySystem Info</summary </details<!-- generated by issue reporter --\nLooks like the id for got duplicated:"}
{"_id":"q-en-vscode-031dac191f99d8ebfc4e8a91b424e267ec09aec1efe4696c66e0b8aa23663e2a","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --originally posted here: when using the OCaml ecosystem (in my example ) the errors can't be clicked on in the terminal: <img width=\"736\" alt=\"Screenshot 2023-01-19 at 10 06 47 PM\" src=\"https://user-\"I often can't click on the file because it isn't relative to where I am, and when I can click it doesn't get me to the line number. I'm not sure how other languages do it but I don't remember having that problem in other ecosystems. I'm also not sure how to fix it. I believe the terminal / vscode will detect filename:linenumber all the time, but the \"File \"\", line 588, characters 50-75\" is not recognized.\nyeah looks like we don't support that case atm, though this one comes close\n<!-- -96bd-47a8-a0e8- --This feature request is now a candidate for our backlog. The community has 60 days to the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our . Happy Coding!\n<!-- -c9e0-7adb-d31b- --:slightlysmilingface: This feature request received a sufficient number of community upvotes and we moved it to our backlog. To learn more about how we handle feature requests, please see our . Happy Coding!\nThe situation is pretty good on Insiders after the many link changes that have gone in since December: ! Keeping this open to also match the part and ideally select the actual range in this case.\nwhat's Insiders? EDIT: the beta,\nI'm enjoying the improvement, but the pattern still doesn't work, e.g.: Thanks!\nis that a real example or did you make up the characters part? I'm not sure what multi-line and multi-character means exactly\nForking proper multi-line support to , we should be able to support the simple case of just going to the first line easily in the one.\nThank you for your work! It's a real example, just going to the first line is good enough for me! The meaning of is \"in the selection within file F, spanning from line N1, character/column C1 to line N2, character/column C2\" (subject to off-by-one errors, I don't remember how the characters are counted).\nOh ok, that makes sense\nFYI the row/col end part of the link should work when is merged"}
{"_id":"q-en-vscode-b060b73725b28ef75b2c0526a97266009d3101d50234e8426669eb5aae620ffe","text":"Click on an error: <img width=\"595\" alt=\"image\" src=\"https://user-\"\nWhat I see when running from sources: !\nin insiders. try with \"Notification Tester\" extension.\nThe problem is that the css rule that adds the icon color looses against the color from the list. !\nFix in"}
{"_id":"q-en-vscode-d0db6a3caa0ab000093a5da0625be8e96d62772759a8d9647b2ddc51e394b92d","text":"Terminal live output is not announced. As far as I tested on Windows with JAWS and NVDA, terminal aria-live event is not auto-announced. I typed \"echo hello\" in terminal, but there was no read-aloud feedback.\nsince the buffer will now be readonly, I think this can be closed. Voice over and NVDA are announcing as I type\nAw, I think there was a miscommunication between us. The issue that I highlight here is not the input announcement. I am talking about the auto output announcement. For example, if you open an external terminal on Mac or or on Windows outside VSCode, executing any command, such as will make your screen reader read aloud the output as soon as you hit return key. The VSCode terminal was behaving like this before the a11y terminal buffer. What you have seen is that screen-reader users want this real-time output announcement per command execution just like before. This no longer works as far as I tested with NVDA and JAWS on Windows.\nThanks for clarifying. I was confused by: Now I understand\nI can repro this not working in the latest stable VS Code as well. I type and it does not read\n-- Aw, that's my bad! Sorry for confusing you. I assume that attribute got broken somewhere for the terminal a11y buffer.\nI have just spent the last 2 hours looking into the output not being read issue and am stumped . since I can repro in Stable, I don't think this is a regression. We're doing the right thing here and debugging showed the element's text content is updated to be the output on enter, but the browser /screen reader is not reading it\n-- I would try the following: if aria-live works when terminal buffer container is marked with instead of (also need to remove role = \"textbox\" for this testing). works, just add textbox role and see if it works.\nthanks for the suggestion. still does not work with that. I'm going to try to find the last commit where this was working - which will be at least a month ago\nin version 1.72 of VS Code stable, it similarly does not work for short output, but does read the message \"too much output to announce for the command for example.\nalso just tried reverting all of my changes (adding the accessibility buffer) and that does not fix it\nnavigation mode also works in 1.72\nthe element doesn't exist in insider's with screen reader mode on while it does exist in version 1.72 <img width=\"935\" alt=\"Screenshot 2023-01-23 at 2 53 35 PM\" src=\"https://user-\"<img width=\"1247\" alt=\"Screenshot 2023-01-23 at 2 53 15 PM\" src=\"https://user-\""}
{"_id":"q-en-vscode-c231bfb75a13f610033e82464976c1d47d6610a1ebb62c0529fdcbab30ba98a8","text":"Testing No matter whether I use or set it to some other keybinding, I cannot seem to invoke this command with a keyboard shortcut--only from the command palette.\nI cannot repro on mac, windows, or linux, could you pls run and provide me with the logs? The terminal has to be focused btw for it to work\nYeah the terminal was focused in my testing. Weirdly enough, seems to work now--definitely wasn't working before though.\nI've run into this as well. Here's the output when turned on Toggle Keyboard Shortcuts Troubleshooting: 2023-01-24 19:45:48.902 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:48.925 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:48.925 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:48.925 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:48.925 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.404 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.404 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.404 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.405 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.405 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.435 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.435 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.435 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.435 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.435 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.482 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.482 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.482 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.483 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.483 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.513 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.514 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.514 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.514 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.514 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.544 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.545 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.545 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.545 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.545 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.577 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.577 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.577 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.577 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.578 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.608 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.608 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.608 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.608 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.609 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.641 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.642 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.642 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.643 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.643 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.673 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.673 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.674 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.674 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.675 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.703 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.703 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.703 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.703 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.704 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.735 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.736 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.736 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.736 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.736 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.767 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.767 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.767 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.768 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.768 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.798 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.798 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.798 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.799 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.799 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.844 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.844 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.845 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.845 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.845 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.877 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.877 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.878 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.878 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.879 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.907 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.908 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:49.909 [info] [KeybindingService]: / Received keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 16, key: Shift 2023-01-24 19:45:49.909 [info] [KeybindingService]: | Converted keydown event - modifiers: [shift], code: ShiftLeft, keyCode: 4 ('Shift') 2023-01-24 19:45:49.910 [info] [KeybindingService]:  Keyboard event cannot be dispatched in keydown phase. 2023-01-24 19:45:49.916 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:49.917 [info] [KeybindingService]: | Resolving shift+F1 2023-01-24 19:45:49.917 [info] [KeybindingService]:  From 2 keybinding entries, matched , when: terminalFocus && terminalProcessSupported, source: built-in. 2023-01-24 19:45:50.063 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:50.063 [info] [KeybindingService]: | Resolving shift+F1 2023-01-24 19:45:50.063 [info] [KeybindingService]:  From 2 keybinding entries, matched , when: terminalFocus && terminalProcessSupported, source: built-in. 2023-01-24 19:45:50.483 [info] [KeybindingService]: / Soft dispatching keyboard event 2023-01-24 19:45:50.483 [info] [KeybindingService]:  Keyboard event cannot be dispatched 2023-01-24 19:45:50.484 [info] [KeybindingService]: + Ignoring single modifier shift due to it being pressed together with other keys.\nThis was on my Surface with Windows 11. I wonder if it has to do with needing to press the Windows function key to get my F1 key? I need to do Shift + Function + F1 as without the Function key, F1 is the screen dimmer. Alt+Function+F1 works for showing the regular accessibility help. I wonder what that code is doing to work?\nThis is expected as to use , the function key needs to be activated. in the terminal does funny things on Windows/Linux, which is why we went with instead here\nOn Windows, when I set the keybinding to , it's not a valid chord and the result from the keybinding service is . That's true even when I set the weight for our action to be greater than workbench contrib. It doesn't work even when I set . Thus, we need to stick with here and the function key has to be used before pressing . !\nshift+f1 isn't working for me"}
{"_id":"q-en-vscode-fd967569376c8bdeacefce2493b4222e43671953cc943f52af122cc37d7e5df8","text":"Testing Here is the description of this setting Controls whether the suggest widget becomes active when triggered via quick suggest or trigger characters. Note that the widget is always active when explicitly invoked, e.g via Ctrl+Space. It was not clear to me what active means. After trying around I figured out that its about the default selection of the first suggestion in the widget. Setting itself calls rather than . Hence I would recommend to explain in the description that this setting is about controlling the selection of the suggestion. Like Controls the selection of the options in the suggest widget when triggered via quick suggest or trigger characters. Note that the first suggestion is always selected when explicitly invoked, e.g via Ctrl+Space.\nIt's not always the first suggestion - there is yet another setting to define what element is default-\"selected\". The problem is that technically this isn't selection (because selecting inserts the item) and also not focus (because focus remains in the editor). I used selection mode because that's how VS calls it but I understand the issue and I open for a better setting name and better description. Maybe we should just talk about focus?\nI think keeping selection isn't bad but maybe the working should say \"Controls whether a suggestion is selected when triggered via quick suggest or trigger characters...\""}
{"_id":"q-en-vscode-10323f00ea42182c6e909934260dce4dac1b875da80b2bb8664523dc7a6018dd","text":"I took one of the suggested results from intellisense and it inserted a literal star symbol into my code ]()\nI also have been seeing this since yesterday, using JS specifically The star has always been present in the intellisense list, but never included in the text filled in.\nAlso reproduced this behaviour while verifying another issue !\nThis issue is caused by an extension, please file it with the repository (or contact) the extension has linked in its overview in VS Code or the for VS Code. See also our . If you don't know which extension is causing the problem, you can run from the command palette (F1) to help identify the problem extension. Happy Coding!\nCaused by IntelliCode:\nThis issue only occurs in Insiders, and does not occur in Stable. Downgrading to an older version of the extension also does not fix the issue. Therefore, I'm wondering whether the completions API implementation changed. the IntelliCode extension (VisualStudioExptTeam.vscodeintellicode). a new JS file. and wait for the IntelliCode completions to show up. If they don't show up, backspace and try again. The IntelliCode completions will have a star in front of them. one of the IntelliCode completions. :bug: In Insiders, the star character is to the editor.\nI've disabled Microsoft Intellicode (AI-assisted development, ID: VisualStudioExptTeam.vscodeintellicode) until they fix it back. All good for know.\nThis is somehow happening in the TypeScript extension. In the UI, when inserting the completion, we have an insert text with a leading . So, the UI is doing what it is told to do. <img width=\"1403\" alt=\"Screenshot 2023-01-27 at 09 30 21\" src=\"https://user-\"Seemingly, in the TS extension the right data is received but somehow a wrong API completion is generated <img width=\"628\" alt=\"Screenshot 2023-01-27 at 09 34 39\" src=\"https://user-\"\nMostly likely regressed via\nConfirmed. This is the faulty statement The OR-statement binds stronger than the ternary operator and therefore the label always replaces the insertText"}
{"_id":"q-en-vscode-677a06efac0f276e33f42f3f0e6296060c2cf5a03933492136801ec8db10519b","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.75.0-insider OS Version: Windows 11 (21H2, .1455) Steps to Reproduce: editor with CSS/SCSS syntax CSS block e.g. to use Emmet CSS completion with abbreviation of more than 1 character that no Emmet completions are provided Emmet completions seem to only be provided for single-character abbreviations: ! For multi-character abbreviations, I am not getting any results from Emmet at all: ! A of the issue suggests that it was introduced in commit as part of More testing shows some weird behaviour when providing completions, with them showing sometimes but not always when expected: Type - text is - Emmet suggests Type - text is - no suggestions Type - text is - no suggestions Type - text is - Emmet suggests Backspace - text is - Emmet suggests Backspace - text is - Emmet suggests"}
{"_id":"q-en-vscode-5e166690d63c4b0c575e4de6106a94dcb1a58be52bae29a178963b9d7f67a963","text":"off the hot exit and auto save settings an interactive window, run a cell vscode :bug: confirmation pop-up to save the IW"}
{"_id":"q-en-vscode-68e7bc6834d333497bcbdc92c50dffcf0fe0e4c02c95682ba79abeef4c62c84b","text":"In virtual workspaces, the file dialog has a button. There's no such dialog in remote windows--the file dialog always shows the remote filesystem. This is an issue for continuing from a GitHub codespace to a local desktop clone when the user is in VS Code desktop. Suggestion: when in a virtual or remote window on desktop, always have the 'Show Local' option in the file dialog.\nOh interesting--there is a Show Local in the Open Folder flow, but not when the git extension invokes the file dialog: <img width=\"671\" alt=\"image\" src=\"https://user-\"\nI see this was an intentional decision for extension access to the simple file dialog in a remote window in for However that makes it hard to go from a remote window to a local one from extension land. Is this something that we can expose to extensions via a new proposed api option for the file dialog?\nDuplicate of , which is an API request to expose \"show local\".\nReopening this to track adoption of in the Continue On flow.\nSteps: Open a codespace in VS Code desktop From the remote indicator, choose Continue in Local Clone Ensure that the repo gets cloned successfully to your local machine"}
{"_id":"q-en-vscode-8d52cfbaa433d11f7535f7310cb399d634bc640d2011209449e7c92033b9b822","text":"Type: <bBug</btypescript server keeps crashing after the latest update VSCODE ! I can reproduce this with all the extensions disabled. LOG: ! VS Code version: Code - Insiders 1.76.0-insider (, 2023-02-03T05:23:17.202Z) OS version: WindowsNT x64 10.0. Modes: Sandboxed: Yes <details<summarySystem Info</summary </details<details<summaryExtensions (11)</summaryExtensionVersion ------ ng-template15.1.0 atlascode3.0.1 vscode-markdownlint0.49.0 vscode-eslint2.2.6 githistory0.6.19 gitlens13.2.0 prettier-vscode9.10.4 vscode-yaml1.11.0 sonarlint-vscode3.14.0 code-spell-checker2.16.0 vscode-stylelint1.2.3 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nPlease try collecting the TS Server log from the crashed instance: VS Code and reproduce the problem the section of the output panel the very top, find the first line that prints to path to the semantic error log file. It should look something like: That file contains the typescript logs. Look through that log file for the first error or stack trace you see. If you can share the log, I can also take a look to see if anything stands out\n, for some reason the log file location is shown as [Object object] (as pointed out by the pink arrow), I checked in the folder pointed out by the red arrow and the folder is empty. ! I have checked this for multiple folder I could see in the log prompt but all of them are empty.\na Microsoft member created same/duplicate issue , in case this could help with the debugging\nConfirming that I also see when using verbose setting and all the log files are empty for me, which is also noted and shown on microsoft/vscode.\nFor me the directory under the object object is also empty, is that the case with you?\nYes, same.\nshould be fixed in the next insider release: microsoft/vscode\nthe object object issue is resolved, however, the file pointed by the log still does not exist. can you give it a try if you can see the logs?\ni did some commenting in the file and found that commenting out setting stops the repeated crash. Note: I am not sure why I have /enabled this setting in the first place.\nOh wow, exactly the same for me. I also do not remember ever setting , but it was in my and removing it fixes the crash for me too!\nAlso confirming that with latest update is fixed but log directories are still empty. I've closed my bug as a duplicate of this one since it's pretty clear these are the same issue now. Are you able to repro using is Also do you know if this would ever have been set on user's behalf, which could impact the severity here if lots of people would have this in their settings?\nThanks and I can confirm that enabling cause the server to crash on windows. Oddly I don't see it on Mac. The server is likely crashing before logging can start too, which is why the log files don't exist I'm not sure how this setting got enabled for you. It's only meant for detailed performance investigations and should never be turned on automatically\nYay, thank you and"}
{"_id":"q-en-vscode-6d54c3215fd52e53057e440cc30c5d5f00e312cce7b969c2c99c286108d26cf5","text":"Bug The the following entry in its : This seems to cause the python extension to be activated whenever a json file is opened, even though the extension does not provide any json language support This was possibly caused by the move to automatic activation events\nI think my initial implementation of implicit activation was perhaps too naive. There seem to be a few issues: It's possible to have a language configuration extension that doesn't contribute any code, and we would now attempt to activate such an extension (this is a noop if there is no main/browser entrypoint in , but results in unnecessary activation if all that's required is a language configuration) the bigger issue is that an extension which contributes extensions, filenames, filenamePatterns, firstLine, or mimeTypes for a language will now get activated anytime the language appears I think it's maybe reasonable to preserve the existing implicit activation behavior if an alias is defined for all of the others, we shouldn't infer , or we should perhaps keep track of the filenames/filenamePatterns etc and activate the extension with the language contribution matching the specified pattern? I know of at least two extensions which are impacted by the current behavior MakeCode Python\n:+1: This makes good sense to me. We will infer only for the one which is the \"main\" language definition, and the best evidence of that is the alias property. I wouldn't do anything beyond that (i.e. tracking filenames, patterns, etc.)\nWould the presence of the property be better evidence of a contribution being the main language definition?\ninjection based syntax highlighting can still have a configuration file"}
{"_id":"q-en-vscode-a1c99d95745e67e4b48bdb8017e023e8c4d67e4a3209690a69bb8d5db159952d","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.75.1 OS Version: MacOS Ventura 13.2 (22D49) I'm using VS Code to debug an attached Node app. The problem is that each time that I restart the Node app, new pseudo targets are to the \"Select Debug Console\" dropdown of the debug console. Problem started in 1.75.0. In prior versions, only the \"Server\" target would appear after restarts (I'm debugging in multiple target environment, so the dropdown is always visible). Since 1.75.0 (and .1), two new targets are to this list (webpack + ) after each restart. Here's how this dropdown looks like after 3 restarts: <img width=\"431\" alt=\"Screenshot 2023-02-09 at 12 39 52 PM\" src=\"https://user-\"A similar issue happens when attempting to debug a React app - each reattach adds a new pseudo target to the list. The targets that are to the debug dropdown are \"stale\" - switching to one of these pseudo items shows an empty debug console. Steps to Reproduce: debugging a JS app that is connected via the attach debug mode the debug console the app using the \"restart\" icon on the debug panel on the Select Debug Console dropdown, and observe new pseudo targets have been to the list (and on each subsequent restart) I tested the same configuration with VSCode 1.74.3, and the problem does not appear.\nI noticed this, I haven't looked into it at all but wondering whether it's related somehow to\nThis is a pain, I often restart my server process and I just see this slowly building list of old server processes. I don't need to keep their logs behind, it just clogs the doprdown. <img width=\"309\" alt=\"Screenshot 2023-02-22 at 8 48 24 pm\" src=\"https://user-\"\nignore my last comment... I found the setting Debug › Console: Close On End was 'off' by default. Turning it on resolved the problem\nYea this looks expected since we have a process ID as part of the session name, so they get unique terminals. Could argue we shouldn't... but we do, and I've found it handy on more than one occasion\nThe problem persists after turning on Debug › Console: Close On End (on version 1.75.1). When the said option is on, the terminal/console pane indeed closes when debug stops, but when I restart the process, the pane re-shows with the \"select debug console\" dropdown populated with the terminated processes. In any case, the problem surfaced in 1.75.0 so I still think there's a bug (or undocumented fix) that contributes to this issue. May I suggest to re-open the issue?\nfollowing up - this is still an issue and there's no fix nor a workaround. Can you please re-open the issue so that someone can give it a second look?\nYep, thanks for the ping. This was hanging out in my notification queue but forgot to reopen it"}
{"_id":"q-en-vscode-7fea204166958b4d06550a1334b0481916acdcf2c8ac7417270e2addcdf68fdb","text":"This is a legit error about a bad URI that came from a debug adapter but I guess we shouldn't send telemetry"}
{"_id":"q-en-vscode-264d912b505a074aa81e978455fa20a2d29dee3486c7e63af91fb024ed0ced49","text":"Shell integration offers many improvements to the terminal experience and makes it more accessible via commands like . reported that he and many others use which does not have shell integration. Should we include a hint in the help menu that the experience could be better when we detect a screen reader & ?\nI think this will improve the discoverability! Voting for this idea.\nOn windows, Create a terminal - via for example. run . Verify that you see a message about migrating to use powershell."}
{"_id":"q-en-vscode-4b570fd5d2243b92601d3d3097bb7aed72d75cc268154861f06a2c05a706cd75","text":"In our menu, we should mention setting\nI actually think we should more broadly mention how to find accessibility settings here -"}
{"_id":"q-en-vscode-7864defa3c4002e55525a8e5df98722442be9b9007db2106bd069c3c28118b71","text":"It should be outside of the output region <img width=\"962\" alt=\"image\" src=\"https://user-\""}
{"_id":"q-en-vscode-7cea5538de366b50ffbed225d86acdb960f0ceb68e95311592cd5f8601e5678a","text":"Testing have search view open drag the explorer icon onto it the explorer (and npm script) views show the explorer icon stays in the activity bar, the view is empty https://user-"}
{"_id":"q-en-vscode-62de6bf46835644b07dc9b38de6902e89ab63a3cdb325bfb7df5730727ec0ff8","text":"Testing Have the following code in an notebook cell. Run the code Scroll the output. Notice as your scroll reaches the top you can continue scrolling and it will scroll the whole notebook file. This is unexpected and I think no other widget in VS Code behaves this way. I suggest instead the following strategy when scrolling: If a mouse is in the Cell then scrolling it will only scroll that cell. You can not scroll the large notebook while the mouse is in the cell. Same way you can not scroll the Cell with the mouse unless the mouse is in the cell IMHO this is what makes the current experience feel weird.\nWe don't often have nested scrollable elements in VS Code workbench and also this is how browsers handle nested scrolling. Current behavior feels right to me. I can see the value of scrolling only when the focus is inside/on the output, there is a catch with that though, users need to move cursor to outside of the cell output to scroll the notebook, and the hit area for that is very narrow most of the time.\nThanks for providing more info. But what other UI element scrolls like the notebook does right now - so that you can scroll the element until it hits the end, and then if you continue scrolling that the outer element starts scrolling? This really felt unnatural to me, but I might have missed this pattern in other parts of the workbench. Or what do other Notebooks products do in this case?\nthe peek view actually behaves this way. That's the only other nested scrollable element that I know of.\nThanks! If I am the only user that it complaining feel free to close this one as designed :)\nI tried using notebook in VSCode again last week, and came to GitHub to open this exact issue. I feel this scroll is extremely unintuitive. Here's an example. A cell is outputting more lines than , so we're currently viewing it's last line. I want to scroll to top. I speed scroll up, but this reaches the top and keeps scrolling the notebook, so now I'm all the way up. I have to scroll back down to my cell. If I go fast, I might miss it. If I go slowly, some other cell's output might catch my scroll and scroll that to bottom, before the notebook scrolls again. Same happens to my cell. I need to be frustratingly precise to scroll this output to the top. Here's a recording to illustrate (I hope you can see it, recorded on MacOS, encoded as AV1; can add YT link if not viewable here): It took me 15 seconds to manage to get that top. And I knew exactly what I wanted to achieve, for the sake of the video. Now here's the same experience in native Jupyter Notebook (exact same notebook and cell). First, I can easily toggle the output scrolling. If untoggled, the output scrolls natively with the notebook, and since it expands downwards, without the notebook jumping, I can immediately see the top, which is right next to the cell. Then I toggle the scrolling again, to illustrate the behaviour compared to VSCode. I initiate scrolling inside the output area, and so only that area is scrolled. No matter the speed, or direction, I can easily reach the top or bottom as many times as I need, without losing context of my notebook. If I do want to scroll the notebook, I need to scroll outside the output area. This is easy because the cell does not take the whole width. This is in relation to what said above: Yes, but the cell does not need to take the complete visible width. In native notebook, cell's width is limited, and there's still plenty of space for code. But even if the cell does take the entire width, the user can initiate the scrolling outside the output area, i.e. over a cell. In this case, it scrolls the notebook, and never the outputs. The outputs do not steal the scrolling from the notebook, like it happens in VSCode. So: scroll initiated on cell or outside the cells area -scrolls notebook only, never output scroll initated in scrollable output area -scrolls only that output area, never the notebook\nshould we reopen this? What do you think?\nsure, that's a reasonable request. It would be inconsistent with how the peek view scroll works, but the possibility for many scrollable regions makes these cases more noticeable in notebooks. - one thing that might help - you can toggle output scrollability per output in vscode as well with , by default\nI have the same issue. The usability of the scrollable element suffers greatly due to the current crazy behavior - when the scrolling inertia reaches the end of the scrollable output and scrolls the whole notebook! If different users want different behavior, this can be an option in the settings, e.g.: current behavior behavior like in the native Jupyter Notebook (in the video above), i.e. the notebook never scrolls if the pointer is inside the scrollable element there may be a 3rd option: the scroll that reaches the end of scrollable area doesn't scroll the notebook, but the subsequent scrolls in the same direction (even if the mouse pointer is inside the scrollable area) - scroll the notebook\nfor verification, please check with a mouse wheel and scrolling with a trackpad, since they behave quite differently. Hitting the boundary of the inner scrollable region will prevent further scrolling for that same scroll action. After a short pause you should then be able to scroll the outer region since the inner region can't be scrolled any further. (similar behavior to a browser)\nVerified on trackpad. Still have to verify with a mouse wheel.\nverified on mouse! there was a pause/delay, but wonder if there should be a larger delay - I find that sometimes I spam scroll and it moved up a little quicker than I anticipated. if it matches browser tho, should be all good ~\nhello , thank you a lot for the work on this issue! I hate to say this, but I think it's not solved: It looks like only a certain \"scroll length\" is ignored, and the rest of the scroll inertia is still applied. Here I do: one fast scroll (as you see, the notebook moves) then I scroll back down then one fast (a bit slower) scroll (the notebook moves) Only if the scroll isn't fast enough the \"scroll length\" will be small enough to be completely ignored. <br /Here I do: one long scroll - all the way until the end of the video: <br /The behavior in the native jupyter notebook is certainly different. Here I do: one fast scroll - different than in VSCode (the notebook doesn't move) one long scroll- different than in VSCode (the notebook doesn't move) one after-scroll (the notebook moves as expected) Could this issue please be re-opened?\nwhich version of vscode was this on?\nVersion: 1.88.0 (Universal) Commit: Date: 2024-04-03T13:28:18.899Z (1 day ago) Electron: 28.2.8 ElectronBuildId: Chromium: 120.0.6099.291 18.18.2 V8: 12.0.267.19-electron.0 OS: Darwin arm64 23.3.0\nok, yeah. looks like I broke the behavior with a related bugfix\nVersion: 1.89.0 - still isn't completely fixed: If I scroll hard enough, the notebook scrolls."}
{"_id":"q-en-vscode-b51fa4b5edf4cbb18338595b53abe51b8c5d4642f288a0e7227b94d14a513eef","text":"Testing into the editor. status bar still says \"Tab Moves Focus\". I think I'd only expect it to say that when I'm focused in the terminal.\nIf you toggled it to on in the editor before focusing the terminal, then this is expected. We remember the toggled state for each context until the window is reloaded and it gets reset to the configured value I realize that's kind of confusing and was not how I initially implemented this\nI never toggled it at all, this was just using the settings."}
{"_id":"q-en-vscode-744c330ea245c8d0a9e32c6ea7681487e3ff21f7dbc7eb2c50fccf4286194969","text":"has indicated that the diff inserted / modified audio cues are hard to differentiate and that the sounds are not as fast / singular toned as they should be\nwhat do you think of this?\nI agree with the suggestion. At least for me it is very difficult to identify the difference between the sounds used.\nSounds good -- will start on this asap.\nI don't hear the cues as often anymore. I also noticed that when using up/down keys on the right diff editor, no audio cue shows up for a deleted section. Example with .gitignore file: Left: Right:\nI can reproduce\nit's getting to the line I expect it to (see logs) but no sound is produced - thoughts? https://user-"}
{"_id":"q-en-vscode-33df85bb83bbf91c35af74a09c41a5273693d10a1d53e8b2955000e6f5a1793a","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.75.1 OS Version: macOS 13 Steps to Reproduce: the following settings to : suggestions still show up: <img width=\"766\" alt=\"image\" src=\"https://user-\"\nwhat you are looking at is the hover. to disable that use this setting\nHi, I want to remove the three dots, disabling hovering doesn't help, these two settings , now they don't, a recent update broke that, please reopen this and take a look .\nOh, you mean \"hint\" diagnostics. Forwarding to for comment\nThanks for fowarding, I got the term from the : !"}
{"_id":"q-en-vscode-104dea625e402c73d732441cd028d1f5ba47926ac357fc12f5ea076001c6db82","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: Windows 11 OS Version: 1.75.1 Steps to Reproduce: Using the following settings: ! ! a typescript file within a method/function following gets output (might have to navigate through some of the suggestions) even though snippets are off: !\nThanks for the patience. I cannot reproduce, please share more of your settings that might influence this. Also share data in a copy-paste friendly way (text over images)\nHere are the settings that I am using to reproduce (all extensions disabled ): Once typing and going to the 3rd suggestion, I get this as a suggestion: !\nThanks for clarifying, looks like we aren't passing the filter here:\nI can't figure out how to get that ghost text to show up for suggestions to verify this. I've never seen that ghost text for anything but copilot completions Do I need a suggestion provider extension?\nYou are right. I made a refactoring after this fix and that broke this feature Thanks for the diligence on this one!\nTo verify, use the steps from Make sure no snippets completions are ghost-texted with and that they are when the setting is set accordingly\nI still cannot get snippet completions to show up\nYou need to say what you are trying because I do see them working just fine <img width=\"922\" alt=\"Screenshot 2024-01-26 at 18 14 32\" src=\"\"\nThis works correctly for me now, no snippets with , and they do show up without it.\nBang on. The whole point is that the -setting is honored also for ghost suggestions\n: <img width=\"1792\" alt=\"Screenshot 2024-01-26 at 9 19 47 AM\" src=\"\"\nthis is about ghost text, please see and the comment linked from there\nthere was a syntax error in my file, so the settings did not get applied. Now it works"}
{"_id":"q-en-vscode-8930793755f1c4ae56cf6cb249bf4fe8c31f9ccea886a086e4b063576ddd19f8","text":"With new API type was but that type isn't used anywhere and looks like a leftover. We should remove it before releasing stable.\nfyi\nMy bad, I must have it and forgot to remove it.\nThanks for the quick turnaround"}
{"_id":"q-en-vscode-13951b5ada64e89fd36815eec81d9376eab029c6930e06f4e854d862ddf89fa5","text":"With Dark High Contrast theme, everything is clear except for warning and errors... Error is shown with an orange x: ! While warnings are exclamation mark in blaring red: ! I mean there's no two ways about it... RED is error, RED+certain amount of yellow(orange is fine) is warning."}
{"_id":"q-en-vscode-6ad2c481df3ee90235717ba42568f7d54f5c80d87266e062a0ed14b5a8867f25","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.76.0 OS Version: macOS 11.6.2 Version: 1.76.0 Commit: Date: 2023-03-01T10:23:45.993Z Electron: 19.1.11 Chromium: 102.0.5005.196 16.14.2 V8: 10.2.154.26-electron.0 OS: Darwin x64 20.6.0 Sandboxed: No Summary When creating a profile with a name such as \"Typescript / Javascript\", the following happens: When exporting as a Gist, no result dialog shows (\"there was trouble\" / \"created profile successfully\") and a validation error is outputted in the logs When exporting to a local file, the text after the last forward slash is set as the filename. With the above example, it'd be \" \" To reproduce: a profile: Settings cog -\"Profile ([...])\" the profile: Select \"Export Profile...\" -\"Export\" the \"name the profile\" step, enter a name with a forward slash in like so: \"Typescript / Javascript\" either method for exporting: gist or local Expected results: If you select local: a system save prompt with the name you entered pre-set in the filename field If you select gist: A success dialog with a link to the website prefilled with your config Actual results: If you select local: only the text after the last forward slash is in the filename field If you select gist: No dialog shows. The following error is in the \"output\" window, under \"Window\": I can only test this on a macOS device, I haven't tried this on a Windows or other Linux device and I feel like (with the area in mind - filesystems etc) these issues could manifest differently or not even be an issue there. I'm assuming validation on the inputbox for the name will be required because filesystems (and Gist? not sure) don't typically support forward slashes in file names. Thanks and if you're reading this, have a great weekend :)\nGists do no accept in the name. Replaced with ."}
{"_id":"q-en-vscode-fc8ae37aef3d5895ce7e62b4d022acccad1a66b1bb7e9130e4828ff78c72cb4c","text":"When fails to get a mac address it still returns a generated mac id, but uses to log the error. That's a bit risky as it rely on VS Code to already have installed a . The default handler throws an exception in a timeout that can crash the process. (see issue ) IMO the better way would be to pass a logger to , e.g. the -move to -add to\nif you support the idea I can provide a PR\nthat all sounds fine to me, thanks for looking into this. I didn't know this could cause a crash since it had been in desktop for so long."}
{"_id":"q-en-vscode-2595a45cae82741b7de76961e5852ab571ecb3f4f8dbe5fc83d0fb60aef06852","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --Is it possible to have a setting that would automatically replace the local tags rather than always prompting? Thx\nEven, I am also looking for a similar functionality.\n<!-- -96bd-47a8-a0e8- --This feature request is now a candidate for our backlog. The community has 60 days to the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our . Happy Coding!\n+1\nWe are now \"moving\" a tag to track the last successful build in our CI/CD process for users who don't have a lot of CI/CD or Git experience. We want them to naturally see the tag slide up as the team makes changes and the server succeeds in processing them. It's very jarring to have this modal pop up and have no way of hiding it.\nTomorrow's (2024-02-14) VS Code Insiders release has an updated dialog that has an \"Always Replace Local Tag(s)\" button. Pressing the button will toggle the so if you wish you can proactively enable this setting in your users/workspace/folder settings. Please give it a try and let me know if you run into any issues. Thanks! <img width=\"265\" alt=\"image\" src=\"\""}
{"_id":"q-en-vscode-8e01bf6d9d7366c304d6d9b9d5d612800c785fcaa5a509b424cb0567abc225e0","text":"Look at telemetry and if not a lot of users are using the Install and Do not Sync extension action remove the action to simplify the Extension Install flow. is the action name. fyi\nOk was run on 59K distinct machines. For comparison extensionGallery:install was executed on 10.4M machines. So Do not Sync is used in 0.5% of cases. I think it is fine if we remove it - we keep it somewhere but not so prominent. what do you think?\nAny suggestions?\nWe already have the \"Sync Extension\" check in the Extension context menu when it is installed - this is great. ! And we already have the \"Sync Extension\" check in the same context menu when it is not installed. Just currently if I click this, there is no effect. So ideally the experience would be that I can uncheck this before installing the extension. And thus we need no new actions. !\nLets try this in insiders already and see if we get any hit back.\nAwesome - thanks!"}
{"_id":"q-en-vscode-9dc824441c4cd0f1c561e9981ecf3ab8dec13949c3d0dd672abea8ffd80f249f","text":"repro: in settings notebook search filter to look at outputs all editors for something that only appears in outputs on the result and wait for the results to convert to rich notebook results list focus has shifted !\nHi - I'm attempting to replicate this bug, but cannot generate the output from a notebook because my build (Code - OSS Dev) crashes when I select a kernel for my notebook with the message . I am getting the extension form How were you able to get the Jupyter notebook running in your build?\nI installed the Jupyter extension via the vsix that is in the marketplace install page.\nThank you for the reply! I still run into the same error after installing the extensions from vsix when I select the kernel. Console output\nHmm, I'm not sure why this would happen. Since this is not really related to this issue directly, could you file a new issue?\nI believe the issue is related to this line in folderMatch.allDownstreamFileMatches() is returning a set of notebook files that excludes the parent file of the current selection."}
{"_id":"q-en-vscode-73c37e0f1ee0c5bf67111db26ab9ad711dbc67380439fd93d72c2a2e99dd43de","text":"a notebook and run a cell that produces an output long enough to scroll, e.g within the output element :bug: the output element scrolls to the top to select the first link"}
{"_id":"q-en-vscode-e40b6524ed9fdcd78f57d7149e0bc311c0e8a945760391672fda1d8ee4cc1ff5","text":"Testing Up / down, go to top / bottom always targets the editor's scroll position, not the cell's scroll position. (Related: )\nIt actually only navigates the input cells. I can then tab into the input cell, but not the output cell. I think for accessibility there would have to be a way to focus the output cell, so that can be scrolled too?\nI also see this, page up/down, ctrl+home/end act strangely\nWe didn't update when clauses for arrow keys and page down/ups so when the focus is inside the scrollable output, it scrolls both the output and the editor."}
{"_id":"q-en-vscode-79437a7c3a99b97b275e43220f32a59e7a837c303ce73a2efa170de38b78a303","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.76.2 Commit: Date: 2023-03-14T17:57:21.103Z Electron: 19.1.11 Chromium: 102.0.5005.196 16.14.2 V8: 10.2.154.26-electron.0 OS: Linux x64 5.19.0-38-generic Sandboxed: No Steps to Reproduce: a git diff view and delete one of the lines. the right pointing arrow in the diff view. https://user-\nI cannot reproduce this. Do you see any errors in developer tools? Does this reproduce with any file and with any change or is it specific to this particular file? https://user-\nI can't reproduce it anymore either. I just updated to 1.77.0. Maybe it just happened to get fixed in that version? I can still reproduce though.\nHmm no I can still reproduce it in 1.77.0.\nFound it! Seems like the issue only occurs after line 99\nHmm and another requirement is that the file has 1000 lines in total or more I think.\nI cannot reproduce. I tried with a file which has 1353 lines and I do an edit at line 600. Do you see any errors in Developer Tools (Help Toggle Developer Tools)? https://user-\nI don't think the devtools console contains anything of interest: WARN Via '' extension 'ms-' wants API proposal 'contribEditorGutterMenu' but that proposal DOES NOT EXIST. Likely, the proposal has been finalized (check '') or was abandoned. WARN Via '' extension '' wants API proposal 'contribEditorGutterMenu' but that proposal DOES NOT EXIST. Likely, the proposal has been finalized (check '') or was abandoned. WARN Via '' extension 'ms-toolsai.jupyter' wants API proposal 'notebookControllerKind' but that proposal DOES NOT EXIST. Likely, the proposal has been finalized (check '') or was abandoned. ERR Extension 'github.copilot-nightly' appears in but enables LESS API proposals than the extension wants. (LOSES): inlineCompletionsAdditions, interactive (WINS): inlineCompletionsAdditions INFO [perf] Render performance baseline is 62ms I've found that the line 100 rule doesn't completely hold for me either, when make the right panel very big it becomes clickable again.\nSo far my hypothesis is that you need a long word wrapped line above the one you're trying to reset. But honestly I'm not sure anymore at this point. https://user-\nHooray, this reproduces when using word wrapping! The root cause is that some properties on the mouse event are not converted from view coordinates to model coordinates. This reproduces with the following word wrapped .txt file:\nSeems like I can still reproduce (with and w/o word-wrapping): https://user- https://user-\nI ran git bisect and I had to skip a few commits due to build problems, but I believe this is caused by The first bad commit could be any of:\nSo merging the PR will make the arrows work but will also bring back the issue of also setting breakpoints? That is unfortunate but I think I'd rather have the arrows working.\n:+1: I think it's better to go back to the state we had in 1.77 and the risk is minimal as it is a simple revert."}
{"_id":"q-en-vscode-704fcd20e716d777ebd1ef37b2a7ff0dff13bdaec1f7de37d146e0eb9f0eca7d","text":"It seems the extension's raw is used for the text here. To combat this I had to restate my extension's name in the extension description, which probably isn't the desired approach. I think ideally the extension name would be featured prominently, and the extension description would be below it or something. <img width=\"1040\" alt=\"image\" src=\"https://user-\"<img width=\"777\" alt=\"image\" src=\"https://user-\"Demo:"}
{"_id":"q-en-vscode-12c44807288efb7a18784394a3ec1d113c8d5eb29abadf9481002b7899c243be","text":"Type: <bBug</bdark v2 theme files from the explorer and drag then in to an editor bug There's no background to the dragged element label so it blends into the editor text ! VS Code version: Code - Insiders 1.78.0-insider (Universal) (, 2023-04-17T12:50:42.586Z) OS version: Darwin x64 22.4.0 Modes: Sandboxed: Yes <details<summarySystem Info</summary </details<details<summaryExtensions (51)</summaryExtensionVersion ------ tsl-problem-matcher0.6.2 comment-tagged-templates0.3.2 docs-view0.0.11 emojisense0.9.1 folder-source-actions0.2.1 markdown-mermaid1.18.1 vscode-eslint2.4.0 kusto0.4.3 typescript-notebook2.0.6 EditorConfig0.16.4 codespaces1.14.5 copilot-nightly1.82.24 vscode-pull-request-github0.63. vscode-drawio1.6.6 vscode-postfix-ts1.12.1 search-editor-apply-changes0.1.1 latex-workshop9.8.2 vscode-tree-sitter-query0.0.1 vscode-position1.1.2 template-string-converter0.6.1 vscode-docker1.25.0 vscode-language-pack-de1.77. vscode-language-pack-zh-hans1.77. isort2023.9. python2023.7. vscode-pylance2023.4.21 jupyter2023.4. jupyter-keymap1.1.0 jupyter-renderers1.0.15 vscode-jupyter-cell-tags0.1.8 vscode-jupyter-slideshow0.1.5 remote-containers0.290.0 remote-ssh0.101. remote-ssh-edit0.84.0 azure-account0.11.3 cpptools1.15.2 hexeditor1.9.11 js-debug-nightly2023.4.1317 live-server0.5. remote-explorer0.5. remote-server1.1. test-adapter-converter0.1.7 vscode-github-issue-notebooks0.0.128 vscode-js-profile-flame1.0.5 vscode-marketplace-extension-browser0.0.7 vsliveshare1.0.5857 code-spell-checker2.20.4 ayu1.0.5 luna-paint0.16.0 sort-lines1.9.1 fast-arrow1.0.8 (2 theme extensions excluded) </details<!-- generated by issue reporter --\nI believe the background normally comes from this line:"}
{"_id":"q-en-vscode-93428c09150e33551d9e3374851265fddbaf87abd8d110f1f8b0ae0ec694836e","text":"This throws the exception below. This happens when logging objects without prototypes, e.g . Instead of the property access should be used"}
{"_id":"q-en-vscode-e94eb279bb36621a7c3a24e185aa4a067cf39142079ff34b2dcdd99acc75d2c4","text":"Open a file with multiple changed regions in git Move cursor inside/next to some change Run \"Show Next Change\" (alt+f3) It reveals and shows a different change- I think it always shows the last one in the editor. It should show the next one after the cursor position I can press it multiple times to cycle through changes, and that's correct, but it always starts with the last change\nI'd like to fix this for April, but it looks like the change could be risky. I will fix it for May.\nFound a low risk fix.\nThis works, thanks! But one thing that my muscle memory is unsure about, did this command previously move the cursor? I'm noticing that if I press it multiple times, then press escape, the view moves back to the place where I started, when I'm expecting this to be like a \"navigate\" action. I can open a new issue, or if it's always worked like this, it's fine\nI went back to 1.74 to check, and it didn't move the cursor there. \"Go to Next Change\" will move the cursor though."}
{"_id":"q-en-vscode-cace4044cc91e38b64563a0bc5ba25beee08bad88c04eefce72174ab1ed2f086","text":"Testing !\nLooking through the code, it seems that it comes from returning multiple from my for the search editor. The problem is with the functionality that allows certain lines to be underlined for mouse definition hover. Instead of underlining the lines individually, it underlines the whole range starting with the start of the first range and ends at the end of the last range. cc\nI think the underlines are by the go to definition feature and not by the link provider. Do you think the API is used incorrectly or is there an issue on the UI side?\nThe UI receives two locations for the definition request. To draw the underline it merges both into a combined range. Looks like it doesn't check if the origin ranges actually contains the position at which the request has been made - I am unsure if that's on purpose or not. So, there might be something in the UI side to fix but the extension also return bogous data. It looks like it return all results when being asks for the definition . That's not how it is intended to work and either a the file itself should be used or no definition should be provided <img width=\"1461\" alt=\"Screenshot 2023-06-13 at 15 32 19\" src=\"\"\nI'm not sure if I understand this correctly. What is the position at which the request had been made and what would be wrong with this being the origin? Makes sense. I haven't actually editing anything in this extension before, so I'll take another look at what was intended when it was first implemented.\nI'm taking another look at this and I have a few questions about the API. I think that the original intention of using a definition provider was this: so that double-clicking the filename in the search editor would pop up references to the code snippets that are directly below it. Given this, I'm wondering: If I want multiple references listed like this, do I get the provider to return more than one and they would all have the same ? What is the difference between the and ? I read the docs and I don't know the difference between the full target range and the span.\nThe has two ends - the origin at which it starts and its target to where it goes. The origin defaults to the word at the current position (cursor) but can be refined if that isn't good enough. The two target ranges are the whole target and its focal point, think of a function where the whole function (comment, signature, body) is and its name is - when navigation a definition link the cursor is placed at the target selection range but the whole targetRange is reveal and/or highlighted etc."}
{"_id":"q-en-vscode-30e8fafc0fdf52ff2a56459be98f72faaef60b1c63f58b313a904ce6659cf485","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --. However, it is very long and explicit and thus decreases readability: I would like to request a new variable, , that would be its shorter alias that also stands out visually due to no alphanumeric characters: This is inspired by .\n<!-- -96bd-47a8-a0e8- --This feature request is now a candidate for our backlog. The community has 60 days to the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our . Happy Coding!\n<!-- -c9e0-7adb-d31b- --:slightlysmilingface: This feature request received a sufficient number of community upvotes and we moved it to our backlog. To learn more about how we handle feature requests, please see our . Happy Coding!"}
{"_id":"q-en-vscode-676b9494f037cb52d5bfb14a80f715fc8b7083909cdfd610c7cc34e2f7cd9227","text":"Testing I'm not clear on what unforwarding a port means so this maybe expected, but we do detect that the process is no longer running but it still appears in the list: !\ncan you please provide the steps you used to get here? If port forwarding is properly set (setting + reload) then yes, the port should be unforwarded and no longer appear in the list.\nRun the above Python script in integrated terminal when using Remote-SSH, it should start appearing in the list along with process ID under \"Running process\" column. Kill process using in a separate terminal.\nAlso confirming I made sure this was done as well.\nThank you for the steps!\nWhen verifying, please test that the port is unforwarded with both and are used.\nThe port is not unforwarded when is used. It is however unforwarded when is used. I'm testing with the vscode-remote-try-node repo and the default launch config.\nThe fix works for me though, with the original Python script I provided.\nI'm not able to reproduce this. It's possible for what you describe to occur if the process is terminated before we have a chance notice that it's running. If you see this again can you please set the log level to debug, repro, and share logs from Remote Extension Host and Window? Closing again for testing.\nI am unable to reproduce the issue with vscode-remote-try-node on Codespaces or Dev Containers, though it did take a bit longer for the port to close on Codespaces. Marking as verified."}
{"_id":"q-en-vscode-11deda212084da3276f17e7ae5973646a969f209612445eebe6e665861aa6f8e","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Codespaces Visual Studio Code - Insiders Version: 1.78.0-insider Commit: Date: 2023-05-02T11:17:01.989Z (10 hrs ago) Browser: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/112.0.0.0 Safari/537.36 Edg/112.0.1722.68 Steps to Reproduce: a GitHub Codespace from an existing repo the \"Developer: Set Log Level...\" command and set it to \"Trace\" there's no \"Telemetry\" channel there, only Extension Telemetry: !\nDo you have any ideas I assume there would be nothing special needing to be done for codespaces, I register this output channel the same as the extension one and the same as the other output channels.\nWhere do you register Telemetry logger in codespaces, is it in server or on client side?\nIt's registered client side.\nMay I know where it is registered in sources?\nCan you please check if this code is being executed in the codespace?\nIt seems to be missing from web too. I can confirm when debugging the web bits locally it is called and registered, but nothing shows up in the output channel list.\nCan you please provide me steps to enable this locally from sources?\nSent instructions offline\nStill not seeing the telemetry output channel in Codespaces (on desktop if it matters)\nActually this is a bug in web, but I can confirm that this is working for me in both web and desktop. Wondering if you have any additional steps to repro."}
{"_id":"q-en-vscode-09a72e874684dad8cc5dd31490588928394c8a10a2c906033055e8faea3d1361","text":"Type a multiline commit message Scroll away Scroll back When the row gets rerendered (I assume) it is sized to one line. If I type a newline in this box, the size gets fixed !\nPossibly a regression introduced with"}
{"_id":"q-en-vscode-42bc6141ee42ee35ba8524ed53d2a8131a908425e5300413f67fc0e66089a761","text":"Just saw , this goes against the design. It's meant to only add the status if it doesn't already exist, if the status changed then it should use a different id and/or remove the old status first.\nI'm not clear on the solution. There're a limited number of Terminal Status IDs: In this case status ID hasn't changed, it's still , only the description which comes with it has updated.\nworking on it, I'll send a PR to you in a bit\nValidate still doesn't happen."}
{"_id":"q-en-vscode-f5e9079860545d2f297cbb711c57d914420489068b403e02fafa232be518886d","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.78.1 (system setup) Commit: Date: 2023-05-04T09:48:08.683Z Electron: 22.5.1 Chromium: 108.0.5359.215 16.17.1 V8: 10.8.168.25-electron.0 OS: WindowsNT x64 10.0. Sandboxed: No Steps to Reproduce: a workspace at a UNC path that isn't set in security.allowedUNCHosts. \"Yes\" when prompted to temporarily allow working from that host. changes to one or more files without_ saving them. VSCode without closing the workspace (may require specific settings to automatically re-open the previous workspace). time, click the \"Learn More\" button when prompted to allow working from that host. Note that any unsaved editors remain, but the rest of the workspace is not loaded as though the user clicked the \"No\" button. I'd argue that \"Learn More\" doesn't constitute a decision from the user and shouldn't proceed as though it was. Depending on the nature of the changes, this may cause a problem for safely saving those changes because if the user tries to re-open the workspace, they are asked to either save or close these orphaned editors. Perhaps a link among the text in the dialogue box would be more appropriate than a button which closes the box.\nI agree that the dialog should not close and/or make a decision when you click \"Learn More\". Unfortunately the dialog closes when you press that button, so best I can do is to open it again asking the same question."}
{"_id":"q-en-vscode-7f56ca1579a3bba4a861f652757bf3656d9f7f98f4ccaa08349a3cdab7723d19","text":"Run the command . Nothing happens Can we show the dev tools or open an output view that gets focus? From describing a user how to help us debugging an issue with storage:\nThe action also runs in browsers where we cannot open devtools automatically. We cannot use the output view because we actually show a real table: <img width=\"1394\" alt=\"image\" src=\"\"But we can inform the user via dialog.\nFor the other action \"Log working copies\" I will reveal the window log where it is logged to."}
{"_id":"q-en-vscode-b357c1cf982282ba4c9b9b2862b6dabd72092f9e0dd2c394e3ba892be0768d79","text":"Type: <bFeature Request</bDepending on the extension we are using, when code suggestion information appears, it is not read by screen readers. Steps to reproduce. Install the codeium extension. Open any project, and start typing or commenting the code; You will hear a sound when a hint appears, but this hint is not read by screen readers like nvda. I suggest that it is possible to activate a configuration in vscode, so that this type of information can be transmitted to screen readers. VS Code version: Code - Insiders 1.79.0-insider (, 2023-05-20T01:18:53.811Z) OS version: Windows_NT x64 10.0. Modes: Remote OS version: Linux x64 5.15.90.1-microsoft-standard-WSL2 <!-- generated by issue reporter --\nI can reproduce. We must not be getting the line content here\nI would like to inform you that I also tested it with the copilot and when listening to the sound that a suggestion was given, the screen readers also do not read.\nYou need to set to and it will work for copilot\ndo you think we should now set that to by default?\nNice defined it and it actually worked for the copilot. My suggestion is that it would not be possible for these settings to go to the accessibility feature? I know that it makes sense to stay where they are because it has to do with the editor, but they are difficult to find, or maybe we have some dock where all the accessibility features are and their respective paths where they should be activated, you get the idea.\nYes, we actually have help menus now for the copilot chat view / interactive editor session. When you focus those inputs, using reveals/focuses the help menu for each. I think for this case, we should enable inline suggestion by default\nSorry, just found out that by enabling the setting, it also worked for codeium. Should I then close this issue?\nI think we should enable this by default, so will use this issue to track that\nI agree with the idea setting the inline suggestion announcement to TRUE by default as it is now stable.\nVerified the setting default is flipped"}
{"_id":"q-en-vscode-2645e96ffbffa25f484dbbd387b3977b7e4b5bf654e19a27cf22efebba5bea1e","text":"Type: <bBug</bOpen terminal Use Terminal Copy Selecton as HTML command () after selecting text in terminal Expected: Selected text in the terminal is copied as html, so formatting (such as color) is kept. I am then able to paste this text elsewhere, outside of VS Code. Actual: [object event] notifcation] & text isn't copied to clipboard at all, clipboard contents remain unchanged VS Code version: Code 1.78.2 (, 2023-05-10T14:39:26.248Z) OS version: WindowsNT x64 10.0. Modes: Sandboxed: Yes <details<summarySystem Info</summary </details<details<summaryExtensions (16)</summaryExtensionVersion ------ npm-intellisense1.4.4 vscode-eslint2.4.0 vscode-npm-script0.3.29 prettier-vscode9.12.0 npm1.3.0 vscode-ansi1.1.6 python2023.8.0 vscode-pylance2023.5.30 jupyter2023.4. jupyter-keymap1.1.0 jupyter-renderers1.0.15 vscode-jupyter-cell-tags0.1.8 vscode-jupyter-slideshow0.1.5 vscode-react-native1.11.0 vscode-react-native-preview2023.5.821 es7-react-js-snippets1.9.3 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nThis was working at some pt, not sure why it doesn't anymore\nCan reproduce: ! It's also a little weird having in the precondition as that makes it never show up in the command palette (I'm guessing that's set to false when the terminal is blurred?)\nSomething's wrong with getting the serialize addon: !"}
{"_id":"q-en-vscode-5a778359675011b9bbacc0d6bcfddb9fb2749ad85f6fe3afbec1768b2af1a7a0","text":"Testing I might be doing something wrong, my contribution: I see no status entry that I can click on. I also tried with an existing command such as . ! Extension code is a hello world: Full :\nnot implemented"}
{"_id":"q-en-vscode-275dc3fc73a4554b5b6346060c5956ddf5165b70d2279a8a5017eef7bf1b3236","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: No. Related to extension api <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.79.0-insider OS Version: Macintosh; Intel Mac OS X 1340 Getting a session for a Microsoft account for the default tenant, and also getting a session for the account but for a specific tenant results in two items for the same account appearing in the Accounts menu. <img width=\"546\" alt=\"image\" src=\"\"Steps to Reproduce: First call with the following scopes: . And another call to with these scopes: I want to point out that the account objects on the returned sessions are actually different: <br<details<summaryMicrosoft Authentication provider logs</summary<p</p</detailscc:\nThese accounts aren't duplicate, they are sessions that exist for each signed-in tenant. I think the best solution would be to change the account label to . This would match what the Azure portal does. <img width=\"201\" alt=\"image\" src=\"\"I think this is an important change to make because this really doesn't look good and is confusing to users. Especially when users are trying to sign out. The screenshot below is after I sign in with a single Microsoft account that has 4 tenants. Having 4 tenants is not uncommon. <img width=\"515\" alt=\"image\" src=\"\"\nYeah this goes back to that \"technically they're different accounts because they're in different tenants\" thing. I'm inclined to agree with you and dedupe per-label.\nAlthough that would present better, I think deduping on the label wouldn't be great unless you made the label more standard across tenants. In that case, I think changing the label to be just email would be best. Then all of these sessions in my example screenshot could be grouped as one. If you deduped per-label with the current labelling it would show two accounts, one that represents 3 tenants, and another representing another single tenant. That would still be pretty confusing. IMO there's two options: Dedupe by email and hide all tenant related stuff. Signing out of that account removes all sessions associated with that email, regardless of tenant. Don't dedupe at all, instead include the tenant display name in the label as .\nsays this works!"}
{"_id":"q-en-vscode-80b2a0669bd1806a7decaccff7b567d78f32ce753864545435eac49e64b8d909","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --"}
{"_id":"q-en-vscode-bb14521cc7280bf65fceafefc8af4c6ed19b818f0a71d8ed25da5b623979b81d","text":"Type: <bBug</bIn previous VS Code version 1.78 when I pressed if inline suggest was visible with suggest widget it hides only inline suggest, if I press again same shortcut it then hides suggest widget. Now with 1.79 it hides both inline suggest and suggest widget. KeybindingService logs: My custom shortcuts to hide inline suggest & suggest widget: I even tried to add to clause and still it hides both. VS Code version: Code 1.79.0 (, 2023-06-07T14:31:03.317Z) OS version: Darwin arm64 22.5.0 Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (12)</summaryExtensionVersion ------ vscode-eslint2.4.0 gitlens13.6.0 prettier-vscode9.13.0 copilot1.88.132 vscode-graphql0.8.13 vscode-graphql-syntax1.2.0 symbols0.0.11 vscode-react-native1.11.0 vscode-thunder-client2.7.6 vscode-fileutils3.10.1 vim1.25.2 quokka-vscode1.0.543 (1 theme extensions excluded) </details<!-- generated by issue reporter --\nfull text search doesn't reveal the context key anymore. Has that been removed/renamed?\nIt is\nsorry I was writing that instead of copy-pasting. Already updated the issue.\nDoes it work when you swap the keybinding configurations? (Like literally exchange the positions in the file)\nStill the same issue even if I swap places :(\nSo, its either the or context key. Can you investigate with the F1 Inspect Context Key Tools to see if those context keys are set. Note that it will print context keys on the console of DevTools\n<details<summaryContext keys</summary</details\nas you can see is and is , but even though inline suggest was visible it returned false.\nSame in the monaco, if the suggest widget is visible, the inlineSuggestWidgetVisible will always be false no matter there is any ghost text or not\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nNow I see us hiding the suggest widget first and the inline suggest second, is that expected? This is true regardless of what order I declare the keybindings in (I do have the Vim extension installed)\nIt should hide first inline suggestions and second time you press it, it should hide suggest widget. Try to swap places of keybindings to see if it fixes it (order maybe matters?). Because previously this is how it worked (in 1.78, but regression was introduced in 1.79). EDIT: Oh I see you mentioned that order doesn't matter. Try to add in clause for command (I hope it works).\nis it fixed for you on the latest insider builds The following does not work for me either\nIt works! !\nGreat, thanks for confirming!"}
{"_id":"q-en-vscode-7787d36d058aed318bd0d8155ac5956aa8980409962c550377bc13f97baf3fb7","text":"Steps to Reproduce: Copilot Chat the file in VS Code +i It takes like 2s for the UI to show up. One time when I was doing a lot with VS Code it took so long (like 10s) that I thought I didn't have the extension installed. Not sure if this is on the UI side or the extension side, but even still, I think the UI should be instant (even if it's perceived fastness) since all I'm seeing is an input box that I will type in.\nI have done a bunch of tricks, most importantly to showing the UI while asking the extension to create a session. Also the UI is more lazily created which helps. However, once there are changes the bottleneck is setting up the embedded diff editor. There isn't much I can do about it <img width=\"726\" alt=\"Screenshot 2023-06-15 at 16 10 35\" src=\"\""}
{"_id":"q-en-vscode-3a0747fc036365f55261de533b65b35ee433c52a76c702b38a5f73491e51707c","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes, actually this issue has nothing to do with extensions <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- Version: 1.79.1 (user setup) Commit: Date: 2023-06-12T16:14:05.102Z Electron: 22.5.7 Chromium: 108.0.5359.215 16.17.1 V8: 10.8.168.25-electron.0 OS: Windows_NT x64 10.0. Steps to Reproduce: Connect to WSL(Ubuntu ) Profiles =Show Contents... <img width=\"286\" alt=\"image\" src=\"\"As you can see, the extensions of the profile are duplicated. The reason I found is I installed the extensions in WSL again. Maybe It's not a bug. But, the profile should distinguish between extensions installed in the host and WSL. Maybe like this: ESLint - host Disabled ESLint - WSL So that I know which place I install the extension and its enabled/disabled status. When I open the detail page of the duplicated extensions, both of them display an \"Extension is enabled on 'WSL: Ubuntu-22.04'\" message. What I expect is: The extension is installed and enabled on the host, the message should be \"Extension is enabled\". The extension is installed and enabled on the WSL, the message should be \"Extension is enabled on 'WSL: Ubuntu-22.04'\". Does it make sense?\nRelated on Stack Overflow: . Let me know if there's any debugging steps or details you'd like be to solicit from people in my answer there.\n/duplicate\nThanks for creating this issue! We figured it's covering the same as another one we already have. Thus, we closed this one as a duplicate. You can search for . See also our . Happy Coding!\nwhat's it a duplicate of?\nI am sorry, I remembered this bug and thought there is an issue already, but there is not. Reopened it.\nI guess duplicate only shows for extensions installed in both local VS Code and remote hosts. !\nFixed: Contents show only one instance of the extension"}
{"_id":"q-en-vscode-e3719f72d237b4a4422394b52588f731696eecef6944c94a5ee81dd8f580e7ea","text":"Add the possibility to persist the content hover sizes\nTo test, please read the following. Currently the implementation of the persistance mechanism of the content hover works as follows: The content hover has a default maximum size which is defined as follows: const defaultMaxHeight = (().height / 4, 250); const defaultMaxWidth = (().width 0.66, 500); Whenever the content hover is resized, the last hover dimensions are stored. Next time the content hover is shown, the maximum dimensions are updated as follows: const currentMaxHeight = (().height / 4, 250, lastHeight); const currentMaxWidth = (().width 0.66, 500, lastWidth); The content hover is rendered naturally within these constraints. If the lastHeight and lastWidth are smaller than the default max dimensions, then the current max dimensions are unchanged. Therefore the next rendering of the content hover will be the same as the initial rendering. Test that the content hover works according to these constraints. There is some discussion going on about how to better implement a persistance mechanism. We could also persist the size for example on a per-token basis in the editor. Let me know what you think is best!\nDo you have an example of how it should work? Regardless of how I resize it, it is not remembered.\nYou may see an example of how it works here. The content hover is resized to a bigger size in the gif. The current max width is increased to a higher value and at the next rendering of the content hover, the widget takes on a bigger size. To see this behavior you should choose a content hover where the content does not fit in the visible widget, where there is a horizontal scrollbar. I opened an issue for persisting the content hover on a per-token basis here: as an alternative idea."}
{"_id":"q-en-vscode-728954f4d98ab1937e4b27d9df95d96321b1739eec71ddf49f5f816f1a61b82b","text":"Testing I started the Troubleshoot issue flow and it took a long time to create my initial profile. All I had was a status bar item which took me a while to discover (I actually thought something else went wrong). I would love to see and editor / web view better guiding me through that process by: listing the steps and adding check marks when done showing me the number of bisects still necessary giving me a final report.\n<img width=\"258\" alt=\"capture\" src=\"\"\nWe should also have a UI to always stop the process (see ) for a situation where this is useful\nI now see that there is a notification for each step. This makes it better, so may be a simple thing to do is to show a notificatio with progress when the profile is created instead of a simple status bar item.\nActually, setting up the trouble shooting profile should be fast and instant.\nWas not for me :-(\nFollowing notification is shown now !"}
{"_id":"q-en-vscode-38ffe00e5885c437eac3c1605f469a7d9b0e40c558c34bda699424d08884f987","text":"Testing After clicking the \"Install\" entry in the quick pick, there's progress for a few moments, then the quick pick flashes back to showing all the previous entries again, before finally landing on the setup options for the remote extension. !"}
{"_id":"q-en-vscode-4f62249deb06b79a9b62cf00830bf0e35a89313e2c0867c6e3168293ab78d63c","text":"Testing Format of keybindings is particularly ugly ! !\nDoes this question accept PR from outsiders?\nyes. VS Code does accept PRs!\nThanks!, and I'm going to refer to styles for other widgets, or if we already have styles we like?\nIt's more complicated than just styles, sadly. We need to use our rich hovers instead of the native hovers... thankfully we did some work recently to plumb a all the way down to the QuickPick. To fix the title bar buttons, you can probably add the hover delegate in into the options of these s: The list will be harder has the buttons get created here: but we can be iterative here and focus on the title bar buttons first. the other issue that has to be addressed is that Monaco: currently doesn't have a HoverDelegate implementation here: so either we need to implement a HoverDelegate in the \"standalone\" code (aka Monaco) or we need to pass in the right value to tell quick pick to not use a hover delegate and use native hovers instead. So yeah... it's a bit more complicated than styles unfortunately\nSorry to reply late, thank you for your help! I will start learning the relevant code here over the weekend, hope it goes well!\nIt is very convenient to create rich hover based on the clues you provide!(Please ignore his style and content) I'm using method for the time being to set the style: I would like to ask for help, is it the right choice to use this method? If yes, I will continue to improve ps:For Monaco, I haven't started looking at the relevant code, but I will as soon as possible !\nOn the VS Code (non-Monaco side) we already pass a hover delegate in here: So you should be able to just reuse that instead of\nI adapted the title bar, but the buttons in the menu options cannot be easily adapted because they are not available e.g. 'Learn More' in screenshot, next I'm going to learn monaco-related code, for monaco's tabbar, do you know which pages can be easily called show?\nwe can start with just fixing the title bar buttons if you'd like and then fix the Learn More ones later. We can be iterative :) We still for sure need this before that PR of yours comes in: from\nThanks for your help, I'll go investigate the monaco related code, but it may be later, I'm also stuck with this issue\nIf you use Node 18, you shouldn't see that issue.\nIt works for me, I would also like to ask for a simple question: In addition to calling api, is there any scenario in vscode that directly runs monaco? I'm trying to find a link to a monaco input call, but failing\nGood question. Here are the steps for running Monaco: What I would recommend is maybe temporarily adding a button to an existing quick pick like the command palette: add a line like: and don't forget to import it:\nThank you for your patience! I will start learning the code\nI investigated the existing code and found some problem if I wanted to implement a HoverDelegate in the \"standalone\" code: We need to use related code, but they are in the , and should not be import this layer, do we need to re-implement a standalone with hover? or lower the code hierarchy to the level, I'm not sure if the conclusion I got is correct, so I still need to ask you for help, thanks!\nso before, I provided a default implementation of the here that does nothing: I think what we should do is remove that and allow to be /optional. If it's don't pass it in to the s so that the native hover is used as fallback.\nIt's better to handle it this way, otherwise it's really difficult to achieve the current situation (especially for me), I'll submit the PR later\nI submitted an initial PR, but I'm not sure if I missed some details, and found that one of the tests timed out\nVerification on macOS would be easiest I think: Open up a few pickers Hover over different components: description detail quick pick buttons quick pick item buttons None of them should be using native hovers. They should all be using custom hovers.\nIs there anything I can do to help? Do we need to replace everything that needs to be replaced with a custom hover,I'm willing to submit a PR to modify these if I can\nIt's already done! Now we verify everything is working as expected before :)\nAre these the old tooltips or the new ones? <img width=\"275\" alt=\"image\" src=\"\"<img width=\"306\" alt=\"image\" src=\"\"This is what I'm seeing.\nYes those look like custom\nActually, the keybinding one isn't... let me capture this in another issue as the class doesn't actually support a hover delegate."}
{"_id":"q-en-vscode-f5188758fd86b319997cfccdc2e73e4bc44b316249fc9527b5a772731215afc6","text":"Testing Steps: Open locally cloned vscode repo with any branch Welcome page Connect to... Continue Working in Nothing happens after this for me... but I do see the commands get run:"}
{"_id":"q-en-vscode-86a5e254ec04ff63f12018c36480856b84f87cd9ab0822f01c8b1df6641a0dd0","text":"Testing a file +enter to run cell in interactive window window GitHub Copilot classic installed and signed in something into the input box and ctrl+enter, this opens a third editor group with multiple synthesized Copilot solutions (feature from GitHub Copilot) close that editor :bug: now there is an empty third editor group which really shouldn't happen AFAIK, also there is screen cheese, it shouldn't really be possible for Copilot to cause this kind of layout issue <img width=\"1256\" alt=\"image\" src=\"\"<img width=\"1245\" alt=\"image\" src=\"\"\nI assigned you for grid layout and I assigned you for editor layout, please feel free to unassign if this isn't relevant to you\nempty editor groups are not a bug per-se, if you configure , groups will remain open when the last editor closes. Also, empty groups can explicitly be created, e.g. via . However, the fact that you cannot get rid of the empty group when you press the X button sounds like an issue to me. When you try to close, is there anything printed to the console?\nOhhhhhhhhhhh good bug! There is definitely splitview/gridview breakage here.\nWait. I could repro once... but not anymore. Can you still repro?\nYes I reproed twice\nI cannot see anything in logs when clicking on the X unfortunately (fwiw, I do not have that setting configured) When I repro this, here's what's in the log (this happens right after closing the last editor in a group that results in the group not also being closed)\nI wonder if this could possibly be a regression from the auto-sizing feature you did?\nWe just need to be able to repro on demand so we can attempt to repro on a version before my changes to confirm this. But I can't seem to repro no more. Can you ?\nI cannot reproduce this, I never get an empty editor group. This is with fresh user data dir and extensions dir.\nif you have a reliable repro that you can share, including your settings if any. Here is what I do: !\nI did some more messing around and was able to repro a similar problem ! <details</details\nOK, I have a trivial and consistent repro. TLDR it looks like the interactive window does not play nice with split view. an interactive window (you can even do this with the Jupyter: Create Interactive Window command) the Settings UI the Settings UI onto the right of the interactive window :bug: the split view is very unhappy Logs <details</details\nThis also repros for notebooks opened in the second group Right click a .ipynb file -open to the side open another editor and drag it to the right side of the first\nOh wow. So it looks like the call to throws an exception when done in the interactive window and that means the Grid will indeed be unhappy!\nfor surrounding some of the calls with .\nArguably, the grid/splitview should also protect itself...\nYeah, maybe. There could be other cases.\nI can't repro any of these now but it would probably help for to verify too\nI still do not have a functional cursor/trackpad so cannot drag/drop maybe you can verify?\nVerified it's working great"}
{"_id":"q-en-vscode-da41ba48ed14558a25e684926033067367175b751e475ebb8861c9d997cee53e","text":"The steps are simple: settings editor in one group untitled editor in the group next to it sizes are equal drag the untitled editor to the left of the settings editor the sizes end up not to be equal. Insiders ! Stable ! Maybe this is related to the settings editor wanting a larger size? I can reproduce this in a small window size (the default one when you open a fresh user data dir)."}
{"_id":"q-en-vscode-eac08b53244e0ccaab0c3fcc12feae9c8daa489519af38491923977e4da7fd35","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.79.2 (user setup) OS Version: Windows 10 22H2 To fix the issue please update the native-is-elevated package. It was fixed two months ago.\nCould you maybe verify using our insiders? You can give our preview releases a try from:\nIt seems working properly now. Thanks!"}
{"_id":"q-en-vscode-1412d412cd4a313dc84d27d05cfdbff654551319233a1d5a9415d896f023111b","text":"in VS Code :bug: sticky scroll has no syntax highlighting <img width=\"691\" alt=\"image\" src=\"\"\nHey I update my insiders version and tested on the file you sent, I am currently not able to reproduce this error: ! I'll try to reproduce this issue. I think I may have seen this happen for a very brief moment.\nIt appears this is due to the following PR which attempted to fix Since the state was no longer recomputed when the new state was equal to the old state, this means that the rendering was no longer done when the model tokens changed but the state remained the same. A fix will be made to revert this change."}
{"_id":"q-en-vscode-410da19bea9edf7e8fb3319105d95875cd56ae01f78bb7a233af56866e8440aa","text":"create a folder with a task that uses: update an extension and use \"Reload window\" or use Reload window directly from the command pallette observe that the task now runs twice: <img width=\"328\" alt=\"image\" src=\"\"\nLikely a regression from one of the task fixes this iteration.\nNot able to reproduce Do you have task.reconnection enabled? Would be helpful to see each of the terminals\nThis issue has been closed automatically because it needs more information and has not had recent activity. See also our guidelines. Happy Coding!\nThis problem still exists, I'm afraid I'm not sure the repro steps are as Alex described, but I still get duplicated tasks\nIt seems for Alex's repro steps to work, the terminal panel needs to be hidden\nWhen the panel is hidden, the tasks don't run at all for me atm. I'm going to merge something that will fix this part of the issue, but in order to truly test it, I'll have to first"}
{"_id":"q-en-vscode-a33c6a421580bb944b5a732d9b3e726099e14b789462f88aed6d66c17f11b00e","text":"Verification steps: Make sure you are on diff editor version 2. Enable whitespace only diffing (the reversed P in the toolbar). Diff two identical documents, change indentation. Verify that a notification is shown.\nI cannot find your reversed P, nonetheless, with diff version 2 and removed \"ignore leading/trailing whitespace\", I get this: <img width=\"1384\" alt=\"image\" src=\"\"and TBH, why should have an option to ignore any diff view? Isn't the purpose of the diff, to really see the diff? Git extensions -more logical diff: <img width=\"551\" alt=\"image\" src=\"\""}
{"_id":"q-en-vscode-b51d4861dfb9e8e9142a9010223b5c890d686575ad590ac0340dfe07e2e97e07","text":"enough cells in the interactive window that it needs to scroll \"interactiveWindow.alwaysScrollOnNewCell\": false to the bottom of the IW and execute another cell :bug: scrollbar height changes and is at the bottom, but the new cell isn't visible"}
{"_id":"q-en-vscode-3264427872058f981b3aba5683e498e4f676c17083bdd63172525e3e8546ff13","text":"Testing Repro: screen reader mode the editor the show notification command, :bug: the editor is still focused the show notification command, the notification center is now focused?\nMy notification was read once fully, then the screen reader started reading it a second time. Part way through that second time, the notification no longer seemed selected (didn't have a blue box around it), and the screen reader just stopped reading it.\nthe looks correct here, but I can repro the focus getting stuck in the editor any thoughts would be appreciated"}
{"_id":"q-en-vscode-5e3590a2603f950e5cb488560456c5ed8f9ede729adf7ef86141758b1fecef81","text":"a md file: between the html elements bug This inserts markdown syntax even though markdown typically isn't supported inside of html like this:\nLikely requires a proper markdown parser to fix\nI don't know if this is the expected result, since I don't see any difference pasting in the tags. I tried using two versions of .\nThanks for testing. I think there's still a bug with . In this case it should likely not paste as a markdown link. Somewhat weird case because it's the pasting itself that makes the html span multiple lines. With the newlines it's two entirely separate html blocks according to markdown"}
{"_id":"q-en-vscode-8c5c6d1e8fee1c98bcbc0cbc9bb7de9f86eeae7070041dfb62326405babb4048","text":"I was just looking at the field of our event and I noticed that a lot of the durations are negative. Looking at the code here It looks like we are computing the miliseconds since 1970. I guess that we are hitting some overflow, and thus the numbers are negative. I suggest that we instead use the class that Joh is using here for example I can go ahead and push a fix for this.\nI am not 100% sure if using would fix this. at the end of the day uses , and you were using . It looks like getTime treats timezones differently. So my proposed fix might fix the problem or it might not. What do you think Here's the Kusto query to see the events with the negative durations\nI have pushed a potential fix for this. To verify run the above query in kusto Make sure that no events are happening in the latest VS Code Stable / Insiders version (thus that the fix indeed fixes it). I am assigning this to September to give it some time for the data to settle in.\nThanks for the fix. Using is the right thing here."}
{"_id":"q-en-vscode-1f9597c016a777ae7ca006363d28931ef8b2c26beed843804ec6f3a5a4aa10e3","text":"Type: <bBug</bOn Mac, in terminal, press Option+F1. It says: Go to Next Command (Command+DownArrow) Go to Previous Command (Command+UpArrow) This has been recently updated as: Option+Up/DownArrows. VS Code version: Code - Insiders 1.81.0-insider (Universal) (, 2023-07-28T12:09:45.183Z) OS version: Darwin arm64 22.5.0 Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (90)</summaryExtensionVersion ------ android-dev-ext1.3.2 Bookmarks13.4.1 openscad1.2.1 spellright3.0.116 zoterolatex0.4.1 mermaid-markdown-syntax-highlighting1.5.2 doxdocgen1.4.0 dscodegpt2.1.15 vscode-markdownlint0.51.0 vscode-eslint2.4.2 vscode-quick-select0.2.9 vscode-deno3.19.1 gitlens14.1.1 EditorConfig0.16.4 prettier-vscode9.19.0 vscode-google-translate1.4.13 codespaces1.14.13 copilot1.99.294 copilot-chat0.5. remotehub0.60.0 vscode-github-actions0.25.8 vscode-pull-request-github0.68.1 easy-snippet0.6.3 path-autocomplete1.24.1 latex-workshop9.13.3 lilypond-syntax0.1.1 scheme0.2.0 better-cpp-syntax1.17.2 google-search0.0.1 vscode-lua-format1.3.8 lilypond-formatter0.2.3 lilypond-pdf-preview0.2.8 lilypond-snippets0.1.1 vslilypond1.7.3 zotero0.1.10 git-graph1.30.0 vscode-docker1.26.0 black-formatter2023.4.1 flake82023.6.0 isort2023.11. python2023.12.0 vscode-pylance2023.7.40 jupyter2023.6. jupyter-keymap1.1.2 jupyter-renderers1.0.17 vscode-jupyter-cell-tags0.1.8 vscode-jupyter-slideshow0.1.5 remote-containers0.305.0 remote-ssh0.102.0 remote-ssh-edit0.86.0 remote-wsl0.80.2 vscode-remote-extensionpack0.24.0 azure-repos0.36.0 cmake-tools1.14.34 cpptools1.16.3 cpptools-extension-pack1.3.0 js-debug-nightly2023.7.2417 remote-repositories0.38.1 vscode-github-issue-notebooks0.0.129 vscode-selfhost-test-provider0.3.15 vscode-serial-monitor0.10.0 vsliveshare1.0.5877 resourcemonitor1.0.7 autodocstring0.6.1 pandocciter0.10.3 shiny-python0.1.2 shinyuieditor0.4.3 quarto1.90.0 r-debugger0.5.4 java1.21.0 vscode-xml0.26.1 r2.8.1 multi-command1.6.0 vscode-deepl1.0.6 abc-music0.4.0 lua3.6.26 latex-utilities0.4.10 chatgpt1.1.2 cmake0.0.17 errorlens3.12.0 intellicode-api-usage-examples0.2.7 vscodeintellicode1.2.30 vscode-arduino0.6.0 vscode-java-debug0.52.0 vscode-java-dependency0.23.0 vscode-java-pack0.25.12 vscode-java-test0.39.0 vscode-maven0.41.0 markdown-all-in-one3.5.1 grammarly0.22.1 (1 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nYes reported this, but I cannot find that issue, so we'll track it here. This is because of this line"}
{"_id":"q-en-vscode-0ec5e90d0da4f19dbf80923af86da866d5b91847144ef2aaf3afd2cbd4e7d69c","text":"Type: <bBug</bIn editor, I can't use and alt+[` do not have specific \"when\" clause and conflict against inline suggestion view in keybindings. VS Code version: Code - Insiders 1.81.0-insider (, 2023-07-28T12:08:04.472Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (89)</summaryExtensionVersion ------ android-dev-ext1.3.2 Bookmarks13.4.1 openscad1.2.1 spellright3.0.116 zoterolatex0.4.1 mermaid-markdown-syntax-highlighting1.5.2 doxdocgen1.4.0 vscode-markdownlint0.51.0 vscode-eslint2.4.2 vscode-quick-select0.2.9 vscode-deno3.19.1 gitlens14.1.1 EditorConfig0.16.4 prettier-vscode9.19.0 vscode-google-translate1.4.13 codespaces1.14.13 copilot1.99.297 copilot-chat0.5. remotehub0.60.0 vscode-github-actions0.25.8 vscode-pull-request-github0.68.1 easy-snippet0.6.3 path-autocomplete1.24.1 latex-workshop9.13.3 lilypond-syntax0.1.1 scheme0.2.0 better-cpp-syntax1.17.2 google-search0.0.1 vscode-lua-format1.3.8 lilypond-formatter0.2.3 lilypond-pdf-preview0.2.8 lilypond-snippets0.1.1 vslilypond1.7.3 zotero0.1.10 git-graph1.30.0 vscode-docker1.26.0 black-formatter2023.4.1 flake82023.6.0 isort2023.11. python2023.12.0 vscode-pylance2023.7.40 jupyter2023.6. jupyter-keymap1.1.2 jupyter-renderers1.0.17 vscode-jupyter-cell-tags0.1.8 vscode-jupyter-slideshow0.1.5 remote-containers0.305.0 remote-ssh0.102.0 remote-ssh-edit0.86.0 remote-wsl0.80.2 vscode-remote-extensionpack0.24.0 azure-repos0.36.0 cmake-tools1.14.34 cpptools1.16.3 cpptools-extension-pack1.3.0 js-debug-nightly2023.7.2417 powershell2023.6.0 remote-repositories0.38.1 vscode-github-issue-notebooks0.0.129 vscode-selfhost-test-provider0.3.15 vscode-serial-monitor0.10.0 vsliveshare1.0.5877 resourcemonitor1.0.7 autodocstring0.6.1 pandocciter0.10.3 shiny-python0.1.2 shinyuieditor0.4.3 quarto1.90.0 r-debugger0.5.4 java1.21.0 vscode-xml0.26.1 r2.8.1 multi-command1.6.0 vscode-deepl1.0.6 abc-music0.4.0 lua3.6.26 latex-utilities0.4.10 cmake0.0.17 errorlens3.12.0 intellicode-api-usage-examples0.2.7 vscodeintellicode1.2.30 vscode-arduino0.6.0 vscode-java-debug0.52.0 vscode-java-dependency0.23.0 vscode-java-pack0.25.12 vscode-java-test0.39.0 vscode-maven0.41.0 markdown-all-in-one3.5.1 grammarly0.22.1 (1 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --"}
{"_id":"q-en-vscode-4a320aff411b7422c9aa446360dfb02e89617720f5b2824e39f39b2ce7cc3a1d","text":"When exporting a profile, a SR user has to navigate the whole list to reach the button. We should either move the button to the top or have a keybinding/command which does this. cc <img width=\"293\" alt=\"Screenshot 2023-07-31 at 11 30 34 AM\" src=\"\"\nIs not the tab should take to Export button when the complete tree is focused instead of focusing the check box?\nI'm not following\nTry to get into the state where Tree is focused not any of the elements of the Tree. For eg: <img width=\"662\" alt=\"image\" src=\"\"See the outline focusing the tree. Press <tabnow. This should focus the Export button right?\ncorrect - pressing tab when my screen looks like the above causes focus to be lost rather than going to the button.\nTab at this state is taking focus to the checkbox in the tree. IMO checkbox should be focused when the row is focused and you tab then. Is it possible to change this behaviour?"}
{"_id":"q-en-vscode-f8ca427bd04e8840f3540c7359e7eaecdf34e00106bcbdf9f0cf5cf5ef32ac1c","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --I have multiple profiles in VS Code and I disable everything I don't need in each of them, including built-in extensions that I will never use with a particular profile. After the successful installation of the latest version I wanted to try the new option; while it works for user-installed extensions, it fails for built-in extensions, read more below. Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: OS Version: , 22H2 (.2070) Steps to Reproduce: the extensions view built-in extensions only the cog icon on any (enabled) extension on error will occur (see attached screenshot) <div align=\"center\"<img src=\"\" alt=\"Error in VS Code\"</div<brNote: all extensions are present at the provided folder shown in the error message.\nSame for me but on MacOS Ventura 13.5 (22G74)\nCan you see if the extension is maybe uninstalled when you follow your steps? That could hint to\n/duplicate\nThanks for creating this issue! We figured it's covering the same as another one we already have. Thus, we closed this one as a duplicate. You can search for . See also our . Happy Coding!\nUnfortunately the latest patch release: and the mentioned didn't fix it, here's the error screenshot again, in the latest version. Also, I've opened the folder itself where the extension is located, while VS Code complains it doesn't exist. <div align=\"center\"<img src=\"\" alt=\"Error in VS Code\"</div\nThis action should not be shown on builtin extensions because they are available for all profiles by default."}
{"_id":"q-en-vscode-675414227f107a2c150aefd121272c46f698c38baec0a8b6391df12d7c498744","text":"VS Code Version: 1.82.0-insider OS Version: Ubuntu 22.04 The path is missing the first character in the path for a specific file on the 'git show --textconv' lines 'ub-folder' vs 'sub-folder' If the command is run manually, with the correct path there are no issues Working (Non-root folder) 2023-08-03 19:10:33.264 [info] git show --textconv [4ms] 2023-08-03 19:10:33.267 [info] git ls-tree -l HEAD -- [4ms] Not Working (Root folder) 2023-08-03 19:12:26.351 [info] git show --textconv [20ms] 2023-08-03 19:12:26.386 [info] git ls-tree -l HEAD -- [35ms] Expected (Root folder) 2023-08-03 19:12:26.351 [info] git show --textconv [20ms] 2023-08-03 19:12:26.386 [info] git ls-tree -l HEAD -- [35ms] Originally posted by in\ncould you please share repro steps that I could use to reproduce this? Or the complete contents of the git output window would also help. I have a suspicion where the issue is but I need the logs to reproduce and fix the issue. Thanks!\nHello Ladislau Szomoru, open vsc go to Source Control (Ctrl+Shift+G) look under \"Staged Changes\" and click on notice left side is always empty and right side shows entire file changed. Thanks, Gregory Porras"}
{"_id":"q-en-vscode-c85ca10952f0a7826830bd9b55e7c758f7e121927baf124ba9b2c522d85e0e48","text":"hover over editor tabs observe some kind of shadow dropping over the text\nThis only reproduces with set to and tabs enabled in general. Hovering over any tab shows the white shadow at the end of the label. With this setting, every tab label has a element to draw a linear gradient (): <img width=\"400\" alt=\"image\" src=\"\"A tab itself is a container of a label and the toolbar with the close button. The close button is set to to not take more space than needed. When you hover, the toolbar is made visible, actually pushing the element into view having the described effect: <img width=\"504\" alt=\"image\" src=\"\"I actually think this is an OK behaviour for when the tab is small enough that the label cannot full show, and that was the intention. However, it is a bit strange that this would happen even for a tab that shows fully and has enough space. If someone wants to chime in to take a look with a reasonably small change, opening for help wanted. But may decide to close out of scope as this does not seem like it was ever a regression.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!"}
{"_id":"q-en-vscode-2622bd5333f2982b2147f7e6abe3aef23eba3778f16c149c23fb9658f28a854c","text":"Insiders 1.82.0 Looks like the command center and associated toolbar buttons aren't functioning when clicked. Weirdly it seems like I can sometimes(?) get a hover state but it doesn't always work. Repeated clicks minimizes the window which makes me think the events are going straight to the titlebar and being handled as such.\nI can reproduce\nMaybe an Electron 25 update regression with draggable regions and custom title bar? //cc\nlooks like behavior on macOS is now aligned with windows and linux, the region no longer emits events. Behavior changed with this commit and verified with the following fiddle Looks like we would need to revert this change to align with Windows and Linux.\nYeah, looks like drag and click isn't working for macOS anymore. It always seems like an oversight/bug but a wanted one. To enable clicking again I have the for the CC even for mac. On the downside it means that the window cannot be repositioned from the CC anymore\nSlack is the same, you cannot drag from their command center either."}
{"_id":"q-en-vscode-d76d100c5224a990325d3aab14c8cb3479b49b208574f7c2d16c60ad761e1e9c","text":"Settings: !\nHow it should be: <img width=\"442\" alt=\"image\" src=\"\"\nHm, strange. I haven't made any changes here lately. I'm noting that is not having an effect on the foreground color used on the tree within Explorer. It does however affect other lists like the one used in the quick pick. I also noticed the terminal tabs tree flicker to use that red before being stomped on again. Not sure what's going on here yet.\nLooks like this is from\nGotcha I might just revert the change and switch it to something more specific - I didn't consider all the other applications of the list widget and checking other themes (it was okay in a couple of the themes I checked)"}
{"_id":"q-en-vscode-d6a3ad66f199c7a003b3991fd53ad846afe8ada7ebf848ac3a4b5e26c973c6e0","text":"Type: <bPerformance Issue</bIn a bash terminal with Shell Integration, execute the following command, which is 5k characters long: On both my machines (Linux native and Windows + RemoteSSH), this commande takes 1s to execute. Make it 10k characters and it takes 4s. At 15k, it takes 8s. I'm using bash version 4.4.23 (the one shipped on SLES15 SP4). It appears to be related to the Bash Shell Integration, specifically the function located . This can even be reproduced outside of VSCode, in plain bash. Just define this function in a shell and execute it with a long argument. As a workaround, it is possible to disable Shell Integration in the settings and open new terminals. With a newer bash 5.2 built by myself, performances are better, but still non-negligible: 5k: 0.5s, 10k: 1.5s, 15k: 3.5s. Note: I get similar numbers with a bash 4.4 compiled by myself, so better performances with newer bash might just be related to different compilation options/optimizations. VS Code version: Code 1.81.1 (, 2023-08-09T22:18:39.991Z) OS version: Linux x64 5.14.21-.24.28-default Modes: <details<summarySystem Info</summary </details<details<summaryProcess Info</summary</details<details<summaryWorkspace Info</summary</details<details<summaryExtensions (1)</summaryExtensionVersion ------ remote-ssh0.90.1 </details<!-- generated by issue reporter --\nThanks for the report, hopefully a bash expert will spot this and help optimize the function\nUnfortunately this still seems to be quite slow for me. Normal bash runs quicky (). Bash in VS Code has had a core stuck at 100% CPU for a few minutes now with no signs of letting up... <img width=\"790\" alt=\"image\" src=\"\"Doing some binary searching, it's \"fast\" until the echo hits 1989 characters (a total command length of 2001 characters), at which point it stalls as above\nActually, it seems any command over 2000 characters in bash stalls and takes a long time to run. Doesn't have to be an echo. It seems like time taken is somewhat exponential, 2001 characters take 21s to run for me while 2100 characters jumps that up to 29s. That seems to be the point at which we use a 'fast' implementation. In reality this seems much slower on macOS at least\nHi. Could you check that whether you have command and on your system? Since I actually provided two fast impl in the PR, I would like to know which is problemetic for subsequent troubleshooting. Edit: never mind. I have reproduced the hang and propose another fix."}
{"_id":"q-en-vscode-d7b15518f07871e888e116ad184cfe0194725fd7bac54383341290f79c768d1f","text":"Workaround This is an issue in Chromium itself. For now there are the following two workarounds: ### <blockQuote</blockQuote1. ### <blockQuote</blockQuote## Bug <!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. -- <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --1. ### VS Code Version Per: affected version is: <blockQuoteType | Data ----------------|----------------------------------------- Version | 1.82.0-insider Commit | Date | 2023-08-11T17:10:45.392Z Electron | 25.4.0 ElectronBuildId | Chromium | 114.0.5735.248 | 18.15.0 V8 | 11.4.183.27-electron.0 OS | Linux x64 6.4.9-1-default </blockQuoteHowever, per: that , that is: <blockQuoteType | Data ----------------|----------------------------------------- Version | 1.81.1 Commit | Date | 2023-08-09T22:18:39.991Z Electron | 22.3.18 ElectronBuildId | Chromium | 108.0.5359.215 | 16.17.1 V8 | 10.8.168.25-electron.0 OS | Linux x64 6.4.9-1-default snap ### OS Version <blockQuote</blockQuote</details## Steps to Reproduce 1. The GUI is invisible: the menubar continues to operate correctly.\n82 updated to Electron 25 and doesn't work on older OS.\nI'm using cpe:/o:opensuse:tumbleweed:, which updated 2600 packages yesterday (which I have installed) yet this problem remains. Nobody in their right mind would consider that to be old. Additionally, I am using which means that no independent part of the OS should be considered old. This is the most bleeding edge system for miles. Regardless, should I perhaps open a linked issue at\nOh, sorry, I thought it's an old OS issue. Before reporting in the electron repo, what is Help Toggle Developer Tools bringing up? Have you tried to do a clean reinstall of VSCode Insider? Perhaps it was just a bad build. Or, alternatively, you could try to download the latest Insider installer and install over your existing broken version (which I assume it's 3-4 days older than what you'd get right now) - that should effectively update it.\nNo problem. Thanks. Apologies for being curt. ! ! Appears that it's reproducible.\nVery strange, indeed. I'm out of ideas. Would be interesting to be able to test with a previous Insider version, but for that you would need to have the old package and disable automatic updates. Electron was updated from 22 to 25 on August 8, so the last version with Electron 22 is the Insider from August 7, which is also the last version that I've tested and still working on Windows 7. Anything after that doesn't start anymore on Win 7. I'm not sure whether the Electron update on August 8 is causing your issue, though...\ndoes the log file explain anything? To me, it seemed a little too verbose to be of much use. I can try other methods to get relevant diagnostic data if you are able to instruct me.\nI see many GPU-related errors, such as: Errors: link failed but did not provide an info log [(48)] (checkLinkStatus): Program link or binary loading failed with no info log. [(81)] Skia shader compilation error Found these: So it's definitely something related to the new Chromium included in Electron 25 and these issues seem relatively recent. The first link recommends to disable hardware acceleration, you could try searching how to do that from command line in the new Insider. Unfortunately, I don't have the time to dig deeper right now.\nper , do you know whether VSCode contains a directory? I haven't been able to locate it.\nremediates the problem. Because indeed this issue is dependent upon , before this update is rolled out, there needs to be remediating measures (or the update needs to be temporarily halted) lest it break a great many peoples' installations.\nCan you please take a look at this? It's Chromium/Electron 25 related, so I immediately thought of you. Please see the response just above for a possible solution. Thanks!\nThanks for investigating the issue, given there are workarounds in place this is not necessarily a release blocker. I will wait for chromium to address this issue and backport the fix to Electron 25.\ncould a patch at least be to the affected releases to delete the relevant directory? Surely users shall be affected otherwise? I doubt we can expect them to diagnose a known issue if the solution is so simple.\nWe don't want to delete the GPUCache for all users and the affected scenario described in is not possible to detect from the application layer. I think the best I can do is to add this issue to the release notes and document the workarounds for visibility.\nWhy?\nIt will have impact on startup performance, for now I would wait to see the impact of this issue across our users and decide on a fix forward. The upstream issue has also been open for a while now, signifying its low impact.\nJust happened again. !\nThis is mentioned in the 1.82 release notes I guess the issue remains in 1.83-insiders, haven't tried it yet.\nIt appears weird to me that solely VSCode appears to be affected by this (on my system) because Discord has been very happily updating frequently, and obviously has its own Electron-provided directory. !\nRelated on Stack Overflow: Related on Reddit: ,\nGot this installed via fedora yesterday This wasted quiet some time on my end, as I first went searching for changes to fedora itself and finding this issue on vscode wasn't trivial.\nLeaving this here for future me, as I now have that problem on a different PC\nThis also happens for other Electron based programs like Element and also chromium and firefox. see\nI'd disagree, . It doesn't read like pragmatism - more like \"didn't want to do a good fix\". Startup time after the workaround felt normal. Super aggravating that this wasn't resolved with a one-time installation task to clear the GPU cache. Hard to know the aggregate cost incurred by people encountering this issue, but \"not small\" is probably an underestimation. Linked-issue count wouldn't be a good proxy for that metric, in that product-support's practice as shown in seems to be linking to the known-issue buried in the release notes, not to this issue. Sorry for the flames, but there you have it. When one's coding environment gets fsck'd, it's no fun, and erodes trust. Please consider.\nI, for one, find great enjoyment in my coding environments.\nThis issue has reemerged with the latest version 1.82.3 for Ubuntu 20.04.1: Removing after the upgrade still works to get vscode running\nIt occurs for me almost every update of the package. It's ridiculous.\nSame.\nAs mentioned before, this is caused by the new update in VSCode I have personally reverted to until this issue is properly solved\nI was under the impression from the chromium bug that this is one time change and not required for every startup. I see that you are using snap version of VSCode, can you check if you are required to delete the cache for every start using the instead ?\nI must rectify my comment: with the snap of version v1.82.1, it was required to clear the cache every startup, with v1.82.2 I don't remember having to do it, while with v1.82.3 it was needed only once on upgrade. I will edit my previous comment to fix it.\nWe had a runtime version bump between to , which seems like a trigger in the snap case. Can you setup deb package, we will have release this week. It would be good to confirm if updating requires to clear the cache in the deb package as well.\nFor what it's worth, I haven't encountered it on the rpm again\nI've tried the following: removing v1.82.3 (snap) and installing v1.82.3 (deb) did require clearing the cache. removing v1.82.3 (deb) and installing v1.82.2 (deb) did not require removing the GPUCache. upgrading v1.82.2 (deb) to v1.82.3 (deb) using apt did not require removing the GPUCache. hope it helps\nI'm not using the snap any more. I use the package provided directly by Microsoft from\nEncountered this problem on Fedora 39, probably after upgrading my vscode rpm which is currently on version 1.82.3. Problem disappeared after running .\nNot with VSCode specifically: but when updating to electron v28 alpha, had this issue. Also solved with (or just deleting the entire Electron dir within ).\nAlso ran into this, thanks for this workaround, it worked.\nI also ran into the issue in VSCodium (which is a special build script for this source code) after an upgrade to 1.83.1 on Debian Trixie and XFCE 4.18. The worked like a charm (which is the equivalent folder on Debian for VSCodium). I did find that I only ran into the issue on my RX 550 (Lexa) and not my Radeon RX 580 (Polaris); I found that the bug disappeared when running the program with appended to it to run it on the latter graphics card.\nThis has happened a second time in 1.83.1.\nGot it again on fedora using the rpm on a system, where I already dropped the cache once Version 1.83.1\nJust installed an update with no vscode update but had a mesa update. VSCode broke again. I assume this is going to happen every time a mesa update is installed (I'm assuming because of what others are stating, could be something other than mesa but seems to be the biggest suspect).\nGetting it after every mesa update. But may be \"GPUCache\" need some harder refresh or detection of brockage?\nThis issue has been fixed upstream via . It might take a while before it trickles downstream though, since it needs to be picked up by Electron first before it can go into VSCode. I did attempt cherry-picking the commit and generating a local build and it seemed to fix the issue.\nThe commit is fairly trivial and can probably be safely backported to all supported Electron trees. This is a very widespread issue, affecting virtually every Electron app running on Linux. It would be great to get the fix as widely disseminated as possible considering how slow the app ecosystem is to update versions. Is there a linked Electron issue for this? Do we need to create one if not or does the Electron project accept drive-by backport PRs?\nI haven't found any linked Electron issues yet. Based on , since the patch is temporary, has been committed upstream, and will eventually be removed, I think the Electron folks will accept the patch. However, do take note that the commit has dependencies on previous commits. On my working local build, I have these 3 commits cherry-picked in chronological order: Therefore, if anyone is going to bring this up to the Electron team, all 3 patches would need to be merged. Add the patch files to the and append the filenames of said patch files to the . Depending on the Electron build version, there might also be some merge conflicts. Based on my experience, those conflicts should be resolvable, but I did not test all of the older Electron major version release branches. For each branch, we would need to double-check if the 2 dependency commits are already included in their respective Chromium version or not.\nThank you for the backporting work to Electron versions 25-28 over the past few days. Downstream applications should now be able to pick those fixes up. Once VSCode picks up the fix, we should be able to close this issue.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nI'll verify when the next release reaches me on openSUSE.\nby chance, were you able to verify the fix for the issue yet?\nthe time period hasn't been quite long enough to be certain. This usually occurs on an update or after approximately two months. However, VSCode did update recently, and this didn't occur (not that it was guaranteed). Consequently, if upstream has fixed the issue, and we're using the patched version, because it hasn't appeared yet, I'll keep this completed until it occurs (and we can reasonably consider this complete).\nConsidering the issue as verified on 1.85.x. 1.86.0 will go out in around a week. Feel free to comment on this issue thread if the issue comes up again after that."}
{"_id":"q-en-vscode-06ae6fc82d2ce6609f5f67300cd120610af61afa72a4c7bda50c3d4628d004f2","text":"Type: <bBug</bPort forwarding over ssh is not working using vscode-insiders. It is working using vscode. How to replicate: some http server on the remote the port to be forwarded and select http as a protocol opening in browser / doing a curl from your local machine the request is stuck pending VS Code version: Code - Insiders 1.82.0-insider (, 2023-08-21T05:32:43.957Z) OS version: WindowsNT x64 10.0. Modes: Remote OS version: Linux x64 6.1.38-59.109.amzn2023.x8664 <details<summarySystem Info</summary </details<details<summaryExtensions (54)</summaryExtensionVersion ------ Handlebars0.4.1 vsc-python-indent1.18.0 dotenv1.0.1 remote-containers0.305.0 remote-ssh0.105.1 remote-ssh-edit0.86.0 remote-wsl0.81.0 vscode-remote-extensionpack0.24.0 remote-explorer0.4.1 remote-server1.4.2 material-icon-theme4.29.0 indent-one-space1.0.0 vscode-todo-highlight1.0.5 nugetpackagemanagergui2.0.6 Handlebars0.4.1 vue-peek1.0.2 vscode-markdownlint0.51.0 vscode-eslint2.4.2 gitlens2023.8.1905 EditorConfig0.16.4 prettier-vscode10.1.0 vscode-jest-runner0.4.66 copilot1.104.347 copilot-chat0.7. vscode-github-actions0.26.1 todo-tree0.0.226 terraform2.27. vscode-test-explorer2.21.1 vscode-nuget-package-manager1.1.6 roslynator4.4.0 vsc-python-indent1.18.0 vscode-python-test-adapter0.7.1 dotenv1.0.1 vscode-scss0.10.0 vscode-docker1.26.0 csharp2.0.376 vscode-dotnet-runtime1.7.0 black-formatter2023.5. flake82023.7. isort2023.11. python2023.15. vscode-pylance2023.8.31 test-adapter-converter0.1.8 indent-rainbow8.3.1 material-icon-theme4.29.0 prisma5.1.0 vscode-scss-formatter3.0.0 indent-one-space1.0.0 intellicode-api-usage-examples0.2.7 vscodeintellicode1.2.30 vscode-conventional-commits1.25.0 volar1.8.8 vscode-typescript-vue-plugin1.8.8 vscode-todo-highlight1.0.5 (2 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nI can reproduce this with SSH, but not with Dev Containers. have there been any changes to SSH recently that might cause this?\nHappened to me yesterday too. Spend a lot of time rebooting my machine and killing processes until I tried regular vscode to find that port forwarding was working fine there. What is weird though is that the browser just hangs. Nothing happens if you try localhost:4000. There is no error it just hangs but if I try a port that isn't being used like localhost:5000 then the browser says that there is a connection error\nI can reproduce\nI have tried the latest version And I still cannot connect to my react dev server using port forwarding into my ssh remote\nI still have this issue in the latest version: To reproduce: node script . a request from client side: side receives only the first chunk: After this, the request gets stuck and never ends. In my case, the issue mostly affects small requests. A 10kb request always has the issue, but a 100kb request seems to be consistently fine. Size of the response also seems to be important. I connected to a distant server (with a ping of ~300 ms), but there doesn't seem to be any problem with the connection itself.\nFixed on the next Insiders. Good find, it was a latent issue that never got hit before. Root causing it was fun\nDoes this need to be verified with Remote-SSH specifically?\nYes, it requires remote SSH. You can verify this issue and at the same time by following Dador's instructions\nserver: <img width=\"319\" alt=\"image\" src=\"\"client: <img width=\"688\" alt=\"image\" src=\"\"Looks verified in latest insiders\nSorry, it does not work for me. To reproduce: Repository at After starting the devcontainer on a remote ssh-host, launch task \"Serve\". Then access http://localhost:4000.\nare you using the latest VS Code insiders?\nJust updated:\nI'm not able to reproduce that, and I doubt it's related to this issue since Containers does not use the code path in question.\nSo, different bug, same behaviour? It worked in insiders before, it still works in stable... it does not work anymore in the current insiders. Does that mean the next release will also no longer work? Edit: Just to make sure there's no misunderstanding: My devcontainer is running on a remote ssh server.\nMost likely. Sharing any contents in the Console (from the \"Developer: Toggle Developer Tools\" command) and the \"Shared\" log from the Output panel would be useful in diagnosing the issue.\nPreparing a new issue!"}
{"_id":"q-en-vscode-754cc54d99cc2e82dd57732356d979d8f275ffd007fa36e9e76e24563af567dd","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --This would replace . Proposal: ts /\nCopyright (c) Microsoft Corporation. All rights reserved. Licensed under the MIT License. See in the project root for license information. *\n-------------/ declare module 'vscode' { // export enum RelatedInformationType { SymbolInformation = 1, CommandInformation = 2, SearchInformation = 3, SettingInformation = 4 } interface RelatedInformationBaseResult { type: RelatedInformationType; weight: number; } // TODO: Symbols and Search export interface CommandInformationResult extends RelatedInformationBaseResult { type: RelatedInformationType.CommandInformation; command: string; } export interface SettingInformationResult extends RelatedInformationBaseResult { type: RelatedInformationType.SettingInformation; setting: string; } export type RelatedInformationResult = CommandInformationResult | SettingInformationResult; export interface RelatedInformationProvider { provideRelatedInformation(query: string, token: CancellationToken): ProviderResult<RelatedInformationResult[]; } export interface EmbeddingVectorProvider { provideEmbeddingVector(strings: string[], token: CancellationToken): ProviderResult<number[][]; } export namespace ai { export function getRelatedInformation(query: string, types: RelatedInformationType[], token: CancellationToken): Thenable<RelatedInformationResult[]; export function registerRelatedInformationProvider(type: RelatedInformationType, provider: RelatedInformationProvider): Disposable; export function registerEmbeddingVectorProvider(model: string, provider: EmbeddingVectorProvider): Disposable; } }"}
{"_id":"q-en-vscode-543d523950ff922cd1f0eb1d9ab4f7b684d5aecdb2975a8270ca13d03d011fed","text":"Repro: a perf profile a ~100mb TypeScript file (duplicate the contents of a large TS file to make this) perf profile ! It doesn't end up working properly anyway: !\nAfter delete and undo: !"}
{"_id":"q-en-vscode-f46c7a0aa053035c19e67a4692da9b3918f0c511f4b47f34e79057b8096e3dcb","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.81.1 OS Version: server 2022 Steps to Reproduce: is a system install. Users login using redirected appdata folders launch code with appdata redirected to a unc path or can launch 'code --user-data-dir servershare$code' error Unable to write program user data , code trying to write to c:share$ Users do not have access to write to c: grant full control to root of C: and launch code a new folder is created c:share$ file created c:share$codeUserglobalStorage is created and code launches does not use the unc path while all other files are written to servershare$code fine with version 1.79.2\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.81.1. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nbased on the above steps looks like the sqlite database is not using the unc path, passing the issue to you.\nCan you try to reproduce with our nightly insider builds? You can give our preview releases a try from:\nSame issue. installed 1.82.0-insider build. Still creates the in the wrong location.\nexcellent finding! I think this is a regression from partitioning state per profile. Specifically: When using , UNC paths only work for scheme but now I think we have as scheme. I am not sure when we made that flip.\nI think this is spreading to a lot more places, the code is very old and was always assuming scheme to be used, which will not work anymore if a UNC path is used: I think we will have to make a pass over usage and find those that potentially use a non-file scheme...\nI think I start to understand the issue better, it seems quite specific to global storage only, the other in the seem to be mainly of scheme. I believe this started with moving out of the environment service and into the profile service:\nThe change to scheme started happening from this change - - I started using this scheme for all user data to adopt atomic write. Since I had reverted atomic write for user data in this commit - - I can revert to use file scheme for user data on main. But going forward, I will bring back user data scheme and probably we should our file service and uri identitiy utilities instead of using fspath and path join.\nbut there is a point where our file service is no longer in charge: when we hand of a to SQLite3 module, we really have to give it a proper path that is a UNC path if the user picked a UNC location. So just using will not help here, we need to find a way that is not breaking UNC paths in all places where we need it.\nThats right, I also observed that we browsing the storage code. May be we should convert this back to file scheme before using in those places?\nyes, doing seems right to me, even though ugly. Curious: with your custom file system provider on , are you converting back to a URI and then handing the off to the disk file system provider? Or do you convert one level above on the level of the file service? Asking, because I wonder if there is already a bug in your provider where we potentially loose UNC paths.\nThinking a bit more about this, I wonder if should statically allow us to register a scheme to behave like a scheme. Because, we do take an existing well defined of scheme and turn that into . This in turn means, every client that deals with such a would need to ensure that the scheme is converted back to before using the . That seems fragile. If we could tell to treat like a scheme, maybe the change would be easier and less fragile?\nIMO this is as ugly as using on URI, because it can cause issues anytime. Even if you know the scheme of the URI converting to file scheme is always safe. Here is the conversion logic: This is handled in the FSP of vscode-userdata. I just change the scheme and pass it to the native FSP."}
{"_id":"q-en-vscode-c825587e90b7d7c54319184cfefdb0db68cd162cfff9ac4421ffdb20125997d0","text":"It should be just \"Copy Output\", the \"to Clipboard\" is obvious? This then aligns with other copy actions across the product Testing"}
{"_id":"q-en-vscode-8e2347e6b4e880b13466128570a1637f9375286dadd85e033d935773cb2a047b","text":"Testing open execute a cell that shows issues select copy output no output is copied I'd except that this \"just works\", at least for its text/markdown output representation"}
{"_id":"q-en-vscode-cacb02222ea2c836493115c2595fb18e6b7429c4a0a8a2b485d5b2929c248a7b","text":"Going through the error logs showed the error below . No steps but it is rooted at this line\nverification - Perform a few actions in a notebook, add cells, run cells, close/reopen the tab, run more cells Open the dev tools and ensure the error is not in the console"}
{"_id":"q-en-vscode-e765ae9f5dd77ecb03bce7bda78475c151b7ba16fa08fd974931a2fe5440973e","text":"Testing Should be exposed using a code action (specifically a source action) instead of as a setting? Some advantages: We can reuse the existing setting instead of adding a new one Sorting will play nicer with other source actions since users can configure the order in which the actions run Users can trigger source actions through the context menu or with keybindings\nHey that's a good idea! I'll discuss it with the team. What do you think ?\nYes, makes sense!"}
{"_id":"q-en-vscode-9dc8a982d530b5654ce8132aeaa3be42b9ddc1db953d2f91b882048cb5951b0f","text":"Type: <bBug</bI am developing a VS code debugger extension by inheriting from DebugSession class. I implement disassembleRequest method to support disassembly view. But now there is a strange bug which I am not sure whether it is the bug of VS code or not. When disassembly view is firstly open, it shows wrong addresses and instructions. But after I scroll mouse up and down to manually refresh disassembly view, it shows correct addresses and instructions. So it seems it is a refresh issue. I have inspected the returned data to VS code and I am sure data is correct. BTW, if the returned data deos not contain source information, then the address and data is correct. I also post a question in stackoverflow: The question contains two picutures which should be able to describe my issue. Thanks! VS Code version: Code 1.82.0 (, 2023-09-06T22:07:07.438Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (9)</summaryExtensionVersion ------ biosdbgv44.1.2 vscode-eslint2.4.2 workspace-storage-cleanup1.1.5 isort2023.10.1 python2023.16.0 vscode-pylance2023.9.10 cpptools1.17.5 hexeditor1.9.12 memory-viewer0.0.1 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nCould you try this in the latest VS Code Insiders build? A bunch of disassembly view fixes just went in\nI test on the latest VS code insiders 1.83.0, and the issue still exists. Please see below pictures When disassembly view is firstly open: ! After scroll mouse up and down to refresh disassembly view: !\nCan you please either share your extension/a reproduction, or share a log of the DAP messages sent and received by your debugger?\nHi Connor, I have sent disassembly message to your email Thanks! Bill Zhang\nHi Connor, below is the implementation of disassembleRequest: protected async disassembleRequest(response: DebugProtocol.DisassembleResponse, args: DebugProtocol.DisassembleArguments) { try { const baseAddress = args.memoryReference; const offset = 0; const instructionOffset = args.instructionOffset 0; const instructionCount = args.instructionCount; const input = { baseAddress: baseAddress, offset: offset, instructionOffset: instructionOffset, instructionCount: instructionCount }; const instructionList: BIOSDBGInstruction[] = []; const output = await callRpcMethod(\"Disassemble\", input); for (const item of output) { const addressHex = '0x' + item.address; const instruction = new BIOSDBGInstruction(addressHex, item.instruction, item.instructionBytes); if (item.sourcePath !== \"\" && item.lineNumber 0) { instruction.location = this.createSource(item.sourcePath); = item.lineNumber; } (instruction); } = { instructions: instructionList }; } finally { this.sendResponse(response); } }\nYep, got it, thanks. My plate is full right now but I'll look into this within the next couple weeks.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nTested with 1.84.0-insiders, it seems that this issue is fixed. Thanks!"}
{"_id":"q-en-vscode-030a6ab80cf93ee5b2bb97e7539a80db4cdae79edf4d2036f8c30e8b129e8218","text":"Repro: Open multiple terminals Select multiple terminal tabs Right click and change color Change color only a single terminal is changed Real context stored with the action was done recently as part of Related:\n/duplicate\nTracking each action separately\nThanks for creating this issue! We figured it's covering the same as another one we already have. Thus, we closed this one as a duplicate. You can search for . See also our . Happy Coding!\nI'm not sure this is worth fixing as color is meant to differentiate between tabs. Same for the icon/ renaming.\nyou might want multiple red though, right now it doesn't work as expected. We were actually doing this wrong by querying which tabs were selected also, after we get the actual instances this was triggered on passed into the action, so we can probably remove the special tab instance command as part of this."}
{"_id":"q-en-vscode-53b96a7dd1fe897bfdb4ef810197dbaca883bcf9726908c98a5d91edb62dfef1","text":"Testing ! Option + Down/Up arrow in terminal is not navigating between commands. These are working only in accessibility view. In terminal the keybindings are cmd + Down/Up arrow.\nI'm experiencing the same issue testing on Windows 11.\nThis text should say in the accessible view - the keybinding is meant to work there, not in the terminal.\nsame issue! All the other commands work fine, arrows/screen reader work fine in the quick pick after command is run. I am wondering though if the keybinding for prev/next commands might be too generic?\nEveryone, the prev/next commands must be run in the terminal's accessible view - not the terminal. Can you pls test that? I plan to make that clear by tweaking the dialog's text.\nYes, it worked in the terminal accessible view for me (as it sounds like it did for too).\nWorked fine for me too! Sorry for the mixup\nThanks definitely see why you were confused though, so will fix that\nto verify, open the accessibility help dialog by focusing a terminal and using . verify that the specify that they're to be applied in the terminal accessible view."}
{"_id":"q-en-vscode-2bfc8e03e596255fcc7ff5ba90c3abddbf23c092156c1bc76ea588ca016aaa66","text":"Testing Start from , on a repo: The git extension will associate as the most likely candidate for 's base. But since does have an upstream, , which, at the point was created, was pointing to the same commit as itself... wouldn't be a better base candidate instead?\nI think that this makes sense to me however would you favour the remote branch (origin/foo) just because it points to the same commit as when was created or should it be a general rule that if a remote is present, we would favour it?\nWill mark this and look out for additional feedback.\nThough more about this and since we are adding and actions to the incoming node, it does make sense for it to be an upstream branch. Marking the issue as \"bug\" so that it gets verified during endgame."}
{"_id":"q-en-vscode-73b27c76ffea4b0dfbdae8a0614ae24e6af0a118ad5323d0144419ec6729ef6f","text":"Testing Testing on Windows 11. opens terminal accessibility help dialog correctly: ! Explore options, decide I want to try Run Recent Command () to open terminal accessible view, opens correctly: ! Try I don't observe anything happen - but maybe I'm misunderstanding what should happen or how I should set this command up?\nSorry this was not clear. Perhaps I should specify which commands only apply when focus is in the terminal? Perhaps that should also work when the focus is in the accessible view... I will think about that.\nTo verify, focus a terminal in screen reader mode. open the terminal accessibility help dialog with and check which keybinding to use for . open the terminal accessible view by focusing the terminal and using use that key binding. make sure it runs whichever command you select.\nI think I'm still missing something in my understanding. Steps I took: Open VS Code repo in latest VS Code Insiders Toggle screen reader mode so that I'm Screen Reader Optimized Open and focus terminal , the terminal accessibility help dialog opens I note that will run recent command, and will get me to previous commands Open terminal accessible view by focusing terminal and using to get to a previous command (this worked well) Once I've gotten to a previous command I'd like to run, I try but again don't observe anything happened: !\nApologies, Brigit, it appears there's an old bug resurfacing in which the keybinding is not listed correctly in this dialog. On windows, run recent command is invoked with when in screen reader mode. On mac, with screen reader mode, it's .\nReopening to fix that\nAh yay, worked well for me, thank you! I'll hold off on labeling as if you'd also like to use this issue to track fixing the keybinding bug.\nActually, I think I'll reopen that other issue. Since the issue was the command not working, it would be great if you'd verify this and I'll link the other . Thanks\nhere's the issue where we're tracking that"}
{"_id":"q-en-vscode-b2b8be862e69fbdac021ed008b04b6febdf2092b9c752203d7d456aa6f47c6ca","text":"When there are multiple debug sessions, the dropdown action item shows in the debug toolbar here as a button that doesn't do anything. You can see it in the gif in We could just hide that item when this setting is set\nI believe it focuses the Call Stack view but only when paused and when having the view active. Tho, I do agree that we shouldn't have it there to begin with."}
{"_id":"q-en-vscode-0e7c1535eb07a9293d68df512fe1017236d924581e1dead8085bb6f6db7a959f","text":"Type: <bBug</bclone repro from and open project folder run the repro extension from vscode open two folders in the workspace (these will not be touched by the repro code, the problem was encounterd in multi-folder workspace and this just reproduces that environment) run command after some time popup with error will appear Expected beahvior: the loop running the dummy task will keep running indefinitely This seems to be caused by using in combination with changes of the workspace folders. When the loop in the repro is changed to only keep writing the file and running the tasks the problem never appears VS Code version: Code - Insiders 1.82.0-insider (, 2023-09-01T11:05:38.381Z) OS version: Linux x64 4.19.0-18-amd64 Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (0)</summaryExtensionVersion ------ </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nI'm not seeing the task you said to run in that folder !\nThe repro steps say to run command , not a task. The command runs the task itself via the API. Here is the command in the command pallette ! And here is the error it produces ! Please try the repro steps again and let me know if you still can't see what I show in the screenshots\nI get this error, which is interesting because the workspace contains two folders. !\nFrom the screenshot it seems like you have one workspace folder containing two file system folders - that still means the workspace is a single root workspace You need to have a multi-root workspace by calling I hope the terminology I use is correct, it is taken from this doc\nPR fixing it"}
{"_id":"q-en-vscode-067966accdde69a221a242943413f234d83bfe3e59d5b557f5735e20c51b8bde","text":"I have a setting in my settings. Everytime I reload, I see a setting being . This keeps on happening, the old setting is never removed. ! Not sure if this should be a candidate. does this code look right to you? Or does migration also need to remove the old setting?\nIf you want to remove the old setting you should explicitly mention that in the migration like below\nCan the migration function be made in a way that people do not forget this?"}
{"_id":"q-en-vscode-c793515b2104700e173ebd3f165ac1f88f28696d2461a0b424ac39625c508b5b","text":"I'm pretty sure some of these keybindings aren't ever enabled with their context key requirements, e.g. I don't think the inline chat can be focused in a markdown cell that isn't in edit mode. Nor will cursor navigation mode be true while edit mode is false."}
{"_id":"q-en-vscode-c84b261302b84e0f7d7d67668fb0dce24a354a4aa3a974d0de99311432eecf29","text":"Type: <bBug</bCC Copilot inline suggestion Accessible View (alt+F2) for the suggestion : this opens a side-by-side copilot suggestion panel instead of accepting the suggestion. I think we need to go back to the original keybindings: for accepting suggestion; for submitting a request. VS Code version: Code - Insiders 1.84.0-insider (, 2023-10-12T05:37:35.298Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (92)</summaryExtensionVersion ------ android-dev-ext1.3.2 aiprm-lang0.0.2 Bookmarks13.4.2 openscad1.2.2 spellright3.0.118 zoterolatex0.4.1 mermaid-markdown-syntax-highlighting1.5.2 doxdocgen1.4.0 vscode-markdownlint0.52.0 vscode-eslint2.4.2 vscode-quick-select0.2.9 vscode-deno3.26.0 gitlens14.3.0 EditorConfig0.16.4 prettier-vscode10.1.0 vscode-google-translate1.4.13 codespaces1.15.4 copilot1.126.0 copilot-chat0.9. remotehub0.62.0 vscode-github-actions0.26.2 vscode-pull-request-github0.74.0 overleaf-workshop0.3.0 cslpreview0.2.2 easy-snippet0.6.3 path-autocomplete1.25.0 latex-workshop9.14.1 lilypond-syntax0.1.1 scheme0.2.0 better-cpp-syntax1.17.2 vscode-bard0.0.10 google-search0.0.1 vscode-lua-format1.3.8 lilypond-formatter0.2.3 lilypond-pdf-preview0.2.8 lilypond-snippets0.1.1 vslilypond1.7.3 zotero0.1.10 git-graph1.30.0 vscode-docker1.26.1 black-formatter2023.4.1 flake82023.8.0 isort2023.11. python2023.18.0 vscode-pylance2023.10.20 jupyter2023.9. jupyter-keymap1.1.2 jupyter-renderers1.0.17 vscode-jupyter-cell-tags0.1.8 vscode-jupyter-slideshow0.1.5 remote-containers0.318.0 remote-ssh0.106.5 remote-ssh-edit0.86.0 remote-wsl0.81.8 vscode-remote-extensionpack0.24.0 azure-repos0.38.0 cmake-tools1.15.31 cpptools1.17.5 cpptools-extension-pack1.3.0 js-debug-nightly2023.10.1117 powershell2023.9.5 remote-repositories0.40.0 vscode-github-issue-notebooks0.0.130 vscode-selfhost-test-provider0.3.19 vscode-serial-monitor0.10.0 vsliveshare1.0.5883 autodocstring0.6.1 pandocciter0.10.3 shiny-python0.1.5 shinyuieditor0.5.0 quarto1.103.0 r-debugger0.5.4 java1.23.0 vscode-xml0.26.1 r2.8.2 multi-command1.6.0 vscode-deepl1.0.6 abc-music0.4.0 lua3.7.0 latex-utilities0.4.10 cmake0.0.17 errorlens3.14.0 intellicode-api-usage-examples0.2.8 vscodeintellicode1.2.30 vscode-arduino0.6.0 vscode-java-debug0.54.0 vscode-java-dependency0.23.1 vscode-java-pack0.25.14 vscode-java-test0.40.0 vscode-maven0.42.0 markdown-all-in-one3.5.1 grammarly0.22.1 (1 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nGood catch. I couldn't test yesterday as inline completions weren't working for me. I think that we can just increase the weight of this keybinding rather than change it back.\nThat keybinding will still work when one is not in the accessible view. LMK if that sounds good to you.\n-- We then reassign the default keybindings for the following: : This is currently . : This is currently .\nWe cannot change the extension contributed ones, so I was suggesting overriding it when in the accessible view. I cannot get inline completions to work in oss atm, so cannot test any of these changes until that works and am leaning toward just reverting the change."}
{"_id":"q-en-vscode-1b69fb8b5614cf4459203c2af60ccc40eec5a9921c797b022244663518920aba","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.83.1 OS Version: 12.6.9 Steps to Reproduce: or the Python and Black Vscode extensions a new notebook and execute any cell or save it You'll a \"Formatting with autopep8\" message in the status bar even though that formatter extension isn't installed! However manually triggering \"Format Cell\" does use black."}
{"_id":"q-en-vscode-4fc668b21de0762b5103ecfa0ddf505044447a39d12f3afa2cf659d88d977f8c","text":"When using the new diff editor with enabled, I very rarely get to the right editor position whenever I click the top action, above the editor.\nYou mean go to reference is broken?\ndiff editors have an action to open the original file (I think contributed from SCM?). When clicking that, expectation is that the current view state of the diff editors right hand side applies to the opening editor so that you stay in exactly the same view state (for example: the selection and viewport). <img width=\"140\" alt=\"image\" src=\"\"\nI really like the collapsed diff editor but this is a very frustrating regression from it which makes me want to shut it off. The issue is 100% reproducible in my usual workflow and I encounter this issue many times per hour.\nI can reproduce. The \"Open File\" command is a mystery to me. Debugging shows me that is called twice, but both times with unhelpful stack traces: ! ! The second invocation sets . However, there is no reveal mechanism.\nFurther investigation showed that the \"Open File\" action is actually implemented in the git extension (even when comparing two files on disk) :/ Which I think is very questionable design, as it should be a core diff editor feature."}
{"_id":"q-en-vscode-eee78ebf28cc36cf950f1df662f400aeab8185f297b4433aede4e45abe2b0fa7","text":"From We should fix block styling in Markdown previews in main as well, from discussing with we want to make the styling consistent across chat, settings UI, and extension webviews\nThe three greys look different to me. I'm unsure whether those are the colours I should be comparing. !\nTo verify this you just need to confirm that there is a background for blocks, since it used to just be a brownish foreground color."}
{"_id":"q-en-vscode-edf042f76eac2a42137494a9c681e29b6a34b9c213e62aa062d718bef0efbfce","text":"Type: <bBug</bWhat does this mean / do? ! VS Code version: Code 1.84.0 (, 2023-11-01T11:29:04.398Z) OS version: Windows_NT x64 10.0. Modes: <!-- generated by issue reporter --\nApparently from by and present since 1.78.\nthis should only be showing up for the quick pick item hovers, not the actions which have a small amount of content."}
{"_id":"q-en-vscode-74f591f463fb6cb4fe16fc49486bb34ce2698d4420f81d1e0c1d4d8472c5b71f","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --<img src=\"\" alt=\"drawing\" width=\"600\"/Currently, we can control when to show the inline suggestion toolbar by setting to \"always\" or \"onHover\", which are defined as follows: \"always\": show the inline suggestion toolbar whenever an inline suggestion is shown, and \"onHover\": show the inline suggestion toolbar when hovering over an inline suggestion. For the new feature, I suggest to add an option \"never\" as follows \"never\": do not show the inline suggestion toolbar.\nI think this should be very easy to implement. Up for a PR?\nI would like to take this.\nAny ideas why this setting isn't in the Settings JSON file? .\nThank you so much for accepting the request and for the contribution!"}
{"_id":"q-en-vscode-d3b3b427b4aac4066ffdbd96672215454832f3231413098e24f31bed8a6bde90","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --extension is used to specify the syntax highlighting for popular languages, such as TypeScript: Currently, the workaround is to push the settings file to Git: In addition, extension is also . It would be nice if VSCode could support these aliases to YAML natively.\nIf you want to make a PR, that would be great!\nI made a PR, I hope it's good, because I have no experience with the codebase:"}
{"_id":"q-en-vscode-8424e64f7f04e5a3e167a53e1a531440b3b0cfd3347cf5a8e7351413cf3108e4","text":"open test explorer notice that the input has the wrong background color, e.g it has two background colors <img width=\"455\" alt=\"Screenshot 2023-11-15 at 14 29 05\" src=\"\"\nWhat theme are you using? I can see the slight difference in your screenshot, but inspecting things I'm not seeing any backgrounds being applied to that action / menu bar\nLike, any light theme?\nStill confused on the latest Insiders for me in e.g. solarized light: <img width=\"725\" alt=\"image\" src=\"\"Does it reproduce on a fresh user-data-dir for you? If so, are you able to do a quick devtools inspect and give a hint as to what styling is being applied that makes monaco input have a different background color?\nI see. It doesn't show with an empty profile but I also don't have any setting that screams test explorer background color\nCould you bisect your user settings or share them so I can play around?\nThis seems to be because I have the test explorer in the secondary sidebar. There is some settings that controls this but I have no clue which one it is <img width=\"2188\" alt=\"Screenshot 2023-12-12 at 18 02 53\" src=\"\"\nCSS is fun"}
{"_id":"q-en-vscode-2256a1ff645c18e4f4601a1c67ae4f3984976510185bf59ae09c98fd4356d011","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --The direct symptom is in the site , but the investigation went through (which the site uses) and ends up here. Steps to Reproduce: to This shows the \"LLVM Opt Pipeline Viewer\" (1) which is a diff pane created by monaco's (2). It's contents change based on the user selection in from the 'passes' list (3). ! see normal operation in action: click in order the passes \"SROAPass\" (first one) and \"Greedy Register Allocator\". Diffs are shown properly: ! reproduce the bug, click in order the passes \"SROAPass\", then \"Debug Variable Analysis\" (one above Greedy) and finally \"Greedy Register Allocator\". First, text diffs are not being decorated despite being present: ! and second, a console shows the error: Preliminary investigation directs attention to these lines in 's : In the (not so rare) problematic scenario, and are non-empty but is, so the code enters the loop and crashes immediately. I suspect something broke in PR , but not sure.\nA monaco-playground link with a small standalone repro for the problem: This creates a diffView and populates both sides with an offending text (specifically, one with non-empty \"unchangedRegions\"). After two seconds it sets the left pane to the same text as it currently has, but this time the decorations disappear and a console error message as above appears. Perhaps this tidbit would be helpful: It seems that upon entering , contains ids that were already invalidated - i.e., do not appear in : ! And so returns only nulls and the resulting array is empty. My current guess is these decoration-IDs invalidation is too early, and should be moved to after .\nThank you for the prompt fix!\nPlease, please, please, include inline diff view in your testing. I use inline view 99.9% of the time, side by side very rarely. Regardless, I hope this Thank you for your work.\nI saw this problem again today.\nCan you please share the output of the dev console? Thanks!"}
{"_id":"q-en-vscode-46def8806d89a618843f9f1dfaf8d93e021078e82e976f453e115949c7d91aaf","text":"Initially noticed by : <img width=\"780\" alt=\"Screenshot 2023-11-17 at 13 53 45\" src=\"\""}
{"_id":"q-en-vscode-073abd1b2d64730abaf4a717b0d97ed9d09fb124bf7bca74db826d5c59e995ec","text":"Type: <bBug</bWhen reviewing outgoing changes for my tracking fork of microsoft/vscode after merging upstream/main into my local copy today. ! Also, based on I was surprised to see the magnifying glass icon still being used for \"View Changes\". VS Code version: Code - Insiders 1.85.0-insider (, 2023-11-22T05:35:23.085Z) OS version: Windows_NT x64 10.0. Modes: <!-- generated by issue reporter --\n/assign\nI'm very sure this is a duplicate of\nThis sounds like there was no insiders for today. Can you check you are running the build from today? The PR just got merged yesterday.\nYes, this is today's. The new codicon appears in the Timeline view, just not in the incoming/outgoing party of Source Control view: !\nIt does for me: !\nAh, we're not talking about the same thing. My screenshot is of the feature recently by and controlled by the and settings. It comes from here:\nare you able to reproduce this consistently? Could you provide steps that that I should use to try to reproduce the issue? Thanks!\nin said it was probably a known issue. Ever since Insiders got that fix I haven't had any problems opening multi-diff for outgoing changes that get listed after I merge from upstream/main into my local main prior to pushing them to origin/main (my tracking repo). So I'm closing this one.\nthanks for the update!\nHmm, just happened to me again with today's Insiders: ! Reopening. I'll hold back from pushing the local changes that I merged in from upstream/main in case there are places you'd like me to look for clues while I can repro it.\nDevtools console:\nIt's this one: ! And specifically which is of course binary: ! I guess the multi diff editor needs to handle this case gracefully instead of failing to show the diffs for the other files.\nthanks! could you please take a look? Thanks!\nAhh, because it is a binary file..."}
{"_id":"q-en-vscode-2e9d383c5976acada63f922762a0b07bc9335ed1836a8fbdd78308908feede8a","text":"Type: <bBug</bA corrupted file is stored in CachedExtensionVSIXs and never re-downloaded, preventing extension from being installed until the file is manually deleted. Repro: Attempt to install an extension while the Azure DevOps service is not available and returning errors. Result in this comment: The corrupted vsix contains the following HTML instead of zipped data ! Shared Output: VS Code version: Code 1.84.2 (, 2023-11-09T10:52:57.054Z) OS version: Darwin arm64 23.1.0 Modes: Connection to 'SSH: ox' could not be established Canceled <details<summarySystem Info</summary Connection to 'SSH: ox' could not be established Canceled </details<details<summaryExtensions (31)</summaryExtensionVersion ------ vscode-ron0.11.0 vscode-eslint2.4.2 prettier-vscode10.1.0 kotlin0.2.32 copilot1.139.0 copilot-chat0.10.2 copilot-labs0.17.1121 vscode-github-actions0.26.2 better-cpp-syntax1.17.2 vscode-peacock4.2.2 vscode-clangd0.1.25 membrane1.6.1 vscode-dotnet-runtime2.0.0 vscode-postgresql0.3.0 remote-containers0.321.0 remote-ssh0.107.1 remote-ssh-edit0.86.0 cpptools-extension-pack1.3.0 hexeditor1.9.12 makefile-tools0.7.0 remote-explorer0.4.1 proto0.0.4 vscode-yaml1.14.0 vscode-gitweblinks2.9.3 rust-analyzer0.3.1748 even-better-toml0.19.2 shellcheck0.35.0 es6-string-html2.12.1 vscode-lldb1.10.0 vim1.27.1 clang-format1.9.0 (2 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\ncc who was looking into this in\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nbumping to see if this issue is still occurring and if the fix has been verified!\nIt's somewhat fixed. You can try again and the second time it will work. But the error message that appears doesn't indicate that. perhaps it should retry at least once if the zip is corrupted before showing this error? Note the bottom right popup asking to manually download where a retry could've sufficed. <img width=\"1304\" alt=\"Screenshot 2023-12-06 at 6 14 41 PM\" src=\"\"\nThanks for the feedback. This bug is to fix not to cache corrupted VSIXs and please validate that. Regarding trying again, I do not think we should be showing this as this might be random.\nVerifying with a code review"}
{"_id":"q-en-vscode-c76ae30243c20ea79405025e773725ca755400e740e3a58ab269469838f62ca4","text":"Repro: an editor to a new window (or as many as you want) the original window that still contains the editor that all windows with the copied editor have input lag ! Version: 1.85.0-insider (user setup) Commit: Date: 2023-12-01T19:00:37.339Z Electron: 25.9.6 ElectronBuildId: Chromium: 114.0.5735.289 18.15.0 V8: 11.4.183.29-electron.0 OS: WindowsNT x64 10.0.\nA regression from"}
{"_id":"q-en-vscode-8796a40bb5afbcf45bee0b0be4543f578a0b4da953d0c4378f48ef0dd09779fb","text":"Type: <bFeature Request</bThe notification suggesting a plugin has a cog with an option \"Don't show again\". It's however not clear whether this suppresses suggesting the specific plugin, or any plugin. Either this message should be clarified (\"Don't show this plugin again\") or a second option should be . VS Code version: Code 1.84.2 (, 2023-11-09T10:51:52.184Z) OS version: Windows_NT x64 10.0. Modes: <!-- generated by issue reporter --\n<img width=\"485\" alt=\"image\" src=\"\"<img width=\"478\" alt=\"image\" src=\"\"\nThank you."}
{"_id":"q-en-vscode-f61e7c5187cb4ac2d0d955ba5cb054a0ed489faab959708a6c6315d6cd6854c7","text":"or VS Code Desktop something to make the autocompletion menu appear. the dash key (-) multiple times to zoom in a lot. Only the text in the menu gets larger while the html elements stay the same size. ! The html should resize with the text appropriately. Zooming in first and then revealing the menu produces the correct sizes on everything. It's possible that \"\" is the incorrect command to call in this case. Perhaps a list of commands could be provided?\nTo verify bind a keybinding to Editor Zoom In/out trigger suggest while it shows, use the editor zoom keybindings"}
{"_id":"q-en-vscode-177dcaaacbb35d31f983e15f144a8f72e9bd0f12e20de289e14f805fa13cec31","text":"Type: <bBug</bWhen first opening vscode or reloading the window several context keys are set to incorrect values. These values may eventually be corrected when focusing another editor and returning to the original one. The keys I have noticed with this issue are: It can be a little tricky to catch some of them misbehaving. It depends on which editors you have open and how they are arranged. ! Steps to repro: Open 2 editors Focus the first editor Pin it Record context keys Reload window Record context keys Diff the two sets of context keys and notice they are not the same VS Code version: Code - Insiders 1.86.0-insider (, 2023-12-18T05:36:59.983Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </detailsExtensions: none<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nBen, maybe you could check for the ones in editor area since you recently fixed some issues there. Maybe its a regression?\nThe following context keys were not being updated on reload: Updating them now on reload\nThis seems to apply to a few additional context keys outside of . Just a heads up in case closing the issue was accidental."}
{"_id":"q-en-vscode-7f5c7c09ebd3688f57599a13e4f271c776e7a59992adec2e94fbd4cf170c5773","text":"Testing The setting says \" speech recognized\" which reads a bit weird. Maybe leave \"speech\" away or says \"Controls whether the editor listens for the 'Hey Code' phrase and what to do with it\"\nBonus points for mentioning that this happens all locally\nYeah, I think I will go with this:"}
{"_id":"q-en-vscode-1492fb87a722a807492361a64f9c71b86f3b912556e882408fa257257f0b8481","text":"]- Identify accessibility bugs before check-in and make bug fixing faster and easier. #A11yMAS; #A11yTCS; #A11ySev2; #DesktopApp; #Win32; #Visual Studio Code Client; #BM-VisualStudioCodeClient-Win32-Jan2024; #WCAG1.4.3; #Color Contrast; Visual Studio Code insider version 1.86.0 Microsoft Windows 11 Enterprise 22H2 Build .3007 Visual Studio Code Insiders editor ctrl+P or Tab till go and press ENTER then TAB till Go to File. that Color contrast ratio of the \"search files by name\" placeholder text is \"2.283:1\" which is less than 4.5:1. Color contrast ratio of the \"search files by name\" placeholder text is \"2.283:1\" which is less than 4.5:1. Similar issue is also observed throughout the application. Color contrast ratio of the \"search files by name\" placeholder text should be greater than or equals to 4.5:1. Low vision users will not be able to see the placeholder text if color contrast ratio will be less than 4.5:1. ! !\nI can't reproduce this one. With the Dark Modern theme I see a background value for inputs of which passes the 3:1 ratio with the placeholder foreground of .\nMisread that the ratio should be 4.5:1. This I can fix.\nVerified the issue in below environment: Visual Studio Code Insider Version: 1.87.0-insider (user setup) Issue is fixed. Hence Closing the bug. please find attachment for reference. !"}
{"_id":"q-en-vscode-695f3a2fe516e031595ac3af498d99e3d0cb4a5759515394cffb777827988cd6","text":"Type: <bBug</bSince updating to 1.86, when I open a git diff in VSCode, I see the latest version of the file on both sides, showing no differences (except that word wrap is off on the left & on on the right... it's on everywhere else)... if I stage a file, open the diff, then unstage the file, then I see the correct diff like I've always seen before likely related: the new +- buttons next to Changes & Staged Changes show the same behavior (although while playing with it, I did have one time where I saw the correct diff on the first file) at one point I restarted VSCode, which seemed to resolve the issue, but it was only temporary. once I started editing files again the issue returned not sure if it's relevant or not, but... in all cases, the changes were saved to disk... the git change indicators in the left gutter of open files are not affected by this issue, they continue to work as the always have, showing up as I edit, clearing when I stage the change VS Code version: Code 1.86.0 (, 2024-01-31T10:28:19.990Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (55)</summaryExtensionVersion ------ vscode-icalendar1.0.1 vscode-m3u1.0.0 vscode-caniuse0.5.0 case-change1.0.2 rtf2.7.0 spellright3.0.122 markdown-checkbox0.4.0 markdown-emoji0.3.0 markdown-mermaid1.21.0 markdown-yaml-preamble0.1.0 vscode-tldr1.0.0 gitignore0.9.0 vscode-markdownlint0.54.0 vscode-eslint2.4.4 rushcore1.0.2 rushnav1.0.2 githistory0.6.20 xml2.5.1 escaping-characters1.0.0 EditorConfig0.16.4 vscode-macros1.4.1 git-project-manager1.8.2 url-encode1.1.0 shell-format7.2.5 matlab3.0.2 vscode-pull-request-github0.80.0 gc-excelviewer4.2.58 vscode-markdown-footnote1.0.0 vscode-favorites1.11.0 rest-client0.25.1 path-autocomplete1.25.0 mediawiki2.1.0 anki1.3.1 markdown-katex0.1.4 vscode-peacock4.2.2 gpg1.1.1 zotero0.1.11 markdown-shortcuts0.12.0 debugpy2024.0.0 jupyter-keymap1.1.2 jupyter-renderers1.0.17 vscode-jupyter-cell-tags0.1.8 vscode-jupyter-slideshow0.1.5 azure-account0.11.6 azurecli0.6.0 hexeditor1.9.13 live-server0.4.13 vscode-yaml1.14.0 vscode-zipexplorer0.3.1 pdf1.2.2 intellicode-api-usage-examples0.2.8 vscodeintellicode1.2.30 vscode-icons12.7.0 org-mode1.0.0 JavaScriptSnippets1.8.0 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nPlease use the Start Extension Bisect command to investigate whether the problem is being caused by one of your extensions.\nI can duplicate this in Insiders with no extensions loaded (... however, that opening & closing revealed more details I didn't realize previously... I only see the buggy behavior with changes made in this session... if I close & reopen VSCode, changes that were staged before VSCode opened render correctly... new changes stop rendering when they're staged (and when I say \"stop rendering\" I mean no red & green highlights, because both sides of the diff tab contain the new code) & start rendering again when the app is restarted (or I switch between Release & Insiders) side notes: I'm no longer seeing the no word wrap on the left in VSCode (& I never saw it in Insiders), so not sure what was up with that... I know that there are often visual artifacts in one side of the diff editor (e.g. valid relative paths get yellow squiggles on one side but not the other), so guess I'll put this one in that category... who cares? I'll see that in the main editor, I just want the diff in the diff viewer I forgot to try Insiders before I submitted because it was so obvious that only 1 thing changed, but I'll try to remember that just to make sure it hasn't been fixed already... but, I don't appreciate being asked to go on a wild goose chase bisecting extensions\none more thing I just noticed... prior to 1.86, if I stages a file, then made additional changes, so that the file is now in the Source Control panel twice... the 2 diff buttons would open 2 separate diffs, one with the staged changes & 1 with unstaged... now I get only one (with the unstaged changes) also, it makes no sense that the bot removed the label, unless there's been another Release since 1.86, and the Releases page still shows that as the newest version\nI've had this issue since February too. Strangely enough, I started an Extension Bisect, stopped straight away (intending to restart) and now I don't have the issue. ¯(°_o)/¯\nI'm hitting this too on two completely different machines (Windows and Linux). The issue is if you click the diff button on a staged file just after staging it, for some reason it opens the working tree diff (which has no differences). If you reload the window so it starts with the file already staged then the button correctly opens the index diff. I reproduced this with no extensions enabled. I think the reason this enormous bug has gone unnoticed for so long (I only just looked into it because I assumed that surely they would fix something as broken as this...), is that it only happens if you have this setting: Do you guys have that set?\nyes, I do have that setting... and if I comment it out, I see the correct behavior\nSorry for not getting back to you on this. Could any of you experiencing this issue provide a screen recording illustrating the issue ()? Thank you very much in advance!\nUgh I can't seem to reproduce it at the moment - sometimes it does work, but I haven't found the exact pattern. So you'll have to use your imagination, but basically you click the diff button for a file (two circles connected by arrows) and it opens a diff view but there are no differences between the files on the left and right of the view. Next time it happens I'll try to record it.\nAre you able to reproduce the issue if you disable the setting? Thanks!\nDidn't seem to make any difference - I still can't reproduce it, sorry. What does that setting actually do? It doesn't seem to make any difference to behaviour and I don't understand the description: ! The Source Control view still automatically updates even if I have that off.\nif I add to my settings I can repro, without it I can't repro... just recorded with gitcap, waiting for it to render so I can get a link\nvideo was way over GitHub's 10MB limit... so dropped it in my OneDrive & tried to include it as an image... !Auy35hyaQsjmnpB70oGx2DvOlOjaXg?e=YLgT5J!\ncould you please re-upload the recording as it looks like it is not available at the link any more. Thanks!\ni will try tonight... on vacay this week, so might need to wait until I'm back on the work PC Monday\nany news on the recording? Thank you!\nwith in settings (without it I see expected behavior) !\nthanks again! I have a fix but it will not make the cut for the August 2024 release so it will be in the September 2024 release.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nFriendly ping! Looks like this issue requires some further steps to be verified. Please provide us with the steps necessary to verify this issue.\nVerification steps: VS Code Insiders to a file, modify it and save it the Source Control view and click on \"Open Changes\" action Confirm that the diff is rendered correctly on the \"Stage Changes\" action on \"Open Changes\" action Confirm that the diff is rendered correctly"}
{"_id":"q-en-vscode-9ad36cd7c3cbf176e3f42eaeb6c74b2bfaba03debdd3d1dedc1f0e1076160fb3","text":"<img width=\"431\" alt=\"image\" src=\"\"The debug view uses monospace font by default other than the workbench font"}
{"_id":"q-en-vscode-940a961707ac5f80b9f156d060c160f2389b926b48aad2aa6014544b899ebfca","text":"Enable Write a long commit message that triggers validation Hover over the diagnostic squiggle <img width=\"405\" alt=\"image\" src=\"\""}
{"_id":"q-en-vscode-6ee2e8d83910813ba82917dd319082cecfc29b4915d99b9df5f5596170bd4ba7","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.86.1 OS Version: Ubuntu 23.04 Steps to Reproduce: an empty text editor the text opening and closing an empty text editor several times that the number of disposables increases each time Opening and closing an editor 1997 times results in 3994 disposables with this stack trace: When opening an editor, the function is called , the function calls adds a disposable to the disposable store When closing and opening the editor again, the same happens: the function is called , the function calls adds a disposable to the disposable store again This probably causes more and more disposables to be in . It seems, disposing the codeEditorContributions disposable () when a text editor model becomes detached might be a possible solution. That way, it would be registered when a model is attached and disposed when a model becomes detached:\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.86.2. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nThanks a lot!"}
{"_id":"q-en-vscode-ba9e610aaad8c4df3322d938125598143ddc1aebd9f3e32f010313018244b9ed","text":"In VS Code action items inside inputs are vertically aligned to top, in the Chat features they are centered in the middle. This happens when putting in a lot of text or pressing Shift+Enter. Panel: <img width=\"425\" alt=\"image\" src=\"\"Inline: <img width=\"608\" alt=\"image\" src=\"\"Quick: <img width=\"621\" alt=\"image\" src=\"\"For Inline and Quick the action items to the right of the input box should also stick to the top.\nSeems like we have a mix of top-aligned and overflow w/ scroll (see below). But I agree that we should be consistent as much as possible. But it does feel like a chat send button should generally but aligned to the bottom, no? We may even consider rendering the toolbar on its own row once we're into multiple lines. That would ensure the text isn't forced to fill a tiny space now that we have 2+ icons. Here's how Teams does it as an example: ! ! ! !\nWould be awesome if the text could wrap around it, especially in the narrow search widget case, I've always wanted that but I think it would be a pain to implement.\nDon't have a strong opinion about top or bottom, I think each can be justified on a case-by-case basis. but middle looks weird. :) For example, inline Chat may want to do top alignment as the icons would align with those outside the input field (poppet and close) Bottom would help with the wrapping."}
{"_id":"q-en-vscode-464cc3ca4d54ac01634733d95c572a66710ca6eacbb0b840dc2de7283312b6d2","text":"use generate cell in notebooks make a request while the request is active press stop a bunch of errors are printed\nIt seems not wise at all to maintain my own *chatcontroller. I'd love to look into how do I ensure I reuse the most of Inline Chat Controller later.\nIs the generate cell feature just with using ctrl+i from within a cell body? Not sure how I can trigger it."}
{"_id":"q-en-vscode-1979159571b638dddbfc213cb87b906f00cf409433bbfb5074a8d2a5e331c865","text":"This only applies in a multi-root workspace. With just 1 view: ! With N views: ! I think we should use the workspace label here. to you recall why this code is there:"}
{"_id":"q-en-vscode-afcded0d1c0e0fddd41fb47ed2a34f04fdcf1ae731e1c3cb87ff707fcb6184bc","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.86.2 OS Version: Windows 10 / macOS Sonoma Steps to Reproduce: at least three windows the command least one of the windows will show a strange close icon, instead of the regular one, like the image below ! I reproduced the same error on Windows, and it also presented a strange icon, but it was a different one. A small vertical rectangle, like that missing unicode chars. I tested the on Insiders release, at least on MacOS, and didn't notice the error. I didn't have the opportunity to test in Insiders on Windows yet. I'll update the issue later with more details. Thank you\nare you using a product icon extension?\nNevermind, a regression from , fyi\nHi , Glad to know you could find the fix. Let me know if you need me to verify it when released. Thank you\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\n/verified"}
{"_id":"q-en-vscode-fc455764b5224924451d69c4e1cf5baacddd3f9113c5b0e9a9f458fd4e089aff","text":"To be able to reach to the files presented in a multi diff text editor the tab model should provide a corresponding TabInput (see ) May be a good idea would be to have a general composite tab input. For the multi diff text editor the elements could then be\nWhen you listen for tab changes the input of the tab is currently .\nNot sure if this is or\nRelated:\nThanks!\nReopening to get API approval. I'm also not sure we will be able to expose the resources in multi diff This will probably complicate your use case. pointed out that the resources can change without changing the input and this is not something any of other inputs do making supporting showing the resources for this input quite difficult.\nCan you verify? To verify please ensure that you see a type now for the multi diff text editor and it properly reacts to resource changes"}
{"_id":"q-en-vscode-843390bd9a8f1b9d6d9e4d70fcd873f166774f12f114faa191915edd19ab406e","text":"This error appears in the insiders dev tools all the time (FYI Cannot read properties of undefined (reading 'includes'): TypeError: Cannot read properties of undefined (reading 'includes') at at at at at at at at at at at"}
{"_id":"q-en-vscode-8553d62cc4c61e9b22dc895befa9d6b47a4fad4cf3716fc97ce0ebb65c532543","text":"Type: <bBug</bIn the below image, clicking the link Learn More brings you to However, that section (id) does not exist on The link is most likely supposed to go to ! VS Code version: Code 1.87.2 (, 2024-03-08T15:20:17.278Z) OS version: WindowsNT x64 10.0. Modes: Remote OS version: Linux x64 5.4.0-173-generic <details<summarySystem Info</summary </details<details<summaryExtensions (32)</summaryExtensionVersion ------ yang0.1.3 jupyter-keymap1.1.2 remote-containers0.348.0 remote-ssh0.109.0 remote-ssh-edit0.86.0 remote-wsl0.86.0 vscode-remote-extensionpack0.25.0 remote-explorer0.4.3 remote-server1.5.1 vscode-openapi4.25.1 vscode-fish1.0.38 ruff2024.16.0 vale-vscode0.18.5 vscode-markdownlint0.54.0 gitlens14.9.0 prettier-vscode10.1.0 code-runner0.12.1 go0.41.1 restructuredtext189.3.0 direnv0.17.0 vscode-azureresourcegroups0.8.4 vscode-cosmosdb0.20.1 vscode-docker1.29.0 vscode-kubernetes-tools1.3.15 python2024.2.1 azure-account0.11.6 vscode-yaml1.14.0 code-spell-checker3.0.1 even-better-toml0.19.2 shellcheck0.37.0 simple-rst1.5.4 vscode-proto30.5.5 </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --"}
{"_id":"q-en-vscode-40af96f7134978390c597c339cac56189c9da928eea9a3b961093fd717e497b4","text":"Type: <bBug</bCC command works only from inside the Accessible View; it does not work when invokved from the code block area of the Copilot Chat View. VS Code version: Code - Insiders 1.88.0-insider (, 2024-03-25T05:49:32.344Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (119)</summaryExtensionVersion ------ vscode-openapi4.25.1 zotenote1.0.1 android-dev-ext1.4.0 aiprm-lang0.0.2 Bookmarks13.4.2 openscad1.2.2 spellright3.0.122 mermaid-markdown-syntax-highlighting1.6.0 doxdocgen1.4.0 vscode-markdownlint0.53.0 vscode-eslint2.4.4 vscode-quick-select0.2.9 vscode-deno3.35.1 gitlens14.6.1 EditorConfig0.16.4 prettier-vscode10.1.0 vscode-firefox-debug2.9.10 shell-format7.2.5 vscode-google-translate1.4.13 codespaces1.16.17 copilot1.175.784 copilot-chat0.14. remotehub0.62.0 vscode-github-actions0.26.2 vscode-pull-request-github0.85. cloudcode2.6.0 overleaf-workshop0.11.0 cslpreview0.2.2 easy-snippet0.7.5 path-autocomplete1.25.0 latex-workshop9.19.1 lilypond-syntax0.1.1 scheme0.2.0 better-cpp-syntax1.17.2 commitlint2.5.1 language-julia1.75.2 google-search0.0.1 vscode-lua-format1.3.8 lilypond-formatter0.2.3 lilypond-pdf-preview0.2.8 lilypond-snippets0.1.1 vslilypond1.7.3 git-graph1.30.0 azure-dev0.8.1 vscode-azureappservice0.25.2 vscode-azurecontainerapps0.6.1 vscode-azurefunctions1.14.0 vscode-azureresourcegroups0.8.3 vscode-azurestaticwebapps0.12.2 vscode-azurestorage0.15.3 vscode-azurevirtualmachines0.6.5 vscode-cosmosdb0.21.0 vscode-docker1.29.0 vscode-edge-devtools2.1.5 black-formatter2024.1. debugpy2024.3. flake82023.10.0 isort2023.13. python2024.3. vscode-pylance2023.12.103 jupyter2024.1. jupyter-keymap1.1.2 jupyter-renderers1.0.17 vscode-jupyter-cell-tags0.1.8 vscode-jupyter-slideshow0.1.5 remote-containers0.352.0 remote-ssh0.110. remote-ssh-edit0.86.0 remote-wsl0.81.8 vscode-remote-extensionpack0.25.0 azure-account0.11.6 azure-repos0.38.0 cmake-tools1.17.17 cpptools1.19.9 cpptools-extension-pack1.3.0 js-debug-nightly2024.3.2217 powershell2024.3.1 remote-explorer0.5. remote-repositories0.40.0 remote-server1.6. vscode-github-issue-notebooks0.0.130 vscode-node-azure-pack1.2.0 vscode-selfhost-test-provider0.3.25 vscode-serial-monitor0.11.0 vscode-speech0.7. vscode-speech-language-pack-ko-kr0.1.0 vsliveshare1.0.5918 windows-ai-studio0.2.4 autodocstring0.6.1 pandocciter0.10.4 shiny-python0.1.6 shinyuieditor0.5.0 quarto1.111.0 r-debugger0.5.4 java1.28.1 vscode-xml0.26.1 vscode-yaml1.14.0 r2.8.2 multi-command1.6.0 vscode-deepl1.1.1 abc-music0.4.0 lua3.7.4 latex-utilities0.4.13 cmake0.0.17 vscode-terminal-here0.2.4 windows-terminal0.7.0 errorlens3.16.0 intellicode-api-usage-examples0.2.8 vscodeintellicode1.2.30 vscode-conventional-commits1.25.0 vscode-arduino0.6. vscode-java-debug0.56.2 vscode-java-dependency0.23.5 vscode-java-pack0.25.15 vscode-java-test0.40.1 vscode-maven0.44.0 zotero-cite0.8.2 markdown-all-in-one3.6.2 grammarly0.25.0 (1 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nI see this when I try that !\nLooks like we should be overriding that or going with a different keybinding.\nNote that this also doesn't work when in the accessible view. I believe has somehow worked around that by tweaking his keybindings / clauses.\nYou are correct. I had to tweak around my keybindings to make it work."}
{"_id":"q-en-vscode-ca2e1986b82b6a9002b4ff66253e1f6e10b6828f40a13a300bf503a592e5ea1e","text":"Just asked hi: ! Expected: ! Version: 1.88.0-insider (user setup) Commit: Date: 2024-03-25T05:49:32.344Z Electron: 28.2.6 ElectronBuildId: Chromium: 120.0.6099.291 18.18.2 V8: 12.0.267.19-electron.0 OS: Windows_NT x64 10.0. GitHub Copilot Chat v0.14. (pre-release)\nbecause of\nTo verify configure have a very wide view port ask AI to change some code make sure the input widget is full width of the zone widget"}
{"_id":"q-en-vscode-a2c2c5a84340dab02d8b6b2137e1005eb7e8e9d9d94df50a190f294f687f5f4e","text":"Testing ! !\ndoesn't reproduce in the panel chat\nI somehow suffer from this change which explicitly targets some monaco tokens...\nThis isn't nice but there's actually no other way to override the foreground color of an editor- because typically you have syntax highlighting. We do this for all the other \"input box\"-style editors. I think I opened a feature request when I did this. We should align this rule with your rule for coloring the decoration so that it can override it properly.\nlabeling so it gets verified\nThanks\nin the latest insider's, still seeing this ! !\nThere is two separate issues with different steps. There is a difference between rendering of slash command and normal prompt text. the only useful/visible information in the original issue description is slash commands (please use words and not images to describe issues!). For verification you have not used slash commands but normal prompt text which is a different case. Please file a separate issue, use words/steps and don't rely on screen captures with just white on white text\nSteps to reproduce/verify use HC light and dark theme open inline chat and type a slash command, e.d make sure the command is readable when selecting its text"}
{"_id":"q-en-vscode-98d5fa281e19d6f511ddce1049b52e1a6a3f7a7be75621528527bf90379b73cd","text":"Testing In the screenshot below, \"Chat: Insert at Cursor\" should be \"Chat: Insert into New File\" !"}
{"_id":"q-en-vscode-72ec8f969b723bd59fe2c7795d7b87a43ee9efda6a763a593be576b2ad4fd7fa","text":"Testing to filter e.g. Python Notice the non-extension entries remain at top Maybe this was by design, but I don't understand it. If I type python, don't I want to quickly select a python related entry?"}
{"_id":"q-en-vscode-42793c308828bd72ea66e76f96e92e509502fc4219828df390e69e98a8ebbb5d","text":"Note that both of these scenarios used to work. In the terminal inine chat, run a command like which yields just a chat message response. Try to tab to it. one cannot Run a command like which yields a code block response. Try to use to go from the input box to a code block response and vice versa. one cannot\nNote that the keybindings are important because it's annoying as a SR /keyboard only user to have to tab through a bunch of things to reach the response.\nThis works for me, I do the following with success cmd+i, paste , press Enter wait for the response press tab repeatedly and reach the response\nthe tabbing part works for me in today's insider's\nI'm able to spam Ctrl+Up on Windows to focus between the inputbox and the chat response, but I'm not able to spam Ctrl+Down to do the same. I'm unable to use either shortcut to focus from the inputbox directly into a code block within the chat response. Let me know if the behaviour is as expected or if there's some other steps I can follow.\nI haven't tried windows yet. This won't directly focus the code editor, just the response.\nthis works on mac, could you pls try on windows?\nThe behavior seems wrong now and is quite disorienting: Input: ctrl+up will focus the response list ctrl+down will focus the response list Response list: ctrl+up does nothing ctrl+down does nothing Response code block: ctrl+up focuses the input ctrl+down focuses the response list I would expect: Input: ctrl+up does nothing ctrl+down will focus the response code block Response list: ctrl+up focuses the input Not sure? Response code block: ctrl+up focuses the input ctrl+down focuses the next code block, otherwise not sure\nThis issue was about cmdCtrl+up/down going between the response list and the input. This is unexpected, but unlikely to be attempted by the user IMO. Input: ctrlCmd+up will focus the response list This works on mac. Not sure why it doesn't on windows. Response list: ctrlCmd+up does nothing I didn't implement this yet, but can in April. Response code block: ctrl+up focuses the input ctrl+down focuses the response list\nverified that this does work. I'm making a new issue about focusing the code block(s) and the bug with in the input focusing the list"}
{"_id":"q-en-vscode-2264476a1dac391001bc378390866d76ef33e8c89269f000807bb802a9dcdf5f","text":"using Default Light + theme run peek references <img width=\"978\" alt=\"Screenshot 2024-03-26 at 17 30 37\" src=\"\""}
{"_id":"q-en-vscode-10a9e53cc9b04e2915d3890970ecd46a58487f7a3fe7b91d92a0e88bee273a42","text":"Sometimes the default width for inline chat seems very wide: <img width=\"854\" alt=\"image\" src=\"\"Maybe it could grow while typing and or allow to be resized.\ndecided against dynamic and going with a more reasonable default width of 360px"}
{"_id":"q-en-vscode-1b3bcf8125e6046d0ce0379f938f8a06f6f3f4a225ed7202bbb948ac7d9c2cc0","text":"Testing Sometimes when I hovered over the button it would show the hover of that button over the button. When moving down to the button it would close the editor hover sometimes because the mouse would move over the hover. This could be solved by setting an anchor for the and hovers"}
{"_id":"q-en-vscode-9f9a2900d3e204a4e6bf9ef9e08bfcf3f07630a552f4ea1831b605479d08fa5c","text":"Testing The keybinding does not work on windows due to quick open. Also the keybinding collides with the command. I wonder if this could be solved without registering keybindings. One idea could be that when the hover is focused the user could just use and on the keyboard to increase and decrease the hover (possibly also arrow keys could work). Make sure to mention in the hover of the increase and decrease actions what the keybinding or keyboard key is.\nThat's a very good point. I think we would still want ideally the + and - to be linked to keybindings so they can be configured by the user. I'll remove the keybindings, so they can simply be set by the users if they so wish. The current code already renders the keybindings if they are set: Perhaps they are not fetched correctly in your case because of the keybinding collisions. !"}
{"_id":"q-en-vscode-09df075b28b8eb7a9ac0a8e21ba4da59b1f907ad5bd92ad6c06d7a7ab4a68f6b","text":"Testing on Linux I have 2 monitors: primary and secondary. I opened VS Code and maximized it on my secondary monitor. Then I dragged an aux window out and maximized it on my primary monitor. After closing and reopening VS Code, the maximized state of both my main window and my aux window were restored, but both were on my secondary monitor. The aux window was on top. I'm using Remote Desktop + Hyper V, so maybe something is going on there: Remote Desktop from Windows to another Windows machine. Hyper V to an Ubuntu VM.\nLooks like which monitor is primary and which is secondary doesn't seem to matter as the same thing happens when swap the windows that aux and main are on.\nI also experienced a variant of this: Main VS Code window maximized in primary monitor, 1 aux window maximized in primary monitor, 1 aux window maximized in secondary monitor Close VS Code Reopen VS Code All 3 windows (main and aux) are restored to max, but they're all restored to my primary monitor I'm testing on Windows 11 Insiders."}
{"_id":"q-en-vscode-cfe4b11a46225c0283b9344c7675890f59bc9163d2e3057fbba92ef1ce9c181b","text":"Testing Steps: Use Windows PowerShell (not PowerShell Core) Turn on the terminal intellisense Type stuff is to the terminal"}
{"_id":"q-en-vscode-2b4614531cc1d93b5bcd5001cc2e7db64b0d262e186b4269d7346a9fbbb8ce3e","text":"Testing ! This happened when updating the file on disk outside of vscode, I'll try to find a consistent repro\na notebook with several markdown headers in insiders the same notebook in stable and add a header :bug: outline in insiders is not updated (stable does update when doing this in reverse) a different markdown header in insiders :bug: screencheese\n!\nThanks I can replicate this issue.\nGiving the issue back to Aaron to verify. My Python install's a bit broken."}
{"_id":"q-en-vscode-a9ec82f7ae288934849e0387bcb67ffb3fc8f1e0e9d46fa1be32b8202241a34c","text":"It seems that works workspace-relative unless an absolute glob pattern is used. I would clarify that from the settings description."}
{"_id":"q-en-vscode-3e49ee3b192fb54089a25de8f37350a01f63a07e449e305e8320b523b90bd597","text":"Type: <bBug</ban existing remote project an embedded terminal in VSC \"PORTS\" tab shows 51 ports This is a vanilla-flavored nodejs service that uses one dedicated port determined by the remote environment (an AWS EC2 instance running Rocky Linux) -- and that dedicated port is not opened until the service is started. I expect at most 2-3 ports to be open at any one time. VS Code version: Code 1.88.1 (, 2024-04-10T17:41:02.734Z) OS version: WindowsNT x64 10.0. Modes: Remote OS version: Linux x64 4.18.0-477.27.1.el88.x8664 <details<summarySystem Info</summary </details<details<summaryExtensions (12)</summaryExtensionVersion ------ remote-ssh0.110.1 remote-ssh-edit0.86.0 remote-explorer0.4.3 java1.30.0 intellicode-api-usage-examples0.2.8 vscodeintellicode1.3.1 vscode-java-debug0.57.0 vscode-java-dependency0.23.6 vscode-java-pack0.26.0 vscode-java-test0.41.1 vscode-maven0.44.0 markdown-pdf1.5.0 </details<!-- generated by issue reporter --\nCan you connect to your remote then share the following: the value of the setting the value of the setting A screen shot of the Ports view\nI appreciate your prompt attention, thanks for looking at this. As you can see in the attached \"Ports view\", this morning there were 53 open ports at startup. I've attached a screenshot of the \"Settings\" view (searching for \"remote.autoForwardPorts\") and the \"Ports view\". I looked in the local \"\" (in ) and see no settings with the two names you requested. I similarly found nothing in the relevant \"\" file(s) on the remote server. Hopefully the attached screenshot of the \"Settings\" view provides the information you requested. ! Settings ! Ports view\nis it possible you at some point ran something that listened on each of those ports? By default, VS Code will auto forward ports, but with the setting set to , which is what you have, we don't know when to stop forwarding the port. We also restore ports (setting ). I think the best thing we can do is change the default for SSH for to , which will let us know when to stop forwarding. I've made the appropriate PRs to get that changed.\nIf something ran that was listening on those ports, it was unknown to me. I don't knowingly do anything with ports on the local (Windows 10 pro) system. I manage multiple ports on the remote system, but each has a 4-digit port number. An aspect that may be of interest is that the startup behavior seems to be specific to this remote folder (). When I use VSC to open a sibling on the remote host (), the embedded terminal shows no forwarded ports at all. Screenshots below show this behavior. ! Startup VSC on ~/tms/backend/gate ! Ports view on ~/tms/backend/gate\nWe restore ports based on the folder, so it makes sense that you're seeing different ports depending on which folder you have open.\nHmm. It seems likely that at some point something was running and listening on those ports, otherwise we wouldn't have found them and auto forwarded them. I think the change to use by default that I mentioned above will resolve your issue in the future, but for now, I would suggest you unforward all those ports and set to .\nThis issue has been closed automatically because it needs more information and has not had recent activity. See also our guidelines. Happy Coding!"}
{"_id":"q-en-vscode-983a2412360bf360239b2635de19c65e7d279019a863783e1993b8211599ef0a","text":"Using the workspace TS version, 5.5.0-dev.: ! ! !\nClosing for verification"}
{"_id":"q-en-vscode-60423e83be56488c703bd71c34d8e30f5cef00ec4234f02ec4b079e89f7afe5f","text":"I got the default font rendered in the parameter hint widget. It looks alien compared to the rest of the hover widget UI. Steps: chat sample line 21 type the following: will see the parameter hints with a weird font (default system?) ! fyi\nI haven't been able to reproduce this Are you still seeing this issue? Does it only show up with specific functions?\nThis issue has been closed automatically because it needs more information and has not had recent activity. See also our guidelines. Happy Coding!\nYes, I still see this. I see this for all parameter hovers. I have nailed this down to be due to this setting I got a new mac, so most likely I do not have Fira Code installed. So the Editor seems to fall back nicely to a good default font, and Parameter Hints falls back to a non coding font. Which looks like the gist of this bug. I will reopen this so you can chime in. Thanks"}
{"_id":"q-en-vscode-d6305dbdac5bb11146f404bab10b4c3c6209dd2b96dba4b35069942b6d6f191b","text":"Type: <bBug</bThe terminal in VSCode - insiders is unable to display special characters in terminals after latest update How it looks in VSCode: ! How it looks in VSCode - insiders ! The font settings are the exact same for both versions VS Code version: Code - Insiders 1.90.0-insider (, 2024-05-03T05:49:15.006Z) OS version: WindowsNT x64 10.0. Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (40)</summaryExtensionVersion ------ better-comments3.0.2 TabOut0.2.2 Bookmarks13.5.0 doxdocgen1.4.0 githistory0.6.20 gitlens14.9.1 copilot1.184.0 copilot-chat0.14.1 latex-workshop9.20.1 rainbow-csv3.11.0 git-graph1.30.0 autopep82024.0.0 debugpy2024.4.0 isort2023.10.1 pylint2023.10.1 python2024.4.1 vscode-pylance2024.4.1 jupyter2024.4.0 jupyter-keymap1.1.2 jupyter-renderers1.0.17 tensorboard2023.10. vscode-jupyter-cell-tags0.1.9 vscode-jupyter-powertoys0.1.1 vscode-jupyter-slideshow0.1.6 remote-wsl0.88.2 cpptools1.20.3 cpptools-extension-pack1.3.0 hexeditor1.10.0 vsliveshare1.0.5918 autodocstring0.6.1 advanced-new-file1.2.2 material-icon-theme5.1.0 r-debugger0.5.4 r2.8.2 markdown-preview-enhanced0.8.13 errorlens3.17.0 intellicode-api-usage-examples0.2.8 vscodeintellicode1.3.1 markdown-all-in-one3.6.2 material-theme3.17.2 (1 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\ncan you share your Also are you using Windows PowerShell () or pwsh 6+? Could you get the version of that via ?\nSame issue for us using bash 5.0.17(1) so might not be pwsh dependent. Terminal prompt now not showing special characters on latest insiders update 1.90.0 e.g. x1b[01x3b34m~/code/project1x1b[00m (main))\nactually got a fix in for bash in\nWhere can i get the The powershell version is: 7.4.2\nI have the same exact problem, and the following suggestion finally fixed it on my end:\nI am not sure I understand what the comment is trying to say. I am supposed to replace my with the one from commit and \nI agree, the instructions could have been clearer. The commit selection does seem a bit random, but what I did was clicking 'Browse files' on the commit and found the from that repository version. Here's a link to get there directly: \nIt was pushed to the official build. I'm having this issue and I dont rememeber ever joining any kind of insider program.\nRelated on Stack Overflow:\nthank you so much! This was what fixed it for me after trying everything yesterday. Some more clarity or directions from the dev side would have been nice, but now its back to normal after all!\nThanks replacing the shellintegration.ps1 from this commit fixed it for me. The location of the file is:\nIn addition to the semi-permanent solution above, I have found that you can simply type (for PowerShell 7) or (for PowerShell 5) in the terminal, and it will launch a brand new, yet identical terminal in the current working directory - but without the gibberish.\nFYI, I see this in the regular code branch Version: 1.90.0 (user setup) Commit: Date: 2024-06-04T19:33:54.889Z Electron: 29.4.0 ElectronBuildId: Chromium: 122.0.6261.156 20.9.0 V8: 12.2.281.27-electron.0 OS: Windows_NT x64 10.0. !\nTo verify: oh-my-posh a PowerShell terminal oh-my-posh, you should not see the escapes as shown above\nSame issue here. oh-my-posh, powershell, issue only with vscode, windows term looks as intended. !\nThis should be fixed in v1.90.2 which was just released\nCan confirm as fixed. Thanks for the work!\nconfirm as fixed and work well without needing to disable setting! !\nLooks good again!\nLooks good to me too!!\nThis morning, I updated my VS Code, activated the setting, but I am still experiencing the same issue (especially when I try to clear the terminal with CTRL+L). I am using Windows 10 Pro for Workstations x64, AMD Ryzen 5 5600X 6-Core Processor, and Nvidia 720 GPU. !\nStill same issue\nwhat shell/prompt is that?\nIt's the PowerShell from the Microsoft Store. Special characters appear when I press Ctrl + L to clear the terminal. I also have the same problem in Git Bash, but it happens when I execute a command. In both cases, I am using Oh My Posh. !\nI was also having a related? problem with PowerShell integration since the 1.90 release. The recent 1.90.2 update has changed the outcome a little, but not fixed it. I get output like the below when I open a new pwsh terminal, and often the terminal stops responding after executing a command or two. I have this PowerShell profile. I thought any of PSFzf, zoxide, or the encoding settings could be the culprit, so I tried removing each one. Oddly, removing any combination of them doesn't fix the problem, but completely blanking out the file fixes it (second image below). It all worked fine in vscode prior to 1.90. ! !\nyou have the experimental suggest setting enabled, disabling that will fix it. What version of PowerShell is that you're using? Currently in Insiders we force suggest to be disabled on Windows PowerShell (5), but if you're experiencing it in a latter version maybe we need to disable it in older versions of Windows too.\nAre you referring to a different setting? (I was wrong; see below.) (Not sure if it's relevant, but PowerShell suggestions do work in Windows Terminal.) ! !\nOOPS, I did have an old version of that setting in my , and removing it did indeed fix the problem. Thank you.\nIs there a separate issue tracking this behavior with ? I do not use OhMyPosh and am still getting the same behavior with VSCode . ! The workaround of invoking from within the VSCode terminal does work."}
{"_id":"q-en-vscode-0b4b0c370edb18abb66d67bce6e09dfa4cbb0f7e1a194a675d909b86ef376007","text":"This extension wants me to install it, also has a reload, and has a restart button. The restart button doesn't do anything. This is just in the VS Code repo <img width=\"476\" alt=\"image\" src=\"\"\nCan you please share the following file F1 Open Extensions Folder Share file\nWhat do you see in Installed Extensions view? Also What version of that extension is running in Running Extensions editor - F1 Show Running Extensions? Is there a message when you hover on Restart Extensions button?\nInstalled <img width=\"470\" alt=\"image\" src=\"\"Running view <img width=\"917\" alt=\"image\" src=\"\"Hover of restart <img width=\"470\" alt=\"image\" src=\"\"Ok, so it does appear I have this extension installed. And it wants me to restart it to uninstall it? And then to install it as a workspace extension? Clicking the restart button doesn't do anything and I didn't request the extension be uninstalled\nThis issue has been closed automatically because it needs more information and has not had recent activity. See also our guidelines. Happy Coding!"}
{"_id":"q-en-vscode-34653015ac83502049a1315f79dfd2bb91a8a1ea7b985075069e81894f34273e","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.89.0 OS Version: macOS Sonoma 14.1.2 arm64 This issue coming from Cc Flutter extension from marketplace /Create Flutter project in Vscode the app by clicking the run button inside the IDE until the app is running on platform (eg emulators) ## Expected behavior Hot reload button will appear :zap: <img width=\"405\" alt=\"Screenshot 2024-05-07 at 12 09 20 PM\" src=\"\"Hot reload button doesn't appear <img width=\"537\" alt=\"Screenshot 2024-05-07 at 12 06 06 PM\" src=\"\"Toggle brings back the Hot Reload Button\nHey are you able to repro this? I set up a test extension that contributes a command the same way that the Dart extension does, and it works fine for me. I can try in the Dart extension too, but I'm wondering whether it's a fluke for this user or something that happens regularly.\nAh nevermind, I guess I do have a repro using js-debug\nHey I can't figure out how this is supposed to work in the command center. It seems that we don't listen to updates of the debug submenu inside the command center. Easy repro steps Put debug toolbar in the CC Start debugging something with node Run \"Take Performance Profile\" The red button to stop profiling doesn't appear in the debug toolbar unless I do something like move it\nWeird, we should receive events and redraw the menu because of the -flag.\nDebugged this for awhile, and I don't know how it's supposed to work, but it basically doesn't load the context keys for extension-contributed menu items because nothing is prompting the CommandCenter menu to reload context keys for submenu items after the menu extension contrib point is handled. create a Menu for CommandCenter, and it has a MenuInfo: picks up the context keys for whatever exists at that moment, and we don't have the extension menu items yet we process the menus extension point and add those items to the MenuRegistry for the DebugToolbar But at this point, something needs to bubble this up to the CommandCenter. It doesn't know that its submenus are changing. The Menu should fire onDidChange at that moment, and I don't see how that would happen. And, the CommandCenter Menu should call somewhere in there. Something is definitely missing. I didn't debug the normal submenu case but I'm guessing the inlined rendering of CommandCenter's submenus is related somehow, I assume in a normal case the context keys for submenus would get updated in some other way. Maybe when you're back we can jump on a call and I can show you what I'm seeing."}
{"_id":"q-en-vscode-73d4e35c02a330b2e5dac712eda77630aabbbd8ddb1586684b687abd470df115","text":"<!-- Do Not Delete This! featurerequesttemplate --<!-- Please read our Rules of Conduct: --<!-- Please search existing issues to avoid creating duplicates. --<!-- Describe the feature you'd like. --Related For migrating towards ESM in VSCode, it seems the tsconfig option could be useful, ensuring that imports are compatible with ESM. For example, this typescript code transpiled to javascript would cause a runtime error because of the export: The rule would enforce a type only import for : Enabling the tsconfig option in VSCode, it seems these files would need to adjusted:\nWe reverted this because we learnt, albeit very late, that it has disabled inlining of const enums. See"}
{"_id":"q-en-vscode-6813a974b05f1eeb892f0d06d8de86d228e913091f4794b77f8c67242cc72864","text":"Personal preference - I do not like the border around the attached context. Does not align with the rest of the UI. Do we really need the border? Have we tried this without the border? ! fyi\nYeah, I've tried this because I don't love the weight it adds either. It does cause some weirdness when there are follow ups and lots of pills. But I'll keep playing with it to see if I can get to a place where we don't need it. !\nThank you David!"}
{"_id":"q-en-vscode-1b50d94adc5b62c5f9230ad6a43eb46ea5273c698acc22d3b3eda22bcb96e03a","text":"Type: <bBug</bI ccould not close inline chats or quick picks anymore. Looks like captured all esc shortcuts. I had used voice input during the session. VS Code version: Code - Insiders 1.90.0-insider (, 2024-05-17T17:53:51.447Z) OS version: Darwin x64 23.4.0 Modes: <details<summarySystem Info</summary </details<details<summaryExtensions (64)</summaryExtensionVersion ------ html-snippets0.2.1 tsl-problem-matcher0.6.2 emojisense0.10.0 github-markdown-preview0.3.0 markdown-checkbox0.4.0 markdown-emoji0.3.0 markdown-footnotes0.1.1 markdown-mermaid1.23.0 markdown-preview-github-styles2.0.4 markdown-yaml-preamble0.1.0 mermaid-markdown-syntax-highlighting1.6.3 vscode-tailwindcss0.11.68 esbuild-problem-matchers0.0.3 vscode-eslint2.4.4 kusto0.4.4 EditorConfig0.16.4 prettier-vscode10.4.0 auto-close-tag0.5.15 auto-rename-tag0.1.10 codespaces1.16.22 copilot1.194.885 copilot-chat0.16. copilot-nes0.0.1 remotehub0.63. vscode-github-actions0.26.2 vscode-pull-request-github0.89. headwind1.7.0 debug-tracker-vscode0.0.15 peripheral-viewer1.4.6 vscode-azureresourcegroups0.9.0 vscode-docker1.29.1 vscode-dotnet-runtime2.0.5 debugpy2024.7. isort2023.13. python2024.7. vscode-pylance2024.5.101 jupyter2024.4. jupyter-keymap1.1.2 jupyter-renderers1.0.17 remote-containers0.366.0 remote-ssh0.112. remote-ssh-edit0.86.0 remote-wsl0.88.2 anycode0.0.73 azure-repos0.39. cpptools1.20.5 live-server0.4.13 remote-repositories0.40.0 remote-server1.6. vscode-github-issue-notebooks0.0.130 vscode-markdown-notebook0.0.26 vscode-selfhost-test-provider0.3.25 vscode-speech0.9. vscode-typescript-next5.5. prisma5.14.0 vscode-xml0.26.1 vscode-yaml1.14.0 kuskus-kusto-syntax-highlighting2.0.2 svg-preview2.8.3 pdf1.2.2 vscode-mdx1.8.5 volar2.0.10 gitblame10.11.0 markdown-all-in-one3.6.2 (3 theme extensions excluded) </details<details<summaryA/B Experiments</summary</details<!-- generated by issue reporter --\nare you certain its that command? It is only enabled when a voice chat session is active. When you set the log level to trace, please check what is printed when you trigger in the chat widget:\nHappened again on a different machines. The main step of the STR seems to be using voice input (I mostly demo the hold down key usage) Both times I did troubleshooting and it pointed to the stop listening command.\nI need some steps to reproduce that you lead you to the issue and the precise location in the UI where this happens (globally anywhere or when you are in a chat input field?)"}
{"_id":"q-en-vscode-4cd351ef524827adf54c037f56382ff26166e337d50ced4f34d516f3db6bc62d","text":"a chat request that yields no code blocks the response with the keyboard in screen reader mode left and right arrows are in toolbar !"}
{"_id":"q-en-vscode-8f15b41c612e0b891d167d6872caf0347e51603463d2abc4d4bb5f9f30414b08","text":"Testing I select a set of cells with Shift+Click, and I hit Ctrl+F to open the find widget. The cells are then highlighted as such: <img width=\"981\" alt=\"image\" src=\"\"I scroll down a full page length, and then scroll back up. The cells are now highlighted as such, with no click or type input on my part: <img width=\"980\" alt=\"image\" src=\"\"It appears that the Code cells lost their yellow highlight."}
{"_id":"q-en-vscode-152706d9665dca6843fc9b42ea1119e41090ed2ae167150b2bc1f98cfb2e7154","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.90.0-insider Commit: Date: 2024-05-31T01:05:41.342Z Browser: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Code-Insiders/1.90.0-insider Chrome/122.0.6261.156 Electron/29.4.0 Safari/537.36 Steps to Reproduce: There is no test-log-channel in the output panel.\nThanks for reporting this I have the exact same issue. In the meantime I have a workaround: I was surprised this does not throw, usually if you create something that was not previously disposed, it will throw an error like \"command was already registered\" or \"notebook controller already exists\". I am currently commenting out my line in my extension. It's not ideal though.\nPR fixing it"}
{"_id":"q-en-vscode-9ec27be45c6eafb19b0393c40ccb61fdd34d2c4153338a50fcab734c694be597","text":"Have the following code Move line request down via Notice how the lines that have been passed change. Their indentation gets modified. The move line down command should not mutate lines it passes.\nSee and for similar/duplicates.\nshould be fixed by\nThanks for the fix!"}
{"_id":"q-en-vscode-59f7159d969af40d87a736a2ed2afd8723eb3be98979441122c98b4ec426f327","text":"Type: <bFeature Request</bCC setting is ignored when accessibility mode is on. While screen reader's TTS can auto-announce Chat response, some users may want to use this new synthesizer. so, we may want to turn off aria-live when this setting is on. Another consideration is that some low-vision users, who don't necessarily use SR, have turned on the a11y mode and they still want to benefit from this auto-synthesize setting. VS Code version: Code - Insiders 1.91.0-insider (, 2024-06-10T05:48:36.787Z) OS version: Windows_NT x64 10.0. Modes: <!-- generated by issue reporter --\nMakes sense cc\n<!-- -96bd-47a8-a0e8- --This feature request is now a candidate for our backlog. The community has 60 days to the issue. If it receives 20 upvotes we will move it to our backlog. If not, we will close it. To learn more about how we handle feature requests, please see our . Happy Coding!\nwould you, yourself, enable this? I worry about turning off regions.\nI don’t know whether I would turn this on permanently, but would like to try this out definitely to see if it gives me any new experience. I may be able to manage to listen to two different TTS simultaneously (one for Copilot; SR TTS for my own drive).\nThat makes sense. It would be great if you could try it out. You can do this now by using the action in the toolbar,\nYes, I have tried this, but this is a manual trigger which does not give me the same experience as the auto-read-aloud.\ndo you mute NVDA speech and use a braille device when you want to use voice input? or do you use a separate speaker/headphones? When testing this experience, the screen reader's speech is getting recorded by the mic and used as chat input for me unless I mute VoiceOver speech.\nThat happens when microphone is too sensitive. SR users need to mute their speech synthesizer or they need to use a different sound card or headset.\nI use a separate sound card for my screen reader.\nwhat are some verification steps for this issue?\nenable screen reader mode, adjust the setting to on, run a chat request, (make sure volume is up), hear the response\nI am still unsure what to test and what I should hear. I followed the steps and didn't hear the chat response. I only heard it when talking to chat in the first place.\nSorry - in order to hear the response, I forgot that you have to use voice input to the chat. So pls try that with screen reader mode enabled and that setting on."}
{"_id":"q-en-vscode-81a7b3d891e25531a500cbcb321379f831d522eb85b1cde200d4e3c40f9c5d76","text":"When I use the command line code --install-extension {id} Cannot find the installed extension Vscode - userdata: $HOME /. Vscode/extensions/extensions. The json. This error only appears in version 1.90.0, but does not exist if you go back to version 1.89.0.\nThis only happened the first time, and it worked the second time I executed the command line to install the plugin. So did everyone else around\nAre you still seeing this issue? If so can you please provide steps to reproduce?\nMy plug-in is a private plug-in that updates through the code command. When I released the updated version of the plug-in, I used the code command to update it, and there would be an error the first time, and it would be successful the second time, which was stable. !\nTLDR: The issue happens when a version that was installed before is installed again I have pretty much the same scenario. A private extension that uses to update itself when a new version is available. This is the error that's logged: I do have an entry for my extension in the file though: I did some testing on a bare project to understand the failure mechanism. What I found out is that if a particular version of the extension is installed before (for example during my testing before I release the extension) then upgrading/downgrading to that version will fail. Using following steps I can reproduce this issue reliably: 1- Create a bare project: 2- Package the extension: 3- Increment the version number, and package again: 4- Install the first package: 5- Update to second package: So far everything is in order. The problem begins when we try to install any of those packages again: 6- Downgrade to first package: At this point the neither 0.0.1 nor 0.0.2 is shown as installed. So installing any of them will be successful. this explains why you're always getting the error on the first, but the second round is successful. 7- Install the first package again: 8- Update to second package again: Tested this with Windows and WSL Ubuntu. Happens only on WSL Ubuntu in my case.\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\nLooks like this is working now in the latest VS Code Insiders. Tried with WSL: And using the latest stable (which doesn't have the fix yet) I can repro the errors that describes."}
{"_id":"q-en-vscode-b5e6bf8c1cfb1c5bc7da550b96891cadf6b6f4e48760a0934bd1fb2445f2effc","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.90.0 OS Version: Darwin arm64 23.5.0 (macOS Sonoma 14.5) Follow the section of the Then, in the the first step (after connecting to the container) is to run: Yarn install should complete successfully, and should run the app. We get the following error: <img width=\"803\" alt=\"yarn-install-failure\" src=\"\"According to the we're using . 20 is the current LTS, and 18 hasn't been in active support since October last year. We can use the node 20 version of the typescript dev container image: . For context: there was a compatibility issue with and node 20, as noted in the issue when we switched to the node 18 image: But it seems like this was resolved in : And we're on , so it should be fine\nI've linked a PR with a fix. As suggested the last time this issue popped up: a CI test for the dev container could be useful."}
{"_id":"q-en-vscode-6505a73ae3352243af3eabd2c49ba0b015e6440a54c219b94203d74cfcba36d9","text":"have a squiggle open inline chat for it the UI flickers when coming up\nI believe the is the same underlying cause as Drilling into this and it seems that setting follow up always triggers a layout with the \"old\" dimensions which isn't going fit. I believe that explains why the editor is looking so weirdly. It only gets very little height left IMO setting follow up should only trigger a content heigh change event so that the owner can issue a relayout but should not attempt to relayout with the existing dimensions. (maybe unless the number of follow up has changed in which case the content height didn't change...) <img width=\"1412\" alt=\"Screenshot 2024-06-18 at 16 47 47\" src=\"\"\ntried this is a notebook and it felt very jittery still (but in a regular file it was fine) should I create a new issue for this? Seems to be due to the inline fix dialog getting placed in one spot then immediately moved <img width=\"1273\" alt=\"Screenshot 2024-06-26 at 10 31 06 AM\" src=\"\"<img width=\"938\" alt=\"Screenshot 2024-06-26 at 10 31 01 AM\" src=\"\""}
{"_id":"q-en-vscode-3827f05d231ae0d4132032e8528b92eefcc8109004cc4f4882d761ae97dce73e","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.90.1 x64 OS Version: Ubuntu 24.04 (Updates 14 Jun 2024) I get the error on a fresh installation of Ubuntu 24.04 and VSCode 1.90.1 Everything works fine with VSCode version . Steps to Reproduce: latest version of Ubuntu 24.04. command. command.\nIn my case the issue occurred when doing , also on Ubuntu (20.04) and also on vscode 1.90.1 Here is a temp workaround to revert to vscode 1.89.1, if that's useful to anyone:\nHi Having same issue on remote VM Ubuntu 20.04 Install errors with Solved by using the flag"}
{"_id":"q-en-vscode-2bc803a9e1dcfb83a1f1bf8c937d63189b8ac2b5195642ca961ec81e46a38fcd","text":"Seeing this logged about 100 times at startup. Let's avoid stuff like this in the log level"}
{"_id":"q-en-vscode-a276a82a2f6e0838e07c0e847aa5c14e0fe1cfa6d52a84f878b96cafadfae5e8","text":"I think the editor used to preserve the vertical position of the line where the the cursor is in. It jumps now: ! Repro steps: Enable Code Lenses for TypeScript Navigate to a symbol in a file that has not been opened yet Notice that the editor reveals the symbol correctly, but without code lenses yet Notice that the editor jumps once code lenses appear Verification steps: Observe that the line the cursor is in does not jump anymore. If I don't misremember, there was a change made here recently?\nNo changes for code lens since long. Doesn't reproduce for me. Can yo consistently repro? Can you try without custom settings?\nThe current logic is to keep the first line in the viewport stable. A good improvement might be to keep the line where the cursor is stable (if it's in the viewport), but that is not what we have implemented. PR welcome.\nOh, somehow I didn't realizing the jumping then. Did you open this file before in the VS Code session? In the first frame of your screenrecording where shows, the code lens feature already reserved some (empty) space to prevent layout shifts (which I don't think the feature knows the first time the text model is opened): !\nBetween reloads the slots are persisted and optimistically shown for a couple of seconds so that the actual extension/provider can compute theirs (which is usually the same)\nI think there is something fishy going on with persisting the code lens cache ().\n<!-- UNABLETOLOCATECOMMITMESSAGE repo --Issue marked as unreleased but unable to locate closing commit in repo history. If this was closed in a separate repo you can add the label directly, or comment .\n<!-- UNABLETOLOCATECOMMITMESSAGE --Issue marked as unreleased but unable to locate closing commit in issue timeline. You can manually reference a commit by commenting , or directly add the label if you know this has already been releaased\nClosing in favor of (which proposes one possible solution to this). To be honest though, I think is what caused my feeling of this happening much more often."}
{"_id":"q-en-vscode-3feab95bbed1b52331aa7739ac310dfbe7ee76cf7c9823ea86d95f681fe06723","text":"Testing text buttons some code and ask \"What does this do?\" starts, and the \"ctrl+esc to stop\" button shows. . actually opens the Windows start menu and streaming continues. It seems like this is a default Windows shortcut.\nThis is the default keybinding (the one we use in all chat boxes) and it looks like we need something else on windows...\nThey have a whole keyboard button for that, why are they taking ctrl+esc too...\nGoing with alt+backspace. Would like ctrl+backspace, but you might start typing while the previous request is running, and still want to use ctrl+backspace to delete a word.\nWorks great!"}
{"_id":"q-en-vscode-a9b16cfcb4b309d03fb94c388012bb4fc2401699b48372f295700b1be2d2eae1","text":"Testing Hover in Chat View doesn't respect underline setting (expect to be underlined): !"}
{"_id":"q-en-vscode-c5edf98a4ee83e5ec09b1c3096f0a254f83bfec67c801e0430bcfff3b9887259","text":"Type: <bBug</bPreparation: create the following file: the terminals panel, create a new shell terminal it to the main editor area to make it a full-size terminal open an editor tab (any file will do, including Command-N) sure the panel is closed (press Command-J to close it if it is open) the above task (e.g. with Command-P, then type ) Expected: the panel opens and shows the running task Actual: the editor terminal is brought to the front, and the panel stays closed VS Code version: Code - Insiders 1.92.0-insider (, 2024-07-11T05:03:58.450Z) OS version: Darwin arm64 23.5.0 Modes: <details<summarySystem Info</summary </detailsExtensions disabled <!-- generated by issue reporter --\nIn this scenario, focus also fails.\nI can confirm that fixes the problem for me. Thanks!"}
{"_id":"q-en-vscode-21f174507210d3776f23bfc41ba46a73d8043f5211e150016ed14be12aa1870d","text":"Testing 1) select a cell 2) cmd+/ -cell commented 3) ESC 4) cmd+/ -nothing 5) arrow-up 6) Cmd+/ -nothing 7) Shift + arrow-up 8) Cmd+/ -two cells commented I don't really understand in which mode I'm in after step 3 and 5. I would expect Cmd+/ to work in those states."}
{"_id":"q-en-vscode-54a87bca7379c61bfc8cbde14450b9b744d3cfe71daba87073eceba63522d767","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: No <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.91.1 Commit: Date: 2024-07-09T22:06:49.809Z Electron: 29.4.0 ElectronBuildId: Chromium: 122.0.6261.156 20.9.0 V8: 12.2.281.27-electron.0 OS: Windows_NT x64 10.0. Steps to Reproduce: a TreeView with an TreeItem the TreeItem's iconPath to the extension. The TreeItem's icon will still be yellow same as in debugging. After inspecting developer tools, this is likely due to how the icon debug-stackframe's color rule is set to !important. The ThemeColor fails to overwrite it due to low specificity. This may not only happen with the icon \"debug-stackframe\" and is affecting customisations of treeitems. Why is some icons' color rules given the !important flag et al? There might not be any easy fix due to this flag's present. I hope the team could look into it, and if possible fix this selector specificity issue (if it won't break vscode, somehow).\nEdit: This not only occurs on special icons like debug-stackframe, but also on very ordinary ones like circle-filled. style attribute { color: rgb(204, 204, 204); } .monaco-workbench .codicon { } Adding !important to the color flag of .monaco-workbench .codicon will fix the issue though.. The problem is worse :C\nI think the best solution would be to get rid of that . it's coming from here. Can it be removed? The css you shared is correct. The rbg value is the one you set in your ThemeIcon. !\nYea, we should scope these rules better (and also move them to css now that we have variables there)\n<!-- UNABLETOLOCATECOMMITMESSAGE --Issue marked as unreleased but unable to locate closing commit in issue timeline. You can manually reference a commit by commenting , or directly add the label if you know this has already been releaased"}
{"_id":"q-en-vscode-dcbdd8e4c7391ac6ba53aea50af07a3b64d4f71112c04060bfbf1cb8f2244f7a","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.92.0 Commit: Date: 2024-07-31T23:26:45.634Z Electron: 30.1.2 ElectronBuildId: Chromium: 124.0.6367.243 20.14.0 V8: 12.4.254.20-electron.0 OS: Linux x64 6.9.3--generic Steps to Reproduce: theme to Light High Contrast a Javascript file final bracket is black, not red as in other themes\nI believe this may be the core cause: The value for hcLight here is just an empty string\nThanks for investigating, looks like this color got forgotten when creating the Light HC theme. I took the same color as used in the overview ruler"}
{"_id":"q-en-vscode-bc64935df97369fe03c8aa593d8488e4dd290e079ba40ed9a98f4cb8ebca67d2","text":"As reported by open notebook open nb from timeline close diff close notebook then I got Can repro from stable"}
{"_id":"q-en-vscode-314f9272723d20737bd1aa7c66611d7c7e5913239f8aebd695329b770f03140c","text":"<img width=\"1072\" alt=\"image\" src=\"\"<img width=\"838\" alt=\"image\" src=\"\""}
{"_id":"q-en-vscode-1d0b0a81275d02fbb5ebd894ac2e10ebc90c6806a11f54d939627e2a38c6776f","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes/No No. The issue occurs in a built-in extension so extensions need to be enabled to see the issue. <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.93.0-insider (Universal) Commit: Date: 2024-08-26T05:04:02.640Z Electron: 30.3.1 ElectronBuildId: Chromium: 124.0.6367.243 20.15.1 V8: 12.4.254.20-electron.0 OS: Darwin arm64 23.6.0 Steps to Reproduce: VS Code behind a proxy server and make a call to the function from the package included with VS Code similar to the one below. After the changes listed in , we began seeing failures from the Microsoft Authentication extension with the following log entries at startup. We rely on that extension to allow another extension we developed internally to authenticate our users. After debugging into the extension source, I found that the issue was originating here. After reading through the previously mentioned pull request, I figured that the call was hanging because the new function was not handling proxy configurations the same way (a common issue in corporate settings). I pulled down VS Code, built from source and tested a change that resolves the issue. I'd like to open a pull request with the fix if you agree that there is a bug in this flow. System Information\nwill be exploring fixing this type of issue transparently.\nThis will be in Monday's Insiders build. Can you please give that a try when it drops so we know for a fact we've fixed the issue for you?\nthat fixed the issue. Thanks so much for the quick response to our problem!"}
{"_id":"q-en-vscode-c17fb1cf41ea8ce271f7bc47f537ec62797e2a57ff8a379f18f7b88e0d8dd855","text":"Testing Repro steps: some changes to a cell, in my case I a cell with some content the changes some lines from the cell that was in step 1, and save. diff view launched from \"changes\" SCM view should diff between staged and working.\nNow the diff view for the changes in the 'Staged Changes' group is the same as the diff view for the unstaged changes\nI am seeing it as I would expect. This is diff launched from staged (where the cell was ) ! This is diff launched on unstaged: !\nWorks for me too, can you share the text file contents in the 3 cases Commited ipynb Staged ipynb & finally changed ipynb\nOriginal steps LGTM\nOriginal file: ! Modification 1: ! Modification 2: ! Observe that the diffs for the Staged and Working Changes regions below are identical !"}
{"_id":"q-en-vscode-fde1ef81a780ce338df6c7d9618a6bfa5625827ef93853cc3c55d41bf66d5a7a","text":"Testing I think we have another issue discussing a similar issue. This hover and the main item hover can conflict with each other: <img width=\"579\" alt=\"Screenshot 2024-08-27 at 10 37 49 AM\" src=\"\"Instead of having a special hover just for the badges I think the main hover accomplishes the goal just fine as the badge and its description shows up at the bottom: <img width=\"698\" alt=\"Screenshot 2024-08-27 at 10 39 05 AM\" src=\"\""}
{"_id":"q-en-vscode-416096d72ad920cb1af1f132058f092b9b0fac54b29765eec7c2d1b643d880b0","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --Version: 1.93.0-insider Commit: Date: 2024-08-29T05:04:03.909Z Electron: 30.4.0 ElectronBuildId: Chromium: 124.0.6367.243 20.15.1 V8: 12.4.254.20-electron.0 OS: Linux x64 6.9.3--generic Steps to Reproduce: a few editor groups in seemingly any split configuration one for text input or otherwise run to type: no result, as no editor has focus any more. This is especially frustrating when some of your other keyboard shortcuts rely on the editorFocus context, as they become unusable after using this feature.\nIn case this takes a while to fix, anyone looking for a workaround can rebind their preferred shortcut to automatically refocus the editor and then \"move\" the cursor to ensure if another window has now covered your cursor, the editor scrolls to reveal it again (also respecting scrolloff setting).\n<!-- AUTHORVERIFICATIONREQUEST --This bug has been fixed in the latest release of ! you can help us out by commenting if things are now working as expected. If things still don't seem right, please ensure you're on version of Insiders (today's or later - you can use in the command palette to check), and leave a comment letting us know what isn't working as expected. Happy Coding!\n/verified"}
{"_id":"q-en-vscode-e7d66e2b3df52dbcf2134b90422dc10faf520f6a5d89007ce877a9132be18e3a","text":"Enable the setting which enabled the native edit context adoption Copy some code which is more than 3 lines long Paste the code. The code that is pasted does not correspond to the code that is copied\nAfter some investigation, I have noticed the following curious points. In the following, please note that to fix the long-press-E chromium issue, we essentially always write the screen reader content which means we always change the document selection. When screen reader is toggled off: When we read the copied text from the the electron clipboard in order to paste, it is incorrect When we remove the logic that sets the document selection for the screen reader content, the pasted text is correct When screen reader is toggled on: Text that is read from electron clipboard is always correct In fact interestingly when we revert the changes we have made for the long-press E issue, then this copy-paste issue no longer occurs! Since the long-press E issue is not important, and copy pasting is much more important, I believe that we should revert that fix. From this I believe this could be somehow related to a chromium bug. cc\nI have made the following PR to revert the long press E fix : This fixed copy-pasting too."}
{"_id":"q-en-vscode-5d64e507f879970c582fdbfd1d65e9db4d050dedd1f082ce7190fa99c90404a1","text":"Enable the setting Open the editor Open the find widget Type several letters The editor is focused as soon as there is a match for the find result It is no longer possible to continue typing in the find widget without retriggering the focusing of the find widget"}
{"_id":"q-en-vscode-db82882439def63f05c7f27c6cc920d4497eff898b4526ab051decff485a02d8","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.93.1 OS Version: Windows 11 Pro Originally from To Reproduce Steps to reproduce the behavior: an empty dart project (can be any language, this is just the one I use of course) the code below breakpoints on both and print calls the breakpoint on the trigger only after the one at (screenshot) the code and make sure the breakpoint stops the program that triggers before Expected behavior For the triggered breakpoint to wait again on restart. Also note that there are two ways to restart a debug session too - a full restart handled by VS Code (which is what happens for Dart above), and a restart that is delegated to the debug adapter (which we use for Flutter to do Hot Restart, so it's faster than a cold restart by VS Code). In both cases, this state should be reset. Screenshots ! <details<summaryRelevant debugging log part </summary</details// CC:\nThanks for creating this issue! It looks like you may be using an old version of VS Code, the latest stable release is 1.93.1. Please try upgrading to the latest version and checking whether this issue remains. Happy Coding!\nInterestingly, I cannot reproduce this with JavaScript using an equivalent file: Unfortunately I don't think it uses DAP and I don't know if there's a way to get a log of what's happening there. It's not clear to me how it could be working there given the DAP traffic captured for Dart. Someone familiar with the implementation of Triggered Breakpoints or the JS debugger might be able to explain what's going on.\nI can on python:\nI am also running into this issue\nThis is probably because the JS debugger has each debugee as its own, child session, and when restarting JS you will retain the parent session (handled by the DA) but get a new child session that doesn't have the data set.\nI am debugging Go. I don't remember seeing it with other languages but I also only rarely run anything other than Go.\nyea, it's reproducable with Python too. Fixed in Monday's Insiders."}
{"_id":"q-en-vscode-3450cd62072d804257a5726f9802d0dddd042ba04c38c153b0586770c0e81078","text":"today's insiders open GHINB and run a cell see errors (many) on the console\nbroke\nWe should also add a custom eslint rule for this. It has been broken for a few weeks and it regresses the output focus context keys."}
{"_id":"q-en-vscode-e5b44f7654273b46bc3820b1d249187b7bf7c1c3e939a88ece4de9171e021628","text":"<!-- Do Not Delete This! bugreporttemplate --<!-- Please read our Rules of Conduct: --<!-- Read our guide about submitting issues: --<!-- Search existing issues to avoid creating duplicates. --<!-- Test using the latest Insiders build to see if your issue has already been fixed: --<!-- Instead of creating your report here, use 'Report Issue' from the 'Help' menu in VS Code to pre-fill useful information. --<!-- Launch with to check. --Does this issue occur when all extensions are disabled?: Yes <!-- If you answered No above, use 'Help: Start Extension Bisect' from Command Palette to try to identify the cause. --<!-- Issues caused by an extension need to be reported directly to the extension publisher. The 'Help Report Issue' dialog can assist with this. --- VS Code Version: 1.95.0-insider OS Version: Windows 11 Steps to Reproduce: : window to message and use <kbdShift</kbd+<kbdEnter</kbd<kbdShift</kbd+<kbdEnter</kbddoes not work in Source control message text area if option is enabled. It seems that it occurs not only in but also when I edit any file. <hrMaybe related to\nI see thank you for filing this issue. I will look into it.\nHi may I ask you what you were expecting to see? I turned the setting on and off and press when I have no commits to make and a notification pops up mentioning that I have no changes to commit. Thank you.\nHi, thank you for checking it out. I expect that a new line will be , the same as with just pressing <kbdEnter</kbd. By the way, the same issue occurs when I edit a file.\nHi thank you for clarifying the question, yes I am able to reproduce this behavior. This happens because we don't send a new line event when we receive the event. I will make a PR to fix this."}