{"_id":"q-en-nixpkgs-5c6c65eca04b10446947d8085f6fac9674a23872d4042d5bae7e40b2107ba2a7","text":"Describe the bug On NixOS 20.09 and unstable, is not in the search path, even though it is installed with the php package. To Reproduce Steps to reproduce the behavior: [example from 20.09 system] Expected behavior [example from 20.03 system] Additional context Debugging with reveals that on 20.09, php is searching in , then , which fails. On 20.03, it searches in , then , which succeeds. On 20.09 and unstable, php74 and php73 both fail. On 20.03, php73 (the default) works, but php74 fails, but I believe it fails for a different reason. This line is new since release-20.03. At first glance, it seems as though is resolving to null here, but I'm not sure what to replace it with: Notify maintainers Metadata Maintainer information:\nThanks for reporting this issue! Your guess was right on spot and it was a really simple fix it seems :slightlysmilingface: As you may have seen: Right after 20.03 branched off we started a big refactoring of the PHP ecosystem. So bugs and issues was to be expected :slightlysmilingface:"}
{"_id":"q-en-nixpkgs-381d4f3bd022fc45e1f3a963a51ff21141f861531b553dcd5cae11bfb14dd7a9","text":"I'm using the example of which renders correctly in firefox (the text in that section matches the image), but not in chromium (I see some placeholder boxes) my font configuration is exactly that given in I'm running nixos-unstable. I don't have flash enabled, or any other modifications to the default chromium package. From a bit of research, I've found that it might be the case that chromium restricts access to system fonts in some way, as a privacy aid - possibly that is to blame. Note: this occurs without any addons (running chromium as )"}
{"_id":"q-en-nixpkgs-4a6f04bb476939757eab596e4c318e0af047adf3e8544928d4ad651325023c32","text":"Describe the bug When cinnamon is started either via gdm or in a . Sometimes the desktop icons do load Always a window spawns telling me that I'm in fallback mode no taskbar, no window decorations (hence I think no compositor & window manager) To Reproduce Here I'm kind of unsure: I deleted all nemo and cinnamon folder in $HOME, .config and .local/share. Still this occurs on all of my machines. IIRC this first happen when accidentally starting a \"virtual keyboard\" app cinnamon \"Virtual keyboard\" cinnamon components crash instantly, the and the behavior is reproducible on further attempts of starting cinnamon Expected behavior Cinnamon starts and is usable Additional context Booting into an old generation did not fix this, however, I'm quite unsure what could be the cause of it. This is some log which I could create by running cinnamon from a . I'm not sure whether this is related to the issue or not. relevant : Output captured (the log is over 2000 lines long, this is from the very end and might be the issue): Notify maintainers Metadata system: host os: multi-user?: sandbox: version: channels(root): channels(wucke13): nixpkgs: Maintainer information:\nI will say that it does start in lightdm. IIRC upstream uses slick greeter and lightdm, so maybe it's just broken for that dm.\nit is broken when started on as well, and it used to work in gdm just a day a go as well. Thinking more of the problem, what else can break the startup of cinnamon if all of these are empty/removed: + + + ? There must be something in my $HOME which alters the startup procedure of cinnamon to break (otherwise it wouldn't work for you?!) but what can it be? Edit: I now gathered a more meaningful error: Edit 2: Found the remaining part thing: dconf. After renaming , cinnamon starts just fine. Holy moly. Running \"Virtual keyboard\" from the cinnamon menu breaks it immediately, though.\nWhich package provides gio for ? We'd need to add that to buildInputs then the error should be fixed\nThat is . Be sure to on the source code to always run a to see what namespaces are imported. From there you track where those typelibs come from to determine the dependencies that belong in\nAny updates on this?\nIt would need to be tested if the issue still persists.\nThe bug is still persistent. system: host os: multi-user?: sandbox: version: channels(root): channels(wucke13): nixpkgs:\nThe bug is still present in 21.05: system: host os: multi-user?: sandbox: version: channels(root): channels(wucke13): nixpkgs:\nThe main problem I have is that I have no idea where to even start debugging the segfault, as it seems to be part of native code\nAny chances upstream has some idea? Not sure how to get in touch with them without making it seem like \"We have a problem and you shall fix it\".\nI ran and I found that it is missing , just install it and the bug will disappear. I think this package is a must or we have a security-issue here\nI can confirm installing caribou fixes the issue. I haven't really looked into it yet, the related line is likely at And I managed to SIGSEGV gjs-console (cjs-console) in a shell with: I search for the output , and found this So I do and the SIGSEGV is gone. So I guess caribou is needed for this case."}
{"_id":"q-en-nixpkgs-749c78a9d855c6c677e5baf80960970da2301fce8486aa55b91ac39497ecefd2","text":", CVSSv3=8.8 (nixos-20.03, nixos-20.09, nixos-unstable) Scanned versions: nixos-20.03: ; nixos-20.09: ; nixos-unstable: . Cc"}
{"_id":"q-en-nixpkgs-dfe78460f143649386dc0f21faa0d370d531760a0fd0c262501f001c65398e2f","text":"With NixOS 15.09 I get this error running LilyPond: I can work around this (clumsily) by using an old version of in LilyPond's . This differs by not including . This seems to be related to Changing apparently can break some users of as well as those of .\ncc\nHmm, perhaps we should add or some similar option to handle these cases easily?\nIt's better to set in most cases, which I just did in and ."}
{"_id":"q-en-nixpkgs-171246af152b99fef20bf6d1eb66375d9cbbce0c3d165d607e1e49a6712dc218","text":"Describe the bug defines the CMake path. The environment variable is included in this, and addons like SDL2mixer append themselves to this path. However, that same file also defines an target, which does not include in its include directories. See for an example of an issue caused by this. To Reproduce Steps to reproduce the behavior: to build an application relying on SDL2::SDL2 adding SDL2mixer to the include path, such as Expected behavior 's should include the environment variable, just like . Screenshots N/A Additional context N/A Notify maintainers (SDL2), (SDL2mixer) Metadata Maintainer information:"}
{"_id":"q-en-nixpkgs-f7843ba261ada76f891f97521061321d01474bb552975cca6be10311f1bda28e","text":"Experienced on gnome3.38 which is currently on master. alacritty 0.5.0 isn't functional at all. 0.6.0rc-1 is. Should we update in order to make our packages set functional? Or should we wait for a real, no rc release upstream?\nyou should update yourself in the maintainers list - I had to go to to see who is .\nIs this a wayland issue on gnome?\nYes.\nSound fine to me.\nThanks for pointing this out. As much as I want to see my first username long gone, the change was not worth it :)"}
{"_id":"q-en-nixpkgs-7f9d7f5b6e9df6d755b618a2472254c1b3c6ebe8c3dc80511b27795ca067ca8c","text":"Updated using I though most of the packages have moved on to openblas, what is pulling in this package? How would I find out?\nA few packages still reference it. should show what derivations use it directly, from those you have instantiated."}
{"_id":"q-en-nixpkgs-78bfcf7d442e0fe3591828407bf6d4792aebae2aab1291d74526be84b96c68d5","text":"Describe the bug option is not working To Reproduce Steps to reproduce the behavior: service and its option in . Add as well for testing. In my case I used something similar to the password read bit is set, for the test generation: (eventually) inspect logs with in a separate terminal a mail The command might return And you might find an error similar to this in your logs Tested on 19.09 and 20.09, on different machines with three different SMTP hosts, various passwords, with and without special characters. As explicitly warn in the doc Path to a file that contains the password used for SMTP auth. The file should not contain a trailing newline, if the password does not contain one. This file should be readable by the users that need to execute ssmtp. I ensured that the password do not contains a leading space or carriage return. Expected behavior should work just as well as its former obsolete attribute. Additional context eventually check again my password in the store: which is yet followed by the character . It's obviously a control character and I cannot find a definition of its meaning. The very same command (original pass file) won't return the . following configuration works, with of course the exact same password as provided in the password file. Which is obviously not the expected setup. Notify maintainers Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:\nMake sure that it doesn't have a trailing either. Editors like vim tend to insert one by default.\nI can confirm.\nIs this still a problem?\nI haven't been able to validate the response using or alternative editors.\nUse to avoid the trailing newline. To check whether there's one, use ; with this command, newlines are displayed as ."}
{"_id":"q-en-nixpkgs-7e2dd9d1869c0b37e1a421e9345bc8dc92126b6e9ff1900ea93c779135deb20c","text":"Seems like nmap changed the licence with the 7.90 release from GPL-2 to NPSL (Nmap public source licence). NPSL is \"based\" on GPL but adds some nasty things such as: Which conflicts with the freedom 0: Discussions about this: Debian: Gentoo: Upstream: The actual license file: They also seem to bundle a GPL library, and their license doesn't seem to be GPL compatible. So we should probably add this new license at least. NPSL looks non-free to me. Non-free is a good default until FSF or anyone has looked at it. We can always follow suit with what Debian does as well.\nAnother option would be to revert to 7.80 for now. The new license seems to be a legal minefield even if you're ok with closed source.\nYeah while looking at the issue in the train I thought of the same as well, will do that now.\nI just read the release annoncement for 7.90: Just to see if the license change was mentioned there. Which it is: Do I read this right they weren't really GPLv2 to begin with and just clarified that? That's like a major thing...\nFirst of all I just realized that I merged the original PR, so I really have to apologize for missing that license thing back then! cc in case you currently have time and energy for that, would you mind taking a look at the issue? I'm pretty sure that you have way more knowledge and insights about that kind of licensing issue than we have :). Is there any kind of statement from OSI or FSF yet? Also, I'm wondering if other distros that are even more focused on free software than we are (such as Guix) have e.g. started a new fork that we may want to use as well.\nMore details from the changelog: I must say that I'm not sure if it's worth to roll the package back since it actually haven't seemed to change really. It just was weird to begin with and just now got a name on the weirdness.\nIt's not just a new name, apparently they also new requirements in 7.90:\nI don't think we should revert it either. Add the new license and mark it unfree until this is clarified.\nOof.... that's actually another reason for reverting.\nI'm in favor of marking as nonfree until the FSF or some other authoritative organisation decides otherwise.\nEither revert, or mark as non-free. Or, revert and add the new, non-free, version next to it.\nThe nmap update has already been reverted. Let's wait what upstream and the legal teams of debian/fedora say. See also the upstream issue\nFedora's stance\nThat's the old license (until version 7.80). This issue is about the new license (in version 7.90).\nFedora the new licence to the bad list and marked it as \"not acceptable for use in Fedora\" due to being incompatible to the open source definition\nSo, now there's an update here that we may want to act on: To quote the post: This means that 7.91 is now available under the same license as 7.80 is. The license that 7.80 uses is still a \"bad license\" because it's still a custom license. But it's a license that we and other distributions have considered good because it was assumed to be GPLv2. So I'm thinking that we could upgrade to 7.91 by now, because it's still not good, but it's not worse than before except we are aware of it now. Then we just hope that they do some good re-licensing in a future release.\nMerged which removed nmap-unfree. This issue is done for now. Thanks everyone."}
{"_id":"q-en-nixpkgs-48d6df343288765639073b6c2560bcab44bba033afd46fe03f5e7adeb01b63b3","text":"This issue tracks the conversion of from Docbook to CommonMark per . If you intend to work on this conversion, please comment below! here that you intend to work on it. (Please also post if you stop working on it.) a new branch of nixpkgs master. : change to . to . the contents from Docbook to . Follow . Use existing sections for inspiration. from the folder to build the manual. the manual on your computer at after running from folder: it looks good, submit a PR and reference this issue.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nI opened a PR here: Thanks in advance for reviewing :-)"}
{"_id":"q-en-nixpkgs-fc7af294ae92797743c832a4bd070673650fa5e986c6f83d3c9ed4e24f01707e","text":"This issue tracks the conversion of from Docbook to CommonMark per . If you intend to work on this conversion, please comment below! here that you intend to work on it. (Please also post if you stop working on it.) a new branch of nixpkgs master. : change to . to . the contents from Docbook to . Follow . Use existing sections for inspiration. from the folder to build the manual. the manual on your computer at after running from folder: it looks good, submit a PR and reference this issue."}
{"_id":"q-en-nixpkgs-662032e7f643781cf5072e7cf48420cfdf46bb2a962ca391b10765c3ca7c5fe0","text":"Describe the bug Milkytracker is missing a menu entry to launch the application Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --"}
{"_id":"q-en-nixpkgs-1242ecf2604c10a0d494fd9a1b007dad04c2f1908293b6b1c0a7b2d5234c2839","text":"We should replace the discontinued source with the more active fork . wrecker/stern hasn't received any commits since 2019. The maintainer of stern doesn't work at wercker anymore, and it seems no one else there is maintaining it (wercker/stern). The community has since moved to stern/stern with improvements to the original source. Project description Multi pod and container log tailing for Kubernetes Metadata homepage URL: source URL: license: Apache-2.0 platforms: unix"}
{"_id":"q-en-nixpkgs-b86c76c18c03ea4ae793e0c45255d4d099325d39338e62afde31b604415b0ec0","text":"Describe the bug spacefm segfaults on start with the following output: invoke spacefm on the command line Expected behavior an up and running GUI program Screenshots see error message above Additional context I noticed that there is no It makes no difference whether spacefm is installed via nix-env or Notify maintainers Metadata** $ nix-shell -p nix-info --run \"nix-info -m\" system: host os: multi-user?: sandbox: version: channels(root): channels(oo): nixpkgs:\nI bisected this crash down to Perhaps the package needs to be updated to support the newer version of gtk3?\nHi. I cannot reproduce the crash with (current master). How about you?\nYep, that same binary segfaults for me. I'm on\nstill crashing with the same error message. Then out of curiosity I ran . The backtrace of the main thread has: 0x0000000000000000 in ?? () 0x00007fa6931c3ceb in ghashtablelookup () from /nix/store/hnk0s558k2h4v8abjzagagflvw1l0cb2-glib-2.66.4/lib/libglib- 0x00007fa69375c50c in gdkx11windowlookupfordisplay () from /nix/store/8pidhwid3ww58wlq705irf48c01dr0hv-gtk+3-3.24.24/lib/libgdk-3.so.0 0x0000000000499915 in getdesktopindex (win=win at main-window.c:4179 0x000000000049bfd1 in fmmainwindowgetoncurrentdesktop () at main-window.c:4225 0x0000000000496343 in handleparsedcommandlineargs () at main.c:1143 0x0000000000421808 in main (argc=<optimized out, argv=<optimized out) at main.c:1526 and now I am wondering about as I am running Wayland and sway. I don't know if it's relevant though.\nDoes it help if you run spacefm with the environment variable set? BTW, I am using X11.\nSeems this a known bug:\nGDKBACKEND=wayland does not help. Found an even older bug report from 2015: Question: this is an upstream known bug. Should I close this issue? (I don't know what the protocol is in this case)\nI am not sure. Maybe we can at least improve something here? Change the description to call out that this does not work with wayland for example? Patch it that it fails gracefully with an explanatory message when not run under X11? Then, it seems the project is also not really maintained anymore. Maybe it's time to let it go?\nAnother idea: a wayland display driver (is this the right wording?) should fall back to x11 support via xwayland if an application cannot run under wayland natively. SpaceFM is a x11 only application, could you try to use and see if it makes a difference?\nit actually runs with GDK_BACKEND=x11! Thanks for the suggestion.\nI have created a PR to make x11 the only backend that is used."}
{"_id":"q-en-nixpkgs-3c1440d3e484b431527da210191f44634f916d244b97ef3b7d68ab17d96d7005","text":"Describe the bug (and, therefore, pretty much any package) fails to build on armv6l. To Reproduce Steps to reproduce the behavior: an armv6l machine. (I actually used an aarch64 Raspberry Pi 4B.) Additional context The segfault comes from , which is invoked by the Linux kernel build system. Here's a minimal C program that segfaults, when compiled with the from : A git bisect reveals that the faulty commit is (). Notify maintainers maintainer of binutils (since that's what seems to be the culprit here) author of Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(gaelan): channels(root): nixpkgs: Maintainer information:\nI guess that this patch should only be enabled on armv7l then. Truthfully that PR was just changes proposed/written by others and someone else is probably more qualified wrt binutils to say what the correct thing to do would be.\nI don't think it's an armv6l vs. armv7l thing: fails in the same way. There's a possibility that the issue is that I'm trying to build on a Pi 4, which is actually a 64-bit machine, with a 64-bit kernel. To test this, I copied the linux-headers closure, and all of its deps, to a Pi 0 W (which is actually armv6l), and got the same bug. It is still possible that the issue arises due to the Pi 4 miscompiling one of the dependences. I'm doing a full build on the Pi 0 to rule this out, but of course that could take a while.\nAlright, also fails when building from scratch on a Pi 0. At this point, I'm pretty sure this patch just breaks building anything natively on aarch32.\nThanks for the testing I'm not sure what's more important, cross compiling to arm32 or native compiling! Hopefully can shed some light here.\nWhat a mess! I guess the patch fixes some relocations breaks others. If we could make minimal programs for both that would be nice, but that's a lot of work.\nAh I see already minimized in , and we can cross-reference Jessica's patch on this issue too. So maybe once we minimize this issue we'll have enough combinations to try that it shouldn't be so bad.\nI have this problem on armv7l as well, on the novena board. 5.9.8 headers required for building nix fail.\nI'm running into the exact same issue using arm emulation via on my x64 machine.\nIIRC GHC was the only package affected by the bug in my NixOS closure when I had it running on an armv7 machine back in 2019 and it was also happening on native builds; had nothing to do with musl or cross. Couldn't we just make a special binutils with the patch for GHC and the (at worst) handful of other programs which need it?\nHello everyone, I'm trying to make use of the recent changes. Here what I've done: I'm using the nix installer from . I guess this should actually not be needed any more as we now have for this merged . Using my Pi4 with 32bit kernel: pi $ uname -a Linux raspberrypi 5.10.17-v7l+ SMP Fri Apr 30 13:20:47 BST 2021 armv7l GNU/Linux on the unstable channel (should not be relevant) and on current running: now passes. W-h-o-o . However, compiling our friendly greeting: results in the following segfault: Seems like this PR has not been all to fix the broken target. Anyone aware of the issue here and how to continue? Happy to update the official Wiki as soon as I've things working ... also including the hydra tarball. Thanks!\ntry cherry-picking commit from\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nThanks . That definitely helped but failing at stage4 now: Hmm, looks like the OS killed it? - maybe OOM? I guess there is a way to reduce memory load by reducing the number of parallel jobs? Do we know the minimal config right now?\nI had no issues building the entire system (with a few tweaks) on armv7l virtual machine with 4 cores and 4G of RAM. You can try building with . Do you use NixOS? If not, is nix installed in multi-user mode?\nYes, OOM often produces exactly this \"killed\" log line.\nAlso make sure you have zram swap (perhaps even with zstd) if you're running into OOM issues. It's free additional RAM space.\nHitting the exact same bug when cross-compiling to armv7l now, in packages that have PIC disabled (zfs)"}
{"_id":"q-en-nixpkgs-5684bbb484476e66b822b1769b32c6aaa3d2d8ddc0be88a73612b9f412bfc8f8","text":"Describe the bug There may be something flaky about the python3{789} build. It's been going on and off on hydra. I can't reliably reproduce the failures seen on hydra locally when building from the exact same commits that failed on hydra, but I have observed the same failure at least once on my machine. ! Expected behavior dask should build cleanly. Notify maintainers Metadata"}
{"_id":"q-en-nixpkgs-a55ecc7875035a29f11dc5bbd94e35bc02992c57d750a8bddc3efc44cdff538a","text":"Describe the bug Just lost on two different NixOS machine my Signal Desktop history because of this faulty update... I don't know, probably doesn't make sense for people to downgrade who are already upgraded but for others it dataloss To Reproduce Steps to reproduce the behavior: Signal Desktop Signal Desktop you to delete db, if you interact in a wrong way it will delete your data Notify maintainers ( and\nSorry to hear that. I recently lost my history too... It's really frustrating! Yeah, unfortunately I don't think there's anything we can do about it (as downstream package maintainers). AFAIK the source of the bug is still unknown, it doesn't seem to affect all users, and all releases have a limited lifetime (\"expire\" 90 days after the release, see my comment in ). Thanks for saying that! This really bothered me too... I just didn't say anything because I was too angry to keep it nice enough. (I raged a bit here but after I realized that one cannot even import old backups I had to take a break.)\nI just got hit by a similar issue by moving from Archlinux (encrypted database) to NixOS (non encrypted database). After many things, the easiest fix to ensure proper functionalities IMHO is installing: This will disable tracker3 and prevent SQLite to be loaded in the process which will replace SQLCipher and cause the SQLite3 API to be the original one (i.e. non-encrypted) and cause many issues down the road I guess. That might not fix everything though :upsidedownface: Opened a PR:\nFYI, in case anyone has feedback/suggestions: We might merge soon and it should cause another \"Database startup error: Error: SQLITE_NOTADB: file is not a database\" error (which is really unfortunate but not easily avoidable).\nHeads-up: I've just merged to hopefully resolve this once and for all and the Python wrapper script should take care of the database re-encryption when launching Signal-Desktop as usual but the user experience isn't that great and you probably want to backup first in case the script fails unexpectedly.\nThank you, I just upgraded Signal on my computer and the re-encryption did work as intended"}
{"_id":"q-en-nixpkgs-0e03f13613f2ca91a7ebf46b67acef566fb1a05485cb223203b40a926f948727","text":"Carry over discussion from IRC #nixos-dev and . Inspired by the doc change in recommending top-level usage of over , we want to reduce confusion for newcomers and unnecessary complexity by making the canonical way to access the Nix library. We may still want to keep to not break external code, and place a or similar at its definition site. There are over files where occurs at least once. Beyond splitting this change across multiple PRs, in what ways can the change be automated? Pinging\nIn the short term, we could add this check in ofbrorg and , and update any examples in the documentation to reduce new usage of .\nDon't forget things like and .\nAh. So 45K occurrences across 11.8K files.\ntoplevel is done\nthe xorg dep generator also doesn’t put it in its output anymore\ntrewide rewrite in\ndoc changes in\nGraph of occurrences over time:\nNice! What scripts did you use to generate it?\nThere's a source code link now :)\nFound another large source of s. Fixing cabal2nix should remove around 24k occurrences in .\nStatus report\nJust found out that counts stuff like as well, so\nIs it ok to rewrite to ?\nThat's the lib output of , so no.\nan easy way to find out is to evaluate both in\nSubfolders of with <10 changes needed in PR\nAround + replacements out of 30k+ are concentrated in - should be trivial to regenerate the file with the new desired contents.\nI'm working on it but it's a bit more complicated than I anticipated.\nWhat problems are you running into?\nPart of code generation of cabal2nix is in an external dependency, so I had to open two PRs. The dependency is itself on Hackage, so we may have to wait until the change propagates to Stackage LTS and then Nixpkgs, unless peti will use an override or something: I'm not familiar with the cabal2nix workflow.\nI think the haskell-updates are mostly done from a local checkout of cabal2nix, so it depends mostly on what version peter is using. You can also probably just ask him, he is generally available for such things on his .\ncould you help here?\nThe Haskell infrastructure has switched to the new style with commit . It will land in this Friday. I have also updated the generator code and released new versions to Hackage, so I guess the Haskell infrastructure can be considered done.\nThank you\nMore accurate regex: (the one above captures stuff like\nWhen is merged, this will be the result Three Nix files remain that use directly.\nThe last usages of it can be found by commenting out where it is inherited in .\nI just got this warning while doing a , but it doesn't give me any info as to which derivations are at fault. Is there any way we could improve the message to identify culprits?\nWith , you should get a hard error which will allow you to use . Or you can update Nixpkgs, since it has been removed in"}
{"_id":"q-en-nixpkgs-7ef3abb1390a145dc1caa7687c3c11ce882c671b9bb6e9b3d55b8a3204d7c959","text":", CVSSv3=7.5 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: . Cc"}
{"_id":"q-en-nixpkgs-aca4f5006797b289d3341848f1a9b6b6b17e39e8769b26d4fd9ec8a30f72b3fb","text":", CVSSv3=7.5 (nixos-20.09, nixos-unstable) CVSSv3=7.5 (nixos-20.09, nixos-unstable) CVSSv3=5.3 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: ."}
{"_id":"q-en-nixpkgs-907dc909eace117134f66dae30b8eb2a49994a646b5ce7d6eabaa5cfe1589f19","text":", CVSSv3=5.3 (nixos-20.09) CVSSv3=5.3 (nixos-20.09) Scanned versions: nixos-20.09: . Cc Cc Cc Cc"}
{"_id":"q-en-nixpkgs-0a2ba3518db2e1615b7c5e685f2b0a42baa15f77a73ce966bb3ea4b9b3eb8526","text":", CVSSv3=6.1 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: . Cc"}
{"_id":"q-en-nixpkgs-6760c3f08d89443dea7bd3a57bfbdcda106553cea19e895fc45fb7b5a7f5f4e9","text":"Describe the bug I believe the source for the package might no longer be valid: Accessing it in the browser shows a 404, and the home page -- -- appears to just be a portal for serving sponsored links. The Wayback Machine shows that the has not been archived since February 2020, and also get the same 404. There was a in Septermber 2020, but I'm unsure if this was the cause. Thanks! To Reproduce Steps to reproduce the behavior: . Expected behavior Installation should not hang, and the font should be installed. Screenshots Additional context appears to be an alternative to Helvetica, so I'm unsure how important this package is. But I figured this error should be documented. As far as I can tell, is the only package using , so if this domain is indeed no longer usable then nothing else should be impacted. Notify maintainers Metadata system: host os: multi-user?: sandbox: version: nixpkgs: Maintainer information:\nbut the link is also dead: so i would ask there where the font is available and if it is freely licensed now. it was also licensed by adobe at some point it seem to ba available from several shady websites, like but that's clearly illegal and shouldn't be packaged in nixpkgs freely licensed alternatives are: (Nimbus Sans)\nI marked this as stale due to inactivity. &rarr;\nCould the helvetica-neue-lt-std package be removed or at least generate a better error than a 404 when it tries to download the font? I ran into this error while doing a nixos-install so it took a while to find the specific package that was causing my build to fail.\nThank you"}
{"_id":"q-en-nixpkgs-adc87b492e63989accb90f7d6e042f9b0537e53be7a268c0ff8995189cd0c46b","text":"Describe the bug ! was something changed about the exported package list? Expected behavior same number of packages or even more Additional context we used to have the most packages, but now AUR is on top again\nMaybe is part of this.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nMight also be a redefinition of a \"project\" on the repology side, i.e. duplicates that weren't detected before. Projects != Packages. Do we know the revisions of before and after? We could check whether anything went missing from in the tarball because that's the source of Repology's data.\nWell it must've happened on January 16th so we could use one commit from the 15th and 17th and see if there are suspicious changes in the that json file in between. Using we can pick a merge commit from the 15th e.g. and one from the 17th e.g.\nWe need to know the state of the unstable channel on those dates, it's days or sometimes even weeks behind master. Don't really have time for this right now but is probably what we need.\nWell it must've happened on January 16th It did appear in Repology on 16. January, but our channel take some time after a commit was done. So i would check like the 9. Jan.. There was a successful channel evaluation on 2021-01-07. Channel evaluation: () packages file: () () () Latest unstable channel evaluation: () So the number of packages decreased by 8,990 between 2021-01-07 () and 2021-01-16 (). There was this change of emacsPackages to on 2021-01-13 () And it's indeed the emacs packages that are missing on 2021-01-16! ! Number of : Number of packages that are in but not in : Full list: For example . It does not exist in packages-2021-02- It does exist today in the search: But not on Repology: We should really be consistent here. The packages that nix can install from nixpkgs channel, and Repology should be identical!\n, and so on were moved to their respective interpreters as & in Partly this is to fix the asymmetry of not having package sets available for the packages. The old names now only exists as aliases which probably explains why they have disappeared from Repology. I think this opens up a question regarding if we should remove from . We've tried this before and it makes tools like a lot slower since these package sets are pretty massive. Maybe we should only expose for the default Emacs package but not for the rest? That way we also don't count every package in the sets N times (where N is the number of distinct Emacs packages).\nI think such tools should show packages only once, even when they exist for different interpreters like emacs26 and emacs27. But it should be obvious that you can use both. The raw package list should still contain all i guess, so tools like repology get all nixpkgs has to offer. But that is a big topic and it's about UX. For example in the search. Now it looks like this: ! But this would be less cluttered and more user friendly: ! (I opened an issue for that ) Same in .\nThe bug is probably here: did update the set names here too but I don't think sub-attributes work in that construct.\nNice catch! I've attempted to conjure up a fix in but it needs more testing before I'm comfortable merging.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nStatus: rPackages is fixed See for example \"tidytidbits\" on or . In the package set list is also emacs27, but i don't find any packages. When i search for random packages that existed before, like \"pony-mode\" i don't find it now in search or repology. The name does exist in which is not found in github search, probably because the file is too big. On repology you find only 1 emacs package in nixpkgs! So the package set emacs27 is still missing from ."}
{"_id":"q-en-nixpkgs-9f888b609632e9c7f54edd4eeb713273d014078e06da88c64ea3f14f426bc437","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on github to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: radvd <!-- The current version can be found easily with the same process than above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 2.18 desired version: 2.19 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-b30f42e8043eaa3a28d9e7a4dccce080dbcb33da5585485f3984a35d6e633021","text":"Describe the bug dmd keeps references to boostrap compiler. To Reproduce Steps to reproduce the behavior: Expected behavior dmd has no references to boostrap compiler Notify maintainers Metadata system: host os: multi-user?: sandbox: version: channels(root): channels(symphorien): nixpkgs: Maintainer information: Note: to prevent future regressions, it would be nice to add to the derivation ()\nI marked this as stale due to inactivity. &rarr;\nStill an issue\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-36c774bdfc8de263bcf40d45bfe02ed28a226c531b4c65ae76b64de9f372dd73","text":"Describe the bug The build of the RPI4 image has been broken for a few weeks now. See: The module closure fails to build with: CC Maintainer information:\nCaused by See\nFor the time being, there is no quick way to work around the issue. From the list of available modules, most are needed. We'd need to be able to filter out a module from the list, which AFAIK for lists is impossible with NixOS modules. So uh... annoyingly it'll have to stay this way until some resolution of the overall issue :/... An alternative, but quite intrusive, would be to split the aarch64 profile into an \"aarch64 base profile\", from which the normal aarch64 and the raspberry pi 4 profiles inherit. Though I don't really like the approach since the raspberry pi 4 profile is, overall, a temporary measure for a somewhat popular and hard-to-work-with platform. It might be more worthwhile , in the long run, to take a final look at 5.10/5.11 mainline and if it is enough to allow a user to bootstrap themselves finally.\ncouldn't we allow missing kernel modules when generating images at least?\nThe interface/abstraction for is lacking. The fact it's a list, and lists are really hard to manipulate once set in NixOS modules makes this basically a hard task. I guess augmenting with a (default true) would help. But really what would be required is to tell per module whether it is needed, and allow overriding data. So needs to somehow allow setting \"yes/no/optional\" per-module, in an ideal world.\nQuick idea but what about adding a or something like that and then use setminus as the required kernel modules? Edit: or yeah depending on how it would work using a dictionary?\nUsing an attrset for availableKernelModules would be the solution. But that breaks existing use cases by default, probably. There might be something coming to handle this issue ; But this is a hard problem to solve correctly AFAIUI. I wouldn't like having the bad abstraction any more than the current breakage that I don't like.\nI suggest then to revert the commit that broke this until we have a better abstraction.\nI've merged that fixes this issue. Once it you can . You can build it on nixpkgs master using .\nI don't think it solved this issue. works around the issue by removing the raspberry pi 4 board-specific image. With merged, the modules are still to the raspberry pi 4 board-specific image, even if they're not present in the kernel build. In fact, there may be more failures! It ends up adding more modules present in mainline, absent from the vendor-specific kernel build.\nrecommends using the generic image anyway, so that's why I considered it fixed.\nAh sorry, I pre-emptively edited this page. And I'm splitting hair right now because technically the issue is still going on. But really it's all but merged-and-solved.\nIt appears to still be broken. Booting the latest image from Hydra ( ) results in the image being unable to find /nix/.../init, and failing to continue booting from there.\nI've booted the same image and it works. What steps did you take? Mine are:\nI performed the same steps, except the last one where I instead used GNOME Disks to write the disk image to the SD card. I don't have a way to reproduce it currently, since the Raspberry Pi 4 I was testing it on, has since been formatted with a different OS to be used in a production system tomorrow. I'll see if I can get my hands on another RPi 4B and perform the same steps again. edit With an exception that I used unzstd instead.\nSame, unzip was a typo.\nI've gotten my hands on one now for a few minutes, and it appears to be getting an error like: It prints this continually, until I press a key, and are presented with the option to reboot or continue. Continuing kernel panics, and reboot, well, reboots it.\nThe SD card used is the NOOBs SD card, but overwritten using the command as above, but with the device replaced.\nBooting with the same SD card using Ubuntu Server 64bit OS or Ubuntu Core works fine.\nIt sounds like it \"works\" up until stage-1, where the kernel is somehow unable to work with the SD card. It might be a combination of that is causing the issue. Do you have other brands SD cards, or same brand but (probably) different batches? It would be helpful to know whether \"Ubuntu Server 64bit OS\" and \"Ubuntu Core\" are using the mainline kernel or not. The driver differences might exacerbate issues on marginal hardware.\nThe underlying issue is now invalid with being merged. I invite you to open an issue, linking to this issue, for your particular problem. It seems you have some kind of hardware-specific issue at first glance.\nThanks, i will test it with other SD cards as well. It appears to only happen with this distribution, but it could certainly still be related to the hardware.\nIt is likely the other distributions are using the vendor kernel, and that the driver differs in a way where only mainline has a specific issue. Using the \"latest kernel\" variant of the image may also be something to look into.\nI've tried the latest kernel variant as well, but I'll create a new issue with more details once I've collected more information.\nI am trying to build NixOS for the raspberry Pi 4 and ran into the same error that started this thread. FWIW, I'm seeing this on nixpkgs and on as well. Would it be prudent to re-open this ticket or file a fresh issue?\nplease open a new issue and ping this issue with the specific problem you're observing detailed. Please detail what you are attempting to do (e.g. whether it's booting through SD or USB), and as much information as you can. Note that there is a known issue (totally unrelated to this one) about USB boot, ()."}
{"_id":"q-en-nixpkgs-5a621837f2060c050f771bbc79aba79967a4d25562d788a0a0058456bf301c4b","text":"Stressful Application Test (or stressapptest, its unix name) is a memory interface test. It tries to maximize randomized traffic to memory from processor and I/O, with the intent of creating a realistic high load situation in order to test the existing hardware devices in a computer. It has been used at Google for some time [...] In my experience it's an absolute monster in finding random memory errors that don't show up over weeks of real world usage. Until ECC is commonplace, running this software is necessary. Metadata source URL: license: Apache 2.0 platforms: linux\ncould you please give a try?\nworked no problem, thanks a lot!"}
{"_id":"q-en-nixpkgs-8891598e48aa52cc9ce04c28d9fde6534f7e2803481935fe39ecafa3852b75ab","text":"Describe the bug When i try to import a config file, i get an error: When i copy the exact same text in , it does work. The file does exist and i can read it, even with the caddy user! So this seem to be a bug. To Reproduce Steps to reproduce the behavior: ... ... ... Expected behavior A clear and concise description of what you expected to happen. Additional context Add any other context about the problem here. Notify maintainers Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:\nWithout knowing anything about my first guess would be this: Is it possible this command is evaluating your configuration? If so, the issue is that the command is being run sandboxed so has no access outside the sandbox to your file. Build time commands can be tough.\nThat would make sense. So NixOS brakes the import feature.\nOut of curiosity can you please paste your relevant config?\nwhat exactly do you want to see? For a workaround, i just copied everything into the NixOS option: But in some cases, it's nice to have the webserver config. together with the website sourcecode:\nThanks for posting. The context helps. I think this is what you want:\nIt could make sense to add an off-by-default option to opt-out build-time configuration check.\nGiven my above solution, I'm not sure what value it would offer in this scenario. What scenario did you have in mind Can this config file hold secrets maybe?\nThen you don't have to use such a workaround on NixOS and can just use it as described in it's documentation. There can be secrets in it, but i don't see how that's relevant here:\nHi, module maintainer here. I was the person who made the original PR for adapting s for Caddy 2, and I had no idea about the directive at the time. The motivation was to generate TLS settings from the module options and merge them with JSON-ified . Please use as suggested for now, and I'll see how I can make this easier in the future.\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-7b1ae0ec6a189bd01069c4032a0b575a14ee2d46413518818339ddb4149eb920","text":"Describe the bug While building the minimal installation iso without using the cache, building smartmontools on fails with: Indeed looking at it looks like the URL should be and indeed produces 11n54ycsy512q4linczzj1x3dyj0q4zvh928rq3lqnkg9jh0kbiv This was in and seemed fine at that point.\nthe difference between the cached and the downloaded file is:"}
{"_id":"q-en-nixpkgs-62e25edff1e53cffba19b60e5cd40f30f8ffbe682b5a6743de339cfaab45ae2c","text":"Describe the bug A clear and concise description of what the bug is. To Reproduce build e.g. or Notify maintainers Metadata Please run and paste the result. Maintainer information:\nI marked this as stale due to inactivity. &rarr;\npkgsCross.aarch64-multiplatform.perlPackages.PkgConfig still fails with the same error on\nHad a quick look: Seems like PkgConfig depends on File::Glob::bsdglob during build (which uses miniperl). Maybe adding a for wrapping to in perl-cross is the way to go? PS: I'm not sure if this is the correct approach, not very familiar with cross compilation. Any ideas on how to fix?\nconsidering the failure in the here actually comes from a failing , so a simple fix might be this: This has allowed me to cross-build on x86_64-linux: I'll post a PR shortly for your review :muscle:\nAs a follow up, the resulting / scripts seem to work: (I have multiarch exec via and enabled BTW)"}
{"_id":"q-en-nixpkgs-4b8423737fedffb50bf09c0e1797b046e93bde891493e0f0175a4ac4840eda2d","text":", CVSSv3=7.5 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: . Cc\nSee also:\nCVE details indicate that is vulnerable until version (included), which is the version we have in . Version seems to be the latest available: .\nI'm actually wondering why this issue was created. We have thrift 0.13 in both unstable and 20.09. Closing and debugging my code...\nI think the issue was rightfully created: it says INCLUDED!"}
{"_id":"q-en-nixpkgs-b55e5da21cf6f0a7527ecd9d60b3be587df12d2e6dde4e1de037847e9fb1890e","text":", CVSSv3=7.5 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: . Cc\nSee also:"}
{"_id":"q-en-nixpkgs-41b5f2fdb04297b0cc8af1214da8b06683d55b551f8625bc942f2c672f89d525","text":"Project description Flawfinder is a simple program that scans C/C++ source code and reports potential security flaws. It can be a useful tool for examining software for vulnerabilities, and it can also serve as a simple introduction to static source code analysis tools more generally. It is designed to be easy to install and use. Flawfinder supports the Common Weakness Enumeration (CWE) and is officially CWE-Compatible. Metadata homepage URL: source URL: license: gpl2 platforms: unix, linux, darwin\ncould you please test Thanks\nWorks for me on NixOS. I wrote a simple c program that it found the vuln in: where it gave:"}
{"_id":"q-en-nixpkgs-d94a4a54949e0ee9854a74ee457315a7dbedef6dd5da980e37a9a9193b96608c","text":"Describe the bug I switched from to on KDE. But the audio control tray icon is gone. Media keys don't work either. Note that can adjust audio volume and switching devices without issue. Audio output is also fine. Not sure it's an issue from KDE or pipewire side. Here is my configuration: and are both running. The latter one has some error but still running as well. To Reproduce Steps to reproduce the behavior: X11 and KDE. sound configuration above. , log out and re-log in. audio icon in system tray is gone. Media keys don't work any more. Expected behavior Audio control from KDE and media keys all work. Screenshots ! Additional context Also mentioned in <!-- Notify maintainers Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Metadata Please run and paste the result. Maintainer information:\nI believe the solution is just to have and installed, so you can install them manually for now and that should work.\nI switched to pipewire some time ago and subsequently forgot all about it (things just worked) and I haven't had the tray icon disappear at any point. I cannot comment on media keys. Keyboard doesn't have them.\nAfter adding and into , and a reboot. The tray icon appears in a wired state. There are two forever-muted icon shown. One prompts the original \"Audio Volume\" window, but shows . Though the audio plays well. ! Another prompts a small window shows several volume slider. The \"Master\" has no effect, but per-application control works. (Another strange thing. When opening this window, any shortcut of KDE doesn't work anymore, including my screenshot key and ) ! Pressing any media key including toggle mute, increase and decrease volume keys, always shows a \"Audio Muted\" tip.\nNote that doesn't work anymore after updating to current nixos-unstable . It shows \"Establishing connection to PulseAudio. Please Wait...\" forever, while are just running fine. Not sure if it is caused by the pipewire version bump\nSidenote: should be removed from your config, according to . I think it enables ALSA. Kind of annoying, but similar to how should be false if using .\nTried but it made no difference.\nWorks fine for me on current unstable with and installed through home-manager. I have the pipewire setup in .\nI'm using pure pulseaudio facing the same problem. After installing kmix and plasma-pa, I also get two audio icons and the other situations like the picture. I believe one of the is for the original plasma which is muted and the other is from kmix. By the way, I'm on an old laptop and the media key is not working until I installed kmix.\nSorry for necroposting. Any updates on this? I have this problem too."}
{"_id":"q-en-nixpkgs-3c2ace2c518f37fbb09924bbfbe89b0206ce25aeb906870a88050aa95251de98","text":"Describe the bug fava does not build on master since the merge of staging-next in : To Reproduce Steps to reproduce the behavior: -env -f ~/code/nixpkgs -iA fava Notify maintainers hi"}
{"_id":"q-en-nixpkgs-f7bcffe2a720bb1ee2a071daa4b6388f8a732b929b7b28a55ce32dbf3abaeb97","text":"Project description The recoll package exposes a python API. It'd be awesome to have a package for that. Metadata homepage URL: related: source URL: Main package already exists - platforms: unix"}
{"_id":"q-en-nixpkgs-7dabef8331e8f26878bea2ec719757eefdbed1336135c51e48b1d8a9e5bc9d7c","text":"Describe the bug Upon enabling docker via sysctl.d is not updated to allow IP forwarding which is required by containers to communicate with outside world. To Reproduce Steps to reproduce the behavior: This will get up and running with set to which in my understanding is set by the itself. to re-apply system-wide sysctl settings and check By now your bridge-mode docker containers lost connectivity. Expected behavior I expect to override relevant values. Something like this: Additional context I discovered this problem when applying other sysctl values on a running system without a reboot. Notify maintainers Metadata\nHi, I stumbled into the same issue. I had to manually run to regain container internet access."}
{"_id":"q-en-nixpkgs-4873a2af18d299a2edb7132217aaa2966c12d07288f716a7d5788d33804d3f35","text":"name: current version: desired version: latest version: maintainers: Please tag this issue in your PR.\nYou'll have to package pydeck as well\nHere's my hack to get things working: And WIP:\nI can PR this for you if you accept maintaining the dependency.\nping"}
{"_id":"q-en-nixpkgs-ed691b76beb23be2d8d529f228d4ca62e4762372bfb54489c4ee201d444fcb87","text":"Describe the bug Postgres already exists as a nix package () bundled with some packages (). Postgres itself has built-in support for Python, but it needs to be enabled during the build process () via the configure flag. The current implementation of postgres in nix does not enable python support, and it also doesn't seem to perform any detection whether python support should be offered. There is already some form of conditional feature support (for example via the attribute. It could be possible to add a attribute that enables python. My knowledge of nix is clearly lacking in this regard, but I hope the idea makes sense. To Reproduce Steps to reproduce the behavior: via then use to initialize and start a server instance psql to enter the postgres instance 'create extension plpython3u;' ! Expected behavior The ability to enable the plpython3u extension. Screenshots This is a screenshot of the current configureFlags of postgres (). Somehow it should be possible to get into this list. ! Additional context I am trying to get a nix-shell working that has a postgresql12 with python support. I am not trying to get a working configuration in NixOS. It should be possible to use to explicitly add support for plpython. I am still exploring this possibility. Others have done something similar for other postgres extensions (). There is a nix pill that describes how to override packages here () and here (). However, it looks like this can get complicated (). More examples of overriding can be found here: Another approach for overriding packages is via overlays (). By including support for python in the derivation itself, it may be easier for new users to get started. attempted to add support for multicorn (), which is implemented using python (), but this doesn't use plpython3u, and is thus not directly related to this issue. Just mentioning it here for others that are trying to figure out how to get python in their database backend. The author of the following article shows how to configure postgres using : I can imagine a solution that may look like this: . Another possibility, is to have something like (see etc.). Notify maintainers Can you recommend a way to get plpython3u enabled in the nix build of postgres? Metadata Please run and paste the result. Maintainer information:\nThe following extract seems to work so far:\nI marked this as stale due to inactivity. &rarr;\nI'm interested in using this but not sure when I'd be able to pick it up."}
{"_id":"q-en-nixpkgs-45e9f09674f500b9e456ea26d564cebfa6a5cb81abb395127782938933b6bf4b","text":"Describe the bug Library is not found when linking on macos. To Reproduce Execute resulting in following output: Expected behavior Deno builds without errors on macos. Notify maintainers Metadata system: host os: multi-user?: sandbox: version: channels(root): channels(suraj): nixpkgs: Maintainer information:"}
{"_id":"q-en-nixpkgs-4bc7acbfd25709a6a97f13da5560ba2324b81772d5b82a3570755cf7337926e6","text":"Describe the bug For sandboxed builds on my mac, any python tests which ultimately import will fail with this error: To Reproduce Somewhat tricky to reproduce since it only affects downstream dependees, and a recent commit disabled the dask tests, hiding this issue. See for an example. Additional context Dask.dataframes infers an optimal chunk size using a package to get the amount of available memory and CPUs. This non-determinism breaks the build. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --cc Metadata This is not valid since I am building from master of nixpkgs. system: host os: multi-user?: sandbox: version: channels(noah): nixpkgs: Maintainer information:\nThat's a clear upstream bug; apparently their value is not given."}
{"_id":"q-en-nixpkgs-92bb6b783e1cc602f130d20a8535f89fb1ac707dc3b3b9e8683bb655230ad778","text":", CVSSv3=8.8 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: . Cc\nSee also:"}
{"_id":"q-en-nixpkgs-9f4c003a786a66e40acbe696d34bac29dd0de301375f4c6549f9f5965407804c","text":", CVSSv3=2.7 (nixos-unstable) Scanned versions: nixos-unstable: . Cc"}
{"_id":"q-en-nixpkgs-77ecefd980871099ff665bb00acfd8038a554b50828ecdf10b4d847b5ecff0eb","text":"Describe the bug Trying to run a nautilus-python extension (syncstate-) fails with ImportError: could not import gobject (error was: ModuleNotFoundError(\"No module named 'gi'\")) To Reproduce gnome3 and install nextcloud-client nautilus from terminal with something like Expected behavior python should find gi. Additional context I think it is part of pygobject3 which is in buildInputs therefore I think it is a problem of a wrong NIXPYTHONPATH Notify maintainers Metadata system: host os: multi-user?: sandbox: version: nixpkgs: Maintainer information:\nI am not sure if it ever worked. The shared module links against pygobject but that still does not have to mean the gi module will be available on Python side. Maybe we need something like the following: At least until we have .\nYes this works, thank you!\nhmmm I just tried with the new nautilus-python, but the nautilus python extensions still fail due to missing gi. Is there a reason that you tried to load gi in nautilus-python and not add the prefix arg as in your comment?\nDid you make sure is updated? It worked for me with gsconnect extension. I chose that method because pygobject is really nautilus-python dependency, not nautilus one.\nnevermind it works now (also with nextcloud), I forgot the in the overlay where I overwrote the package from master. Thanks for fixing it!"}
{"_id":"q-en-nixpkgs-6aadc1b69a16bb01246a3c5aed3e108023e8fa7654aab73f945f346737c54151","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on github to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: ghc <!-- The current version can be found easily with the same process than above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: desired version: <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR.\nToday I tried to install , but it failed at downloading : complained about hash mismatch. It seems like URL has changed.\nI used the following overlay as a workaround:\nShould work with ."}
{"_id":"q-en-nixpkgs-c416c576c1ac4692206020926708f963c89af30e4dd401836734cc193284dab1","text":"nixosTests.blockbook-frontend fails to build Discovered in Refs Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Maintainer information:\ncc\nTests are \"fine\". It's the main that's causing the issues (failing to build) and it was not working for some time now (as I've discovered also ). I did some initial debugging and broke the . I've notified upstream but I guess there's little I can do about it so it either gets fixed or we should remove the package altogether.\nHm, I suggest we mark blockbook as broken for the time being:"}
{"_id":"q-en-nixpkgs-e52a2279dfd9ec33fd0f540165def08233c835a36ae42525f74d473a4dfc4005","text":"Describe the bug nextcloud default nginx config did not include the ` config see: but expires is set in Expected behavior no expires, the cache option causes problems with the usage of Deck"}
{"_id":"q-en-nixpkgs-d1dcf88f6d830297228cda21f09d359b7d36163d5d402a7eb82c9f2408cd3477","text":"Describe the bug The phosh lockscreen will accept any pin. To Reproduce Steps to reproduce the behavior: phosh screen with 1 even if it isn't your pin! Expected behavior Probably use the numeric password of the user Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Solution The PAM configuration seems to be at fault. I can't PAM my way out of a paper bag, but I tried one attempt: Why ? Here's why: That doesn't explain the issue. Only a plausible fix. With this, only the user password is accepted. We should get someone who understand PAM to look at this. Metadata Ran on top of . Maintainer information:\nUh, so the problem is that does not exist, so there are no requirements enforced? Or what does our pam file for phosh look like? The better approach would likely be to add here:\nSorry, I probably misexplained a bit here: we're not using the manjaro pam bit at all. So I suspect those are not good at all. When using the manjaro bit verbatim, you just can't unlock! I guess adding to that list you linked to could do it too? If the other snippet is removed.\nOuch, that pam file is not testing for anything meaningful, so it shouldn't come as a surprise. Login works if you are not root. pamnologin is a PAM module that prevents users from logging into the system when /etc/nologin exists is designed to succeed or fail authentication based on characteristics of the account belonging to the user being authenticated. pamsecuretty is a PAM module that allows root logins only if the user is logging in on a \"secure\" tty, (Quoting from their individual manpages) Because we have no pam.d file."}
{"_id":"q-en-nixpkgs-a78d729b251507e71ff78bd31c3da8cc5beace659e928a6032b689cec87e042f","text":"Describe the bug json-glib seems to be a new transitive dependency of tracker/gtk, which the Paradox launcher needs. Somehow it isn't being pulled into the Steam container. To Reproduce Steps to reproduce the behavior: Steam a Paradox game (eg. Stellaris) to start the game Expected behavior Instead of the launcher working as usual, it doesn't launch at all. Launching it directly yields the following error: Additional context Adding json-glib to steam's fhs wrapper libraries is sufficient to start the launcher, but should it be propagated by gtk implicitly? Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Metadata Please run and paste the result. Maintainer information:\nwhich game? I tried eu4 and CK3 and both launch fine\nStellaris. Funny that it seems to work with the others, I suppose it could be something specific to my system?\nUnlikely, steam games run in a bubblewrap sandbox, so your host system doesn't matter much."}
{"_id":"q-en-nixpkgs-9bc7c535284084b3e0a52d6c619095d97bfa5e639503a93a9e9074b7bdd5ff8f","text":"It used to be possible to build (it is possible on 20.09), but these days it fails with: I’ll try to bisect it…\nBisecting points to by Nathan, do you have a hunch?\nCargoculting what I saw in the GHC code, trying this right now:\nSeems to work, fix in\nThere appear to be multiple linker bugs affecting at least Rust and GHC with Musl . Both bfd or gold had issues, gold seems a little worse. I would suggest trying lld, it may fix it.\nlld with gcc is currently not supported and I think last time I tried we can't build OCaml and all its dependencies with clang+lld fully yet…\nThe second commit of (\"match leading/trailing arguments\") should fix this\nThat's great news!\nFixed by"}
{"_id":"q-en-nixpkgs-6853ba43be1c26c39f644f3dedce2694437c2027368e5cf0bfeca5c44d3d57f1","text":", CVSSv3=7.8 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: . Cc Cc\n09 is still vulnerable\nNixOS 20.09 is EOL, so this issue isn't actionable anymore."}
{"_id":"q-en-nixpkgs-d25ede83cad91608e0de8cf3868f808a91dae043acbfd7a22778c7f0209bbfb7","text":", CVSSv3=7.5 (nixos-20.09, nixos-unstable) CVSSv3=7.5 (nixos-20.09, nixos-unstable) Scanned versions: nixos-20.09: ; nixos-unstable: . Cc Cc"}
{"_id":"q-en-nixpkgs-2de419a81dec9dc2c6ce6909a1c6b4261657622e5278631cc1cb5a932d0fc5c5","text":"Describe the bug fails to build on master/release-21.05. The build failure ( is not found anymore by cmake) appeared with the update of from 2.9.10 to 2.9.12. Addressing CC"}
{"_id":"q-en-nixpkgs-46a6afbfb48ffe96d96bcf89dcc898b900b7d31b1c3d7e152aa671a70346f18d","text":"Describe the bug The rhdf5 library does not work on latest release. To Reproduce Expected behavior A shell with R and rhdf5. Additional context Instead I get: Notify maintainers Metadata Maintainer information:\n\"6.topic: rust\" should be removed, this is an R package, not Rust.\nThis was fixed in"}
{"_id":"q-en-nixpkgs-0b3d179eb12989379012ce3cf0bd439b2be1187ee4c9485594ccb1ddc5507faf","text":"Describe the bug I believe the auto-cpufreq daemon should start on boot if enabled. To Reproduce auto-cpufreq daemon: . the system. . auto-cpufreq as dead/inactive. Expected behavior auto-cpufreq service should be started on boot. Additional context I was using the unstable service in 20.09 and it worked as expected. In 21.05 stable, the service starts dead/inactive and my system is very slow. I manually start the service and I'm back to the NixOS I know and love. I was able to fix this by adding the following to my configuration: Notify maintainers Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(pnelson): channels(root): nixpkgs: Maintainer information:\nSorry for being late, should fix this.\nI have the exact same issue now, on 23.11.\nSame here, also facing this problem on 23.11...\nThought this might be caused due to some false overwrites in the module config, so I decided to try out rewriting the systemd service. Doesn't work either...\nFor others who might still be hitting this, I thought I was too. When checking the logs I saw that power-profiles-daemon was starting up, causing auto-cpufreq to shut down. I fixed this by adding: From looking at this code, and be kind with me here since I'm a nix beginner, it is .  I think the best action would be to add an assertion to give users a prompt like with tlp. If this looks correct I can open a PR for it"}
{"_id":"q-en-nixpkgs-3d3cb1bf3c17fd3cd18b80aef8009431bd2467f55a09251eebad3b02924d8860","text":"As reported on we need some patches: In particular: gnome-settings-daemon needs bluetoothd needs cc\nAt least for g-s-d we're already on 40.0.1, which seems to have the fix? Both commits referenced (!232, !234) are there, but maybe I'm misreading. Not sure if we want to address bluetoothd with , but regardless I have a commit ready.\nAbove is no longer necessary with 5.60. I'll open a bump for that instead."}
{"_id":"q-en-nixpkgs-35e3a95ae05980479eaba39193774cb7b64ebbf3b35c8d4d5575878c56bcce4d","text":"Describe the bug The current downloads 2 patches, both of them 404. I realized this when I tried to build my system from scratch without substitors. To Reproduce a system that somehow depends on mercurial (not exactly sure if this is reproducible with just or similar, this was also on a content addressed machine, and reproducing the issue in a more isolated way would require another toolchain bootstrap which takes several days) the system from scratch using see errors like this: Expected behavior mercurial builds without errors Screenshots See log output from reproduction. Also ing the URLs as in the derivation linked in the description result in 404. Additional context The exact commit the system was build from is nixos/nixpkgs/, though as the same path URLs are used in nixos/nixpkgs/master, this is probably only partially relevant. Notify maintainers cc Metadata nix shell nixpkgs#nix-info --option trusted-substituters -c nix-info -m system: host os: multi-user?: sandbox: version: channels(nmelzer): nixpkgs: Maintainer information:"}
{"_id":"q-en-nixpkgs-df0b4671da53af9481a034dc3460f9c803b82c11fec036a098560278d304b9fc","text":"Describe the bug I'm on a mission to get (and thence ) compiling on M1 Macs, and having fixed , the next stumbling block is 4.1.1. Sadly, the build fails with , presumably related to . To Reproduce Steps to reproduce the behavior: Expected behavior A clear and concise description of what you expected to happen. Screenshots Additional context This works fine when built for x8664 on the same machine. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Metadata system: host os: multi-user?: sandbox: version: channels(steve): nixpkgs: Maintainer information:\nSubmitted PR which disables Fortran support for OpenMPI on aarch64-darwin. Would that fix the issue for you or do you require Fortran?\nThis seems to be the build failure:\nI actually have no idea if I require Fortran. I mean, presumably it's enabled for some reason, so I'm a little wary of disabling it on only one platform/arch combination: that feels like an atypical workaround.\nTo reproduce: I wonder if the gcc maintainers can offer any insight?\nseems to hint at the stack-protector hardening, so perhaps something like could work around that? (no idea what's the real cause though, and I have no Darwin machine at hand)\nYou were right! Sent a PR in"}
{"_id":"q-en-nixpkgs-bfb4e8b9db14d10f329721576c2eb8d6ebb4147bb69b55501afc4c9f82bc23fc","text":"This tracks the conversion of from DocBook to CommonMark per . here that you intend to work on this. (Please also post if you stop working on this.) a new branch of nixpkgs master. : change to . to . the contents from Docbook to . Follow . Use existing sections in the NixOS and Nixpkgs manual for inspiration. the CommonMark back to DocBook xml with: from nixpkgs root dir with: a preview server with: your work at on steps 5-9. it looks good, commit both the new file and the generated file, submit a PR and reference this issue."}
{"_id":"q-en-nixpkgs-b3c7731867d7c979f862eabf5a0bcd8d9c8f653e127f9b37f22a72ec49a20432","text":"has two forms of input for the field. and . I personally use the latter as it is the output of and better than having cleartext passphrases. (This also works with a random AES key - so long that it is shared) The current implementation lacks the distinction and only works with passphrases. I believe this should be ."}
{"_id":"q-en-nixpkgs-3cc1ee2546405ebaf916a730f257ee2ce476fefd50d03223ce75226d69f90f4a","text":"Project description\nI marked this as stale due to inactivity. &rarr;\nping! Is this still an issue?\nhas been The command still does not have documentation in the Nixpkgs manual.\nIs satisfying? Could use approval to get merged.\nIt's a good step, but it doesn't (yet?) tell me what the options are.\nNow it does."}
{"_id":"q-en-nixpkgs-c99467ff4867c2f2cca4ae7623088c2356a96b2d6a1681d9faa8bb84871a1986","text":"Project description Emote is a modern emoji picker for Linux rocket. Written in GTK3, Emote is lightweight and stays out of your way. Metadata homepage URL: source URL: license: gpl3 platforms: linux"}
{"_id":"q-en-nixpkgs-4b3e0825c7f418486d4ba73f3cb3778608f267e1ff84375a70af6c70349dadd7","text":"Looks like it is missing python3Packages.setuptools...\nworks for me - please provide output from\nfixes that, but then you will be running into\nI have informed the author of archivy, and he's looking into it.\nThis is because in the version requirements are loosened and an incompatible new version of flask is used to run archivy. Removing this change to the version constraints for flask would fix the problem."}
{"_id":"q-en-nixpkgs-9e18c5cdc4a8ae4ad1ff780d6df4f9ac39f79463443d2dd6b73afa0e0814614d","text":"gPodder crashes and dumps core when trying to import an OPML file: Steps to reproduce the behavior: \"Subscriptions -Import from OPML file\" the crash I'm using KDE Plasma, not Gnome as a desktop, in case that's related to GSettings discovery. gdb backtrace from the coredump: <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result. Maintainer information:\nThis crash does not appear in the flatpak verision of gPodder, so I assume it is related to our Nix packaging.\nLooks like something on my system sets the right environment variable to prevent this crash: I tried to add dconf, gsettings-desktop-schemas and hicolor-icon-theme but nothing seems to be picked up by the . Maybe has a better idea? The crash only happens when I do with nix-shell.\nThat would be schemas from gtk3 if it crashes on trying to open file chooser.\nYeah but gtk3 is already in so something is off here with\nIf it is Python package, it needs strictDeps false\nindeed fixes this issue, see PR\nThis is quite a footgun, can we disable from somehow?"}
{"_id":"q-en-nixpkgs-046df70540cc6db7baca741585f35fb23e8a5663acec56f3997ed7e277ac4625","text":"See the : the options aren't declared by any file (it's actually , but that's stripped at some point). I think this is most likely because they are declared inside a : but I don't know the module system well enough to troubleshoot this. This was causing the unstable channel import to because of an which should be fixed by This is an issue of the module system, so pinging\nfixes this, thanks for the report!"}
{"_id":"q-en-nixpkgs-830ecea0cb685531ee0955a17c02af1e8159e747ea78fba2737d9b3ce5fb83af","text":"This project is no longer activily maintained! there is a spiritual successor, so maybe package that instead cc maintainer\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-69568eadcc8d60dc3f808812eb03d0e5b5e72b07cfae76935f7a5d3f41780599","text":"NixOS's wrappers for gcc also apply for gcj, which is a Java compiler. The wrappers add the flag to gcj's command line, however the Java language does not have a preprocessor, and means a different thing for gcj. This means that gcj cannot compile anything. Another issue is that gcj links to zlib by default, but does not pull zlib into the environment. This means that, even if the former issue is silenced with , compilation still wouldn't work.\nA similar problem for gfortran:\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-0af86241f29b7a4801588018a0812989b5676dcd9c78c6c55224d1b2adaf6e1f","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on github to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: <!-- The current version can be found easily with the same process than above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 4.6.3 desired version: 4.11.2 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-0de0e7173e1dc7303657f058a84cbc91eeb65a98b06b1c01bf1fcd7e271490ef","text":"After enabling the GNUnet service with , it prints a lot of errors at startup: Additionally, attempting to stop the deamon with reveals that it is apparently not even running: Please run and paste the result."}
{"_id":"q-en-nixpkgs-a5856cb7d2d6faacc803720f0e3895cf45fb7ca4c91389d2c15d65101e25a969","text":"When adding to and running , emacs 25 builds and fails with 4 unexpected TRAMP mode test errors: The full output contains null bytes and weird control characters that don't let me paste this into a gist easily. Is that part of the test or indicative of the problem? cc\nok... i think this will contain the build output? again if i open it in emacs there are lots of control characters. had to zip so github wouldn't complain.\nI disabled for emacs25pre, for now, on my system. I hope this will be resolved upstream before next release ...\nIt's quite possible that tramp-related tests fail because there is no proper networking support in the build chroot, i.e. it's not possible to resolve host names or service names. Another source of trouble could be the lack of an UTF-8 capable locale in our default build environment. Anyhow, it would be really good to report those errors upstream to make sure that the next Emacs release has them fixed in one way or another.\nI've been working with the Tramp author for the past several days about these failures, and we've identified that the problem is somewhere in the environment that the tests are presented with, but we don't yet know what exactly. I'm still digging into it."}
{"_id":"q-en-nixpkgs-7ab4a1529ef51c2cc2aec2e1cdadb0cf91a595f1c0dc8b492bfccc80d8374259","text":"I tried updating it, but no luck. I think the new glibc 2.23 is the reason for configure fail: I've only been able to find arch having the same issue, no patch yet. cc\nThis is strange. I do reproduce the build problem, but when I compile the configure test and run it myself, it succeeds (and measures bytes). (Of course I do compile it with the stdenv from 16.03.)\nComments on note a workaround.\nTemporary workaround on arch:\nFixed then?\nFYI the fix to Xen build system is committed. See ;a=commit;h="}
{"_id":"q-en-nixpkgs-8eafcd7d432d35c763e9bed6b773b609e673e30fae0cb394682eb700b8d0f604","text":"I want to use the TTS Mimic on a Debian 11. After installation, I Tried reading a text. Cli usage give me an \"alsa lib error\". Steps to reproduce the behavior: -env -iA -t \"hello world\" A clear and concise description of what you expected to happen. If applicable, add screenshots to help explain your problem. Add any other context about the problem here. Please run and paste the result. Maintainer information:\nworks perfectly on NixOS 21.05. Can you elaborate more about your sound setup? While geared toward NixOS, this thread might give useful information to debug further: Also try running ALSA tools or CLI audio applications natively from Debian and via to observe any differences in behaviour.\nThe above works for me too on NixOS, at least. I'm not familiar with how ALSA and PulseAudio in nixpkgs integrate with non-NixOS systems; perhaps a thread on Discourse might be of more help.\nI was able to reproduce your issue on Ubuntu 20.04. It seems this is . A few workarounds are mentioned in the linked issue.\nIt seems mimic just needs a proper wrapper: should give you some output.\nOk, I found a solution with inspiration. alsa-lib-1.2.4 doesn't containt but if i installed alsaPlugin (nix-build -A alsaPlugins '<nixpkgs'), this works for me : However, not really user friendly."}
{"_id":"q-en-nixpkgs-93fb3fb51b0d59bf5896b5450edcec65c55e4ce651a3053dd582f8daea58c911","text":"Project description A cross-platform TUI database management tool written in Rust Metadata homepage URL: source URL: license: mit platforms: unix"}
{"_id":"q-en-nixpkgs-6f9a01426ba98b588ad30f7d272f88aa4b2d46f32904f0d2e643cb64ed038a83","text":"After upgrading to Firefox 92 on NixOS unstable, anything that requires a menu or a tooltip in the browser (hamburger menu, right click context menus, tooltips on tabs, etc.) stopped working. The menu itself isn't rendered, the whole UI freezes and sometimes the cursor changes to a hand, indicating a clickable button, but nothing happens when clicking where the button should be. Clicking outside the area where the menu should appear restores Firefox to a usable state, until another menu button is pressed. Steps to reproduce the behavior: Firefox the hamburger menu button on the top right corner The main settings menu for Firefox should appear. The hamburger button remains in a \"clicked\" state and the entire UI freezes, the mouse cursor might change to a hand, but no menu is rendered. Running Firefox on Sway with and Maintainer information:\nI can reproduce this too, but I don't think it's NixOS' fault; we ought to report this upstream to sway and Firefox.\nYou can manually apply this patch to firefox to fix the issue: I'm testing that now.\n~You should be able to use this overlay to work around the issue for now:~ This partially fixes it\nAlright, this overlay fully fixes the issues while we wait on a new release: I'm going to make a PR adding this to our pkg for now.\nThe two patches applied, menus and form overlays are visible again and the freezing is gone. Thank you\nUnfortunately, it seems the fix has not worked for me. It may be due to using the home-manager module; however, I have overriden its to use the patched version and the only noticeable difference is that now a tiny vertical line appears where the context menus should be. The freezing caused by it and the fact that the menu itself is still unusable are still present, though.\nThere were multiple similar bugs, perhaps?\nAs a data point: I'm running Firefox 92.0 with the above mentioned patches on Sway on NixOS unstable and everything works great.\nDo you happen to use the home-manager module or do you run Firefox as a regular system package? That may be a differentiating factor and may point towards an issue with my home-manager configuration instead.\nThrough home-manager.\nI have tried to add your method of explicitly wrapping the package and yet, the same behavior remains, unfortunately:\nI wonder if this could differ based on GPU model. (different rendering code paths on driver side)\nInteresting. It is working for me on amdgpu, but I saw the issue yesterday on an i915 gpu, didn't investigate much, whether the patch already made it onto the unstable channel.\nYes, it is in already but not in yet.\nI have tested this on amdgpu, nouveau, and i915 without problems. I'm wondering whether is actually using the patched version? Try to clone the repo, do and see if it works correctly. Alternatively, this should give you a firefox with the patch so you can make sure:\nAh, I managed to reproduce it on i915 now. It's not immediate for me, after a bit the menus seem to break. Amazing.\nMaybe we also need this GTK patch:\nAlso worth trying:\nInterestingly, the firefox-unwrapped package in my system has a different hash, namely . Compiling from the nixpkgs repo as you suggested does give me the version with the hash as you stated, and this version does behave correctly in regards to the menus. Something really strange is happening with my system updates then. Adding to this, I only got the version with the \"black vertical line\" as I described above (I attached a screenshot for reference) after I garbage collected a few generations of testing, which really shouldn't happen if I'm understanding the way nix works correctly. In any case, this is not an issue related to the Firefox package itself. !\nI successfully have this working with firefox from mozilla's pkgs.\nSo I was able to fix this by removing the package from my home-manager config and adding it to my nixos-config. Not sure if this helps anyone here debug the issue, but it's worth a look. EDIT: Well, menus are fixed. Tooltips are still messed up.\nI've been running Firefox out of a local nixpkgs build but I haven't tried messing around with the system config yet since then, I'll try your suggestion. One thing of note is that some sites' dropdown boxes are still broken, even with the patches, while everything else has been working fine.\nI don't know what exactly is wrong but I am getting the random flickers on Firefox 93 as well. Has anybody found a solution?\nI still get this issue sometimes. It's odd because it's very hit-or-miss for me; restarting Firefox often fixes it, and restarting it again results in it being \"broken\" again. The same thing happened for a long time with the screenshot tool in Firefox: sometimes it works, sometimes it doesn't, and restarting Firefox sometimes \"toggles\" if it does or doesn't work. I haven't been able to debug any further, though.\nI haven't experienced this issue anymore since Firefox 93; both the Bugzilla issue and the GTK merge request report this issue as fixed, does anyone else still have this problem?\nCan you share your exact config related to Firefox so that I can try it? Last time I checked I still had the same problems in 93.\nIn , add to the list of packages. Then, in : EDIT: shouldn't make a difference, but for completeness' sake, I also have this in :\nThanks! Do you use any environment variables such as ?\nI actually didn't use that, but maybe I should. is set by using the package.\nOh no. The flickers are still here for me with the following config.\nThat is troubling... might it be a GPU-related issue?\nI marked this as stale due to inactivity. &rarr;\nI think this can be closed as it has been fixed for me since 94 or something."}
{"_id":"q-en-nixpkgs-d7790ee2ed0194a2bf7f742382028587e22b796cdf9012dda59a48f046e4da60","text":"Currently we have two expressions that both build the kernel, whose only interface differences are in whether or not the .config is generated or provided. These builds should be merged with the two different interfaces simply being two ways to call the same underlying function, reducing code duplication and allowing the generic kernel builds to benefit from the IMO better structure of the newer manual kernel builds.\n+1 on getting rid of duplication What is the \"better structure\"?\nStructure may have been the wrong word, but the manual kernel build uses multiple outputs (suboptimally currently, which was part of the problem with the old PR) and uses the official kernel build system for building and installing, while the generic one (last I checked) had a manually maintained shell script to handle the install and parts of the build.\nFixed by"}
{"_id":"q-en-nixpkgs-98567d407acb303d913ae96abb70a2e1fb68cf086e7e55fba89724f5ae3d5d8b","text":"Practical consequence: with (this log snippet is cleaner thanks to ) I checked that it really is the patch by applying Some bugs are discussed in the upstream ML thread: I'm not sure about the best approach for us so far; the rebuild amounts complicate this. Perhaps just temporarily fixup the kernel build and get a properly fixed cpio with the next iteration? I expect that three to five upstream commits will be needed; maybe just the whole diff since the last release.\nI think we should just fix this in master; the channel isn't going to advance anyway, and I'd rather not ship a known-buggy version of by fixing just the cpio used by the kernel. Is that reasonable?\nMaybe the current channel version with public code-execution exploit (linked above) is a worse bug than this, but I don't know how commonly this bug manifests. ( channel has advanced, by the way)"}
{"_id":"q-en-nixpkgs-5f75f7da0756b53930bde3ac5b4263981822b0a592552ed26de613dfb1a1b7c1","text":"fails to build as There is an invasive change that fixed it in newer version: . But maybe it's a good time to remove old version? Maintainer information:\nStarted with commit\nIs there a reason to still keep Ardour 5.x in nixpkgs? Edit: The reads: Is that still the case ?\nYes it is, but I'm not sure it's worth the effort to fix Ardour 5.\nI prepared a PR which drops ardour_5 if we decide to go this way:"}
{"_id":"q-en-nixpkgs-887efb1b4853673627a00bd915d24cca3f92ff518690868c4766e1721012f72d","text":"insync-v3 server doesn't seem to start. See that there is no insync process running. Please run and paste the result.\nI believe this is caused by the Insync binaries expecting specific versions of some system libraries like glib. A similar problem has been reported on the Insync forums by a few Gentoo users. See and Since Insync releases are pegged to Ubuntu/Debian/Fedora releases, I'm not sure there's a good fix for this. Fortunately, I was able to get insync v3 working in a nix-shell environment that is pinned to an earlier commit of nixpkgs (). I uploaded a reproducible flake here: The system icon shows up as a red circle, and native Wayland support doesn't work, but it seems fully functional other than that.\nUsing nixpkgs pinned at , I've been able to successfully run insync version (the version currently in nixpkgs) and version (latest available version). This suggests that there was an update elsewhere in nixpkgs that wasn't backwards-compatible with insync. Pinning nixpkgs works for now, but it's not a long-term fix. Maybe we can try to identify which libraries need to be pinned and explicitly pin those? Or maybe use binaries targeting a distribution which is more up-to-date?\nI've noticed this as well a while ago, and even figured out what the issue was, but was too lazy to report it until now. Sorry about that. When you run it should tell you why it can't start. In my case it had some issues harfbuzz, which most likely comes from the upgrate to version 3 . Harfbuzz removed some functions that were not officially documented, but are used by qt webview, which is why it fails. I think this is going to be fixed automatically at some point with either an insync update or an update of qt webview in nixpkgs. But until then, your best bet is to use insync from an older nixpkgs commit. It might even be a better idea to completely remove insync from nixpkgs and providing it as a flake, that has a working nixpkgs version pinned to it.\nFor me just produces :(\nHi how can I test this pinned version? Thanks!\nAny fix for this? insync start --no-daemon Traceback (most recent call last): File \"\", line 19, in <moduleFile \"site-\", line 829, in call File \"site-\", line 782, in main File \"site-\", line 1259, in invoke File \"site-\", line 1066, in invoke File \"site-\", line 610, in invoke File \"\", line 87, in start File \"/usr/local/lib/python3.7/site-\", line 627, in execmodule File \"\", line 11, in <moduleFile \"/usr/local/lib/python3.7/site-\", line 627, in execmodule File \"\", line 7, in <moduleImportError: /nix/store/41ms51yaqhqwfp83z8dms55g4abfzihp-insync- version `GLIBCXX3.4.29' not found (required by /nix/store/z8hz4nlh48niad995hs4w33wwryc3lkr-qtbase-) [2847] Failed to execute script insync\nI'm afraid it has been a while since I used Insync because I couldn't find a fix for this type of issue. Instead I have been using which is\nI reached out to insync back when I made my last post and begged them to look into the issue. I offered to pay for a fix. I was told it was to the developer backlog. I've switched to using .\nYeah, it is frustrating that the Insync developers don't seem to be interested in providing a distro-agnostic version of their product. A flatpak or appimage distribution of insync seems like it should be an obvious win for them, but they haven't committed to anything. Here's hoping; I liked using Insync when I was able to.\nare you achieving continuous (real-time) sync with those tools? Neither of the two tools supports that by itself , . If you're achieving it by using the tool as part of a more elaborate solution, would you mind sharing how you're doing that?\nI wrote an overlay to pass by the error 403 of url down and I get it installed. But doesn't open yet The error:\nUpdate about . It worked when i pinned a rev from github. To fix the broken URL I created an overlay to override the version and URL of current version on nixpkgs.\nI am trying to fix it for the latest insync binary... But it doesn't work, it crashes, badly If I don't do these: Insync will crash with: Something about GLIBC_2.36, while nixos is still at glibc 2.35. But it shouldn't be a issue, otherwise insync would only work for distro that is glibc 2.35... Hope it gives some hints, if anyone wants to continue from here, feel free!\nI made build and insync-v3 run again, but it would need to patch two more python packages... to continued.\nUpdate from , the package should be resuscitated soon, coming to unstable and hopefully in time for 23.05!\nthank you for persevering on this and getting insync working again!\nthanks for this package! Could you also please add the .desktop file and icons? The base insync package has them but they are not to the wrapper package and are not currently user visible. Thanks again!\nAlso the insync binary is currently named insync-${version}. Can you please strip the version from the binary name? Otherwise the desktop file won't work. Thanks!\nNo problems. I will look into these!\nMight have the desktop files solved."}
{"_id":"q-en-nixpkgs-82d026a518c70f70fc61ee445ba8bffefa32d402d7db3741cde8bba09ed6a01f","text":"\"Source\" link in search result on for Poetry package is pointing to poetry2nix source code instead of poetry. Steps to reproduce the behavior: on \"Source\" link in \"poetry\" package. This will point you to instead of When clicking on \"Source\" link in \"poetry\" package search results I want to get to \"poetry\" source code If applicable, add screenshots to help explain your problem. Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result. Maintainer information:\nThis is really a bigger issue in how source code locations are extracted and it isn't unique to Poetry/Poetry2nix. The root cause is related to how is extracted: In this case is resolved to where Poetry2nix inherits descriptions from /. You can explicitly pass to which I just implemented in :\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-0d5ee661795a678c6d0e17b7053293838938353f52679efadcbd9010c9fefac0","text":"is \"no longer maintained\" or distributed, and the releases are unavailable: <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result. Maintainer information:\nA closed source substituting another one? Well, it can be fixed by aliasing them to a failure."}
{"_id":"q-en-nixpkgs-55f45d3a2c34150eba072263e3d303db6fa49fdab0c9058ddd83d398b08a9c54","text":"I would like to do this myself, but I'm getting some errors. I fixed the version fetching in , but now I am a bit stuck: No change to the actual package file is seen. I was able to work out that those errors are coming from inside but that's as far as I got. <!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on github to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: bluejeans-gui <!-- The current version can be found easily with the same process than above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 2.23.0.39 desired version: 2.24.0.89 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: (tagging because just merged my bluejeans version fetch PR ()) and (tagging because seems to have committed to the package recently) Please tag this issue in your PR.\nthis is why it fails:\nsomething like this will work:\nOh! I thought the I saw when checking was from something else. Thanks.\nThanks for picking this up!\nFirst step:\nlooks like my work paid off and the bots are taking it the rest of the way"}
{"_id":"q-en-nixpkgs-1f1eecdcce8357ae5a5f625910eb0248c793fe9b6f1c11c489e95a70c14f536b","text":"To make sure that we are on the same page: Kernel: (run ) 4.5.0 System: (NixOS: , Ubuntu/Fedora: , ...) (Flounder) Nix version: (run ) 1.11.2 Nixpkgs version: (run ) During testing of I managed to get at least two different hashes for . After some discussion with it seems that this derivation can produce non-deterministic output. Hashes of outputs are always the same. They are not. Build several times.\nHow would one work around that? It basically means the system cannot be rebuild :cry:\ncc\nYou can work around it by updating SHA checksum in the affected package locally.\n:confused: Is there any documentation you could point me too. I know how I would to it for packages that I install as a user: out nixpkgs repo the last commit that build for unstable on hydra that out the exa's file 5 run But I think it's also possible to override it in a nix expression, like the one in , since I installed it as a systems package. The thing is that all ressources I find either explain how to do it for that they link to\nI changed my to include that: but to no avail :confused:\nSadly we don't have a mechanism to override (it needs to be supported). However, you can rebuild your whole NixOS from a modified repository:\nAnother way to do it is to call patched nixpkgs from your :\nThanks for your help! I really appreciate it. I tried your last suggestions, but that leads to: It looks like some network hick up, but I tried it multiple times and I opened the github repository from the same machine and it loads fine. Maybe it has something to do with this line: ?\nHm, unfortunately I have no idea what's going on -- you've fixed the hash, so it seems like another problem in .\nStill thanks for the help. I learned a lot!\ndepends on a git repository in which doesn't work with\nmaybe fetchcargo could take an argument to pin git-based dependencies on a certain commit?\nI can confirm that this is still an issue when building : through AUR/exa-git. cc\nAre there any updates on this issue, please?\nShould be fixed."}
{"_id":"q-en-nixpkgs-e17a626b142357c820bcce448da247c0e6983027407024293d9f3076fd6e0b8f","text":"To reproduce: set to rebuild system with two users that don't have uid specified rebuild system by specifying uid of the first user to match the generated uid of the second user They will now share the uid. Expected behavior: error should be raised cc\nthis also happens if you force the uid for 1 user, and dont set the uid at all on another, i need to force the uid on every user to avoid collisions\nyour issue should be fixed in , I'm still verifying if my problem is covered :)\nThere might be some other edge cases, but now this issue yields:"}
{"_id":"q-en-nixpkgs-c1c2c0245757112c0199043946acee66eb141344ee4df372abe0c1b1b01f9b71","text":"RStudio () fails to build from nixpkgs master with a 404 on downloading sources. Steps to reproduce the behavior: <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Maintainer information:"}
{"_id":"q-en-nixpkgs-b2409eb103e9ad3260b7a2254bda021c8cc6716f12072f080948e0a33f1ac7d9","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on github to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: chirp (chirp-daily) <!-- The current version can be found easily with the same process than above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: desired version: <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: None (!) - I can probably volunteer to keep this up to date? The last PR was submitted by though. Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-bb8e95ef6b5326ecd2dd6e449b64a01f42bf7ebcbe4919b7e4741887ff49361c","text":"A URL like generates the following error log: Steps to reproduce the behavior: to download a file that has been uploaded Should download the file Are you seeing this problem on your non-beta versions of Discourse? It looks like maybe the X-Accel-Mapping isn't applying and then the request to download a file is getting fed back into Rails.\nI've confirmed this is also an issue on Discourse 2.7.5 for NixOS.\nThis can be fixed with: However, I think this is probably a bad fix as is, because you might be able to trick discourse into serving arbitrary files in the /nix store. Two possible solutions: Discourse so it serves files with paths like /run/discourse instead a much more specific location regex in nginx to mitigate the issue I noted above Do you have any other ideas?\nThis works and should be safer:\nMy hunch is that it's both safer and cleaner to patch Discourse. If I follow the code correctly, I believe this is what would need to be changed to just return : There is already a match in place in the nginx config, so nothing more should be necessary.\nThat's the same place I followed the code to, so that is probably correct. I'll look into patching it."}
{"_id":"q-en-nixpkgs-429a986f492f37fb99a6da24f4fc0ad91a4622b3d3c348ddd7e8d76c04e3cb0a","text":"When trying to build the swift package, it inevitably fails at the same point in two of my machines. In particular, it seems to fail at the following point (): Steps to reproduce the behavior: the following a long time for it to ultimately fail with the above message. I expect the swift package to successfully build. I have not tried to build it using any other method than the way outlined in <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Maintainer information:\nIt looks like swift is using a bundled llvm source tree, so we will have to apply the patch from\nOut of curiosity do you know if there is a version of nixpkgs that contains a Swift package I could possibly build? Or is it somehow dependant on how my system is configured?\nCertainly, that patch was only needed since bumping our userspace linux headers from 5.12 to 5.14, so I would expect any commit before that would work. According to Hydra, () was the most recent success."}
{"_id":"q-en-nixpkgs-3acc230d8dd4e86cca5996ae8dfae030642d0e71164a23e909685f27fd8e9b87","text":"To make sure that we are on the same page: Kernel: (run ) Linux lapole 3.18.25 -NixOS SMP Thu Jan 1 00:00:01 UTC 1970 x8664 GNU/Linux System: (NixOS: , Ubuntu/Fedora: , ...) 15.09.952. (Dingo) Nix version: (run ) nix-env (Nix) 1.10 Nixpkgs version: (run ) 16.03pre75806. I installed zsh-5.2 from nixpkgs channel to get away from a 5.0.8 bug. After I changed my login-shell to 5.2, ssh logins were no longer functional. Ie. PATH wasn't fully initialized. Reverting to 5.0.8 fixed the problem. I'm not sure why there's a difference between regular login shells and ssh though. Running yields: For 5.0.8 I get the expected /nix/store/... path instead of the strange $out path. The line with $out seems wrong.. Checking the commit log I find Dropping into nix-shell for zsh-5.2, setting and manually running I get: So it seems to me that variables in is not expanded leading to zprofile not being run. It could be something else of course (not everything seems to add up without more investigating), but the $out thing seems wrong either way.\ncc\nThis is fixed by the merge of . Cherry-picked to ."}
{"_id":"q-en-nixpkgs-ea3a7bbc675952cf59a57738ae67f04caa5f9a0c2274eaeaf4bac31d950b6f1a","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on github to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: broot <!-- The current version can be found easily with the same process than above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 1.6.2 desired version: 1.6.6 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-9f75da30ae160ccbf0439ef9e32f998b229f2540022b58d71f27299abd388ef0","text":"My hydra instance started to report hash mismatches in rust packages (agate, ion, nushell, ox, ripgrep) (I'm cross-compiling to aarch64-multiplatform). It appears that the commit introduced this. See also the associated PR: Steps to reproduce the behavior: nixpkgs @ (just building the attribute would be enough) all deps... mismatch for cargo dependencies. No hash mismatch. I don't understand why the linked change would cause this. Maybe you know why this happens? Please run and paste the result. Maintainer information:\nHuh, this is really interesting, I'm going to revert that change for now.\nReverting has fixed this. Hopefully the right solution can be found :).\nWould you have time to test a re-do of the original PR? I think I know how to get it to work correctly."}
{"_id":"q-en-nixpkgs-f1b8b1a2e652726477dada066c66ea30c7f60a9b9a8643153134b12ebfad61c4","text":"= 2021.07 does not build successfully on Darwin. A bisect over commits shows this: 2021.10 - - does not work 2021.07 - - does not work 2021.06 - - works 2020.12 - - works Steps to reproduce the behavior: this repo: on Darwin. (I'm on MacOS Big Sur). line 4 from to . . Observe the build failure in the output: For the full log, see .\nCaused by Please use to show the files that a working rakudo contains on darwin\nTested on and it works. Thanks!"}
{"_id":"q-en-nixpkgs-88a9baed4439798cedd8d589b9441de4a424f1acef37e266fbfacf5dfb5502fe","text":"Building tree-sitter grammars on MacOS results in a fatal error with the post-installation fixup stage due to differences in how is invoked on MacOS. The error reported is: Steps to reproduce the behavior: or with any other tree-sitter grammar. the fatal error in the post-installation fixup phase where it reports that no files are specified for . The grammar should build and execute the fixup phase without errors. This error does not occur on non-MacOS platforms. I reproduced the error on MacOS 12.0.1. I suspect this error was introduced with this commit but I'm not 100% sure.\nSome discussion:\nI can also replicate this with on macOS 12.0.1."}
{"_id":"q-en-nixpkgs-555ed3c35a275fe5145c8e4df08a7838915d92b403c5c7e9d6488221a33c14a6","text":"When is installed with , works fine. When is installed with , then prints an error message. When installing with , I get this log: So, it might be that git-with-svn-2.6.3 is fine but not . System: NixOS: = () Nix: = Nixpkgs version: () /cc\nI think you have nixpkgs and nixos channel. One might be coming from root. So on one channel it works, on the other it doesn't.\nThis is a bug in the Perl bindings generated as a part of our Subversion build. The library depends on , but it has no RPATH set that allows it to find library. Curiously enough, is the only dependency with that problem: other library like are found properly using a full path into the Nix store. Cc: and\nand it was working fine in :\nAs I said, this as nothing to do with Git. It's that provides broken Perl bindings.\nRan into same problem. git bisect log of checking ldd /nix/store/-subversion-1.9.3/lib/perl5/siteperl//x8664-linux-thread- follows. there's some big binutils related changes in the middle yet."}
{"_id":"q-en-nixpkgs-ab5b0c999c894e745efd04bc6e204c38fac615991cfc44df4cc4fd4a2c679d3e","text":"The Haskell package build fails due to missing C library. I assume it just needs a dependency on . The describes how to add non-Haskell dependencies to a Stack build, but what if I'm not using Stack? The build fails, outputting:"}
{"_id":"q-en-nixpkgs-4ef781a2cdd1b0fd66bf8920c0ad550e71cdb7dd64441f08082f3d35a5d73beb","text":"MySQL python binding is broken in 16.03 channel. Is seems to be OK in master. Apparently, is the commit in master that fixed that issue. If it is OK, I’d like to cherrry pick that commit into the branch (this will imply a massive re-downloads from the cache for the users). System: Nix version: Nixpkgs version: cc\nany objections for a full rebuild of 16.03? (Maybe there's a workaround for this bug)\nThe error in seems related to this.\nCherry-picked to 16.03"}
{"_id":"q-en-nixpkgs-a4df8bd7a3b26553227e58572ea86506f566d6d640cdaea47e177240fdf64804","text":"The derivation of creates a different .desktop file from the \"official\" one that has a different name. This is probably done in order to differentiate the two versions of in case someone installs both of them, but the problem is that while it creates a new one, it doesn't delete the \"official\" one. Also, the new .desktop file should also set , otherwise games don't launch in Wayland. (Though maybe it should be a wrapper instead) (And, unrelated, but why and not ?) The same also applies to PATH: there are both and (which is a link made by the derivation to in the folder). For consistency, there should be made a , otherwise if both dolphins are installed, one overwrites the other in PATH. Steps to reproduce the behavior: dolphinEmuMaster; an app launcher; are both the original \"Dolphin Emulator\" and the new \"Dolphin Emulator (master)\" that open the same app (the master one doesn't set `QTQPAPLATFORM=xcb). PATH there are both and (which is a link). There should only be the \"Dolphin Emulator (master)\" desktop file, and in PATH there should be only (and instead of ). <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --"}
{"_id":"q-en-nixpkgs-40761a4870190e8740ac145924076ff1393a138a2c11a9cebeb54f9e2104c9fd","text":"After enable HBase module , and rebuild. The service unit cannot start, it says JVM option not exists. Steps to reproduce the behavior: HBase by adding config . , it should be failed. shows VM option is Unrecognized. I found the same issues on the StackOverFlow, it seems jre 15 dropped the option: I fixes this by adding this line, but it should be fixed in module. I can do the fixes PR if you want. Maintainer information:\ngo for it. I'm also not maintaining this anymore, so may be best to also remove from ."}
{"_id":"q-en-nixpkgs-37469eae0b0b5ffaff1e291094fb5f7c6f619d7879f0f6e01f7babbba61cdc04","text":"Found in It does build on and . Refs ZHF <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result. Maintainer information:\nIs there a general fix for the error? I'm seeing this with (this issue was the only one in nixpkgs that matched that string, sorry if I'm completely off base, I just have no idea where to go on this)"}
{"_id":"q-en-nixpkgs-1ef2aacf1286226aee0f163fc81c9bdf47312c059399144d430f26c9a3b11753","text":"Broken package: 'groove' Install package. Run 'groove-simulator'. File -New Grammar... Try to save, get an exception. system: host os: multi-user?: sandbox: version: channels(maestro): channels(root): nixpkgs: Permalink to broken package: It seems that the broken Nix package depends on OpenJDK17. Found that running their latest version (5.8.1) from SourceForge with the much older OpenJDK8 works just fine. Did not test with OpenJDK11. Probably just have to depend on an older Java version. Package version can also be bumped to 5.8.1. Don't have sufficient knowledge of Nix to fix this myself. Thanks\nProbably should be assigned to\nI think I've fixed it. I'll submit a PR soon. Also, I tested OpenJDK11 and it crashed there as well."}
{"_id":"q-en-nixpkgs-440499c1694f2aad664bd2b4095ab82ab2169806fe1dba1c8607176a082b4d74","text":"Project description Pipeable Ruby - forget about grep / sed / awk / wc ... use pure, readable Ruby Metadata homepage URL: source URL: license: mit platforms: ruby\nLooks like you got this in"}
{"_id":"q-en-nixpkgs-fc8877bfab441980fe5ba517a43f261574ba85300666cd31e5aac717031d75a7","text":"Project description Teapot is an interesting project that implements a compact spreadsheet in roughly 140k, with the following features: UI: Ncurses terminal UI, FLTK windowed GUI, and a batch processing mode all available. Cell Syntax: Cell locations represented as lists of coordinates. Eg, instead of \"B1\" or \"R1C2\", cells are referenced as \" Three Dimensions: This 3-coordinate style adds a third dimension to the spreadsheet, reducing the need for hiding columns or rows of intermediate results as is sometimes the case in traditional spreadsheets, among other things. Relative References: This also enables both relative cell and relative formula references that make refactoring the spreadsheet easier. Type-Safety: All values have a type, and all operations and formulas check types. No more adding a float to a string and getting a (probably undesirable) result. Iterative Expressions: Modeled on clocked expressions, these solve problems with circular dependencies in traditional spreadsheets and effectively make teapot Turing-complete. Platform-Independent File format: Defined by XDR and portable across machines, especially floating point representation that may differ across platforms Legacy File format support: Portable ASCII file format, standard CSV, SpreadsheetCalculator SC and Lotus 1-2-3 WK1 format. Extensibility: Easily extended by new functions in C. Internationalization: Can be internationalized and comes with English, Dutch, German and Polish message catalogs. Interoperability: Interfaces to troff/tbl, LATEX, HTML and CSV which are well-suited for further processing, be it graphing, typesetting, accounting or anything else. Metadata homepage URL: source URL: (bottom of page) license: GPL3 platforms: unix, linux, darwin\nI tried to build this from source, but I'm getting an issue with it. Here is what I did so far, in case someone wants to try it: I am going to try with the deb and see if it works.\nI got the deb working successfully (even created a desktop entry for the GUI client, although no icon). I'm going to keep trying with the source today (it is an issue regarding mbslen not being recognized), but if I can't get that, then I will post the working one.\nis just a warning, it can be fixed with a simple patch. The more serious problems are compiling with FLTK support and build the docs. Maybe more patches are needed. (On the other hand, it can be useful to package the already-given binary targz.)"}
{"_id":"q-en-nixpkgs-95b2ceb2c0682688c4a18e18702267ffb13c95ca305e0ebee9a1866afc67892b","text":"The service seems to be executed before the module is loaded, causing automatic modeswitching not to be enabled. The following is seen in the : Restarting after login resolves the problem, as does adding the following to the guest's configuration: Steps to reproduce the behavior: a nix guest with X11 enabled in vmware fusion windowed mode, resize window by dragging appears stretched/blurry or centered in window (dependent on host configuration). Display resolution is automatically switched to match dimensions of window. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Maintainer information:\nHey, I'm no longer using VMware with NixOS, so would be great to find new maintainers for the vmware-guest and open-vm-tools. you've worked on vmware support recently. Would any of you be able and interested in taking over?"}
{"_id":"q-en-nixpkgs-9637787a60f7a55e0ac047ba04527aceb4ae9cfd1d613ffa93513fe4d58ed934","text":"shairport-sync module creates the user without a default group, running into enabling causes the following error:"}
{"_id":"q-en-nixpkgs-cc9318edc0dbadfffdc8652c6b4824d98e08e273bf28c1cb8a6861894da19b74","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on github to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: dstp <!-- The current version can be found easily with the same process than above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 0.3.0 desired version: 0.4.0 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: I'd love to update the package myself without annoying you :) If there is a documentation about how to upgrade packages, i'd love to learn it and try my best next time, thanks in advance!\nFor simple version updates, I update the version number and to get the new hash I run (in the nixpkgs top directory): For go packages, we also have to update the hash, that one is tricky, and what I do is to try to build the package. That will fail, but it gives us the new hash, I don't know a better way. There is also a bot () creating PR for those simple version updates. It uses repology to find the latest version:\nAwesome, thanks!"}
{"_id":"q-en-nixpkgs-868fa31c6bb92517d56780f4332126348bf6af3b1cf9d3a86878fe0e6f65bf70","text":"Project description Tile World 2 Metadata homepage URL: source URL: license: gpl2+ platforms: linux"}
{"_id":"q-en-nixpkgs-9ca255853cef7edbedfe55b69d6d5730b381c1f1f733b0236cde667e02fb557e","text":"Project description Graphical application for generating different color variations of a Materia and Oomox themes (GTK2, GTK3 and others), Archdroid, Gnome-Colors, Numix, Papirus, Suru++ icons, and terminal palette. Metadata homepage URL: source URL: license: gpl3 platforms: linux"}
{"_id":"q-en-nixpkgs-207755b239efdfee87ebec596ce81c5bcb8ed336abde22813f5855885fd5d574","text":"Project description Cubiomes Viewer provides a graphical interface for the efficient and flexible seed-finding utilities provided by cubiomes and a map viewer for the Minecraft biomes and structure generation. The tool is designed for high performance and supports Minecraft Java Edition main releases 1.0 - 1.18. Metadata what does this mean? homepage URL: source URL: license: GPL-3.0 platforms: Linux"}
{"_id":"q-en-nixpkgs-f132d9e3cfc0898ab3350446733feb1997e10a6c81985e97e389e55761828c99","text":"Since won't build () From the above logs: Steps to reproduce the behavior: <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nCan reproduce on my ZFS system, not on my ext4 system though.\nI can't reproduce by running on any system.\nI was able to build from that commit nix store is zfs backed\nIt's more relevant whether is.\nThe renameat2 syscall used isn't (fully?) supported by ZFS: I'm not sure how that relates to this issue only occurring with , I guess it's related to the sandbox.\n/tmp is tmpfs for me\nThanks for filing the upstream issue!"}
{"_id":"q-en-nixpkgs-d3e9d2ed2b466c1dd19fa179cc701cbb37ae34fe8e797102043c62fe9b406d1e","text":"Project description Junction lets you choose the application to open files and links. Junction will pop up automatically when you open a link in a desktop application. Use the mouse or keyboard navigation to choose the application to open the link or file with. Metadata homepage URL: source URL: license: gpl3 platforms: linux"}
{"_id":"q-en-nixpkgs-ff93313c40e8241a976049652c1983ae4edac80f5ea043ac946c9d0545b1c5fc","text":"Recently variety stops working on NixOS. Pictures are downloaded well but the wallpaper can't be changed and can only be set in system setting. That's because of the hardcoded shebang which is taken from on the first installation. So when the bash is updated the script can't find it. It should be set as something like . Not sure if this is a nixos bug. Steps to reproduce the behavior: variety bash It works after the bash update. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nIIRC now we use . I am patching it now."}
{"_id":"q-en-nixpkgs-2a691bdb43d1f69ead1c99f6c1593b702b91e030a544887a0c19da4b54a80966","text":"No GNOME Shell Extension can enable desktop icons. . to enable desktop icons. Should be able to enable desktop icons in GNOME Shell, one way or another. About the time of NixOS 20.03, i was still able to have desktop icons using some old version of extension. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nI can confirm this. The systemd journal shows (once every second):\nThis probably requires patching all paths referencing other libraries or binaries. Which of those extensions shall I have a look at first?\n, maybe the one that is easier to fix? I did not have an opportunity to test , so i cannot decide between the two. Thanks!\nHaven't had the opportunity to check out any of them so I also cannot decide. Any one of them would be fine for me, so I guess start with the one that is easier to fix (if there is a difference). If I can be of any help let me know.\nWIP: I have the following override and the following patch: But I somehow can't get it to find nautilus. It will always end with a null exception while trying to access nautilus stuff (can't copy the stack trace out of the testing VM, sorry).\nstill not solved :\nMade a PR for , a similar technique can be used on the other desktop icon extensions. Would like if someone could test that everything's working. Note the schema path bit, that was causing . For any path you cannot directly list with you most likely need to do the thing (or otherwise find a way to make it find the schema source)."}
{"_id":"q-en-nixpkgs-4df336f9ca1fa2bc2a48cfb498bd236f1f4e3ea87fe2e7bdc253ee0177a6a0d7","text":"If I only do: then I get virtualbox + nvidia: If I enable virtualbox guest additions: then I get only virtualbox!!! and nvidia does not work anymore on my machine: this is because virtualbox guest additions is overriding services.xserver.videoDrivers: this is not necessary since by default the lists are merged workaround: then I get virtualbox + guest additions + nvidia: Ideally this should work out-of-the box, debugging this kind of behavior is very obscure, and also different configuration options should cohexist"}
{"_id":"q-en-nixpkgs-88263ceb2da15c5f1e555a3155a6e7d61332ad3ae280aeb9610d55f96040248e","text":"vulkan-tools fails with an assertion error from vulkan-headers Steps to reproduce the behavior: the nixos-unstable-small channel vulkan-tools being installed # nix-shell -p vulkan-tools --show-trace error: assertion '(version == (vulkan-headers).version)' failed at /nix/store/cqbcz8k1wq604f8b8nc8khamyh2cl6dl-nixos-22.05pre348006./nixos/pkgs/tools/graphics/vulkan- 10| # the rest of nixpkgs. 11| src = (assert version == vulkan-headers.version; | ^ 12| fetchFromGitHub { … while evaluating the attribute 'src' of the derivation 'vulkan-tools-1.2.189.1' at /nix/store/cqbcz8k1wq604f8b8nc8khamyh2cl6dl-nixos-22.05pre348006./nixos/pkgs/stdenv/generic/make- 204 (attrs ? pname && attrs ? version)) { 205| name = | ^ 206| let … while evaluating the attribute 'buildInputs' of the derivation 'shell' at /nix/store/cqbcz8k1wq604f8b8nc8khamyh2cl6dl-nixos-22.05pre348006./nixos/pkgs/stdenv/generic/make- 204 (attrs ? pname && attrs ? version)) { 205| name = | ^ 206| let\nFix at , thanks for the report. I thought we had tests for this stuff...\nfixed it for me. Thanks!"}
{"_id":"q-en-nixpkgs-3d5b17446b3f0669b0b0eed2db3c0897a9c3ebde568d5ca94ac243f85d3b32f4","text":"Trying to install tetrio fails (see screenshots) Steps to reproduce the behavior: tetrio-desktop from nixpkgs-unstable (I did it using a flake) The package should install with no errors. ! <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nOops, it looks like I accidentally used a non-direct download link for the source. It also looks like audio is broken, so I'll start a PR soon."}
{"_id":"q-en-nixpkgs-e7db5ff2b76c3ad1ed8a839168b21ddfe967c64e71d7e02179d3470421dc3e2d","text":"As of , the build of is broken. This is currently breaking the build of other packages including and certainly many others... ~I suspect this was triggered by , , or one of the redis upgrades.~ I believe this was triggered by . Steps to reproduce the behavior: nixpkgs at failure The build to succeed. n/a n/a <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nI tried patching around the issue with but that gives me a test failure:\nshould fix that. Also please use postPatch."}
{"_id":"q-en-nixpkgs-0edec83b6ade513a5cfc78972086224790d0580882f49d32b976bedfb907ba81","text":"Project description Fprintd replacement which allows you to have your own backend as a standalone service. Metadata homepage URL: source URL: license: GPL2 platforms: linux"}
{"_id":"q-en-nixpkgs-95c43717e355b49bdbcc0ca7a5ab7d68235315a48d7e26b6b01c3d46127a37a7","text":"nix-shell -p supercollider [anders $ nixos-version (Flounder) [anders $ nix-env --version nix-env (Nix) 1.11.2 [anders $ nix-instantiate --eval '<nixpkgs' -A lib.nixpkgsVersion \"\"\nyou use supercollider ?\nno, sorry.\nSame here:"}
{"_id":"q-en-nixpkgs-5bb05e1586298b8cc3cbfbfa520f53ac36239e72c0b335126edbbe2523a823a7","text":"When using with the I get the following error. ! I'm assuming this is caused by the following code. Since it doesn't include the module. Steps to reproduce the behavior: the to open Libreoffice Writer The Languagetool plugin doesn't cause a runtime error. Please run and paste the result.\nHm, jre_minimal tests seem to confirm it; does adding \"java.logging\" there help?\nHappy to take a look by the end of the week :)\nThat'd be great if you could do that. I tried building it on my laptop. But it was still building after ~3 hours. :)"}
{"_id":"q-en-nixpkgs-6ce3edd4d9a9944a8d55665d2c12adb574f72d247400b588e30c4c65296d27b2","text":", CVSSv3=5.5 (nixos-21.11, nixos-unstable) CVSSv3=4.9 (nixos-21.11, nixos-unstable) CVSSv3=4.9 (nixos-21.11, nixos-unstable) CVSSv3=4.9 (nixos-21.11, nixos-unstable) CVSSv3=4.9 (nixos-21.11, nixos-unstable) CVSSv3=4.3 (nixos-21.11, nixos-unstable) Vulnerability in the MySQL Server product of Oracle MySQL (component: Server: Compiling). Supported versions that are affected are 5.7.36 and prior and 8.0.27 and prior. Easily exploitable vulnerability allows high privileged attacker with network access via multiple protocols to compromise MySQL Server. Successful attacks of this vulnerability can result in unauthorized ability to cause a hang or frequently repeatable crash (complete DOS) of MySQL Server as well as unauthorized update, insert or delete access to some of MySQL Server accessible data. CVSS 3.1 Base Score 5.5 (Integrity and Availability impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:N/I:L/A:H). Vulnerability in the MySQL Server product of Oracle MySQL (component: Server: Federated). Supported versions that are affected are 5.7.36 and prior and 8.0.27 and prior. Easily exploitable vulnerability allows high privileged attacker with network access via multiple protocols to compromise MySQL Server. Successful attacks of this vulnerability can result in unauthorized ability to cause a hang or frequently repeatable crash (complete DOS) of MySQL Server. CVSS 3.1 Base Score 4.9 (Availability impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:N/I:N/A:H). Vulnerability in the MySQL Server product of Oracle MySQL (component: Server: Stored Procedure). Supported versions that are affected are 5.7.36 and prior and 8.0.27 and prior. Easily exploitable vulnerability allows high privileged attacker with network access via multiple protocols to compromise MySQL Server. Successful attacks of this vulnerability can result in unauthorized ability to cause a hang or frequently repeatable crash (complete DOS) of MySQL Server. CVSS 3.1 Base Score 4.9 (Availability impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:N/I:N/A:H). Vulnerability in the MySQL Server product of Oracle MySQL (component: Server: Parser). Supported versions that are affected are 5.7.36 and prior and 8.0.27 and prior. Easily exploitable vulnerability allows high privileged attacker with network access via multiple protocols to compromise MySQL Server. Successful attacks of this vulnerability can result in unauthorized ability to cause a hang or frequently repeatable crash (complete DOS) of MySQL Server. CVSS 3.1 Base Score 4.9 (Availability impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:N/I:N/A:H). Vulnerability in the MySQL Server product of Oracle MySQL (component: Server: Replication). Supported versions that are affected are 5.7.36 and prior and 8.0.27 and prior. Easily exploitable vulnerability allows high privileged attacker with network access via multiple protocols to compromise MySQL Server. Successful attacks of this vulnerability can result in unauthorized ability to cause a hang or frequently repeatable crash (complete DOS) of MySQL Server. CVSS 3.1 Base Score 4.9 (Availability impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:N/I:N/A:H). Vulnerability in the MySQL Server product of Oracle MySQL (component: Server: Security: Privileges). Supported versions that are affected are 5.7.36 and prior and 8.0.27 and prior. Easily exploitable vulnerability allows low privileged attacker with network access via multiple protocols to compromise MySQL Server. Successful attacks of this vulnerability can result in unauthorized update, insert or delete access to some of MySQL Server accessible data. CVSS 3.1 Base Score 4.3 (Integrity impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N). Scanned versions: nixos-21.11: ; nixos-unstable: .\nSee also: , , , , , ,\nFixed: nixos-21.11: mysql-5.7.37, mysql-8.0.28 nixos-unstable: mysql-5.7.39, mysql-8.0.30"}
{"_id":"q-en-nixpkgs-72ad20613edf4d55cb28ac482ebb640df89ab3cf116e9a6445cc696878214021","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 4.1.6 desired version: 4.1.7 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-45ccc65dd67c3e9e6ae78e353afa2441c13233f30da64954c21c42dd0cfd9bfc","text":"After Wireplumber was enabled by default in , audio and video playback completely broke on my machine (NixOS-unstable, GNOME on X11, x86_64-linux). Preview thumbnails of videos did show, but clicking \"play\" on audio/video tracks does nothing. I've observed this directly in Firefox, Element Desktop, Freetube, Rhythmbox and Signal. Reverting by putting into my config resolves the issue. Steps to reproduce the behavior: Firefox to play a video since you authored the PR Please run and paste the result.\nI've also verified that this isn't resolved by .\nI am also experiencing issues with video playback, I am glad i found this issue. seems to have fixed it\nKnown regression with pipewire 0.3.48 and wireplumber 0.4.8. Should be fixed by .\nI've tried to apply the fix from with: but it didn't resolve the issue for me.\nThat is strange. What's the store hash of your wireplumber build after this?\nWith the nixpkgs revision , I get: I haven't applied any overlays that influence wireplumber. (Checked by directly building wireplumber from your Nixpkgs PR branch)\nThat hash does seem to match. I wonder if it's a different issue then... Can you also run ?\nBy the way, if you want to, we can continue this on Matrix (I'm there).\nHonestly at this point I'd probably just report a bug upstream, or ask around in\nOkay, thanks, I'll try that. Do you happen to know whether wireplumber has some sort of \"verbose/debug log\" build option which might make it easier to find the bug?\nTry WIREPLUMBER_DEBUG=10\nYou should try zapping all config and cache. Especially if you use bluetooth (in which case the bluez cache).\nI've been having this issue, too, on unstable. What I've noticed is that, on reboot or logout/login, playing a video works just fine the first time. But on the second and subsequent times, video won't play. A stil image from the video is displayed, but the video doesn't play at all. Here's what I see with : For me, running fixes this problem immediately. That is, videos suddenly start playing.\nIs this with the patch or without?\nI also wanted to note that I've reported the issue upstream now (), maybe a few more log dumps from other people with this issue will help the developers over there\nthere might be relevant details here\nHow are you enabling Pipewire? Something like the snippet below?\nThat's an interesting question. I didn't explicitly enable Pipewire anythere, my sound config is just: Basically, I have no idea of audio systems on Linux and I just assumed that it was some sort of backend for PulseAudio. On my system however, the Pipewire and Wireplumber services are there. Maybe that's some sort of bug in the NixOS audio submodules?\nWith , I've confirmed that something in my config indeed causes to be set to true, but are all false.\nI've backtraced this and is set by , which is in turn enabled by default on Gnome (nixos/modules/services/x11/desktop-). It might be worth to open an issue for that.\nThis is really useful debugging information! Take my next paragraph with a grain of salt, since I am not a PipeWire expert. Some parts of PipeWire relate to screensharing (which is probably why enables it) and supposedly should work even without the ALSA/PulseAudio/Jack emulation, but sometimes there are bugs in the session managers; for example, was a bug in the old session manager with a similar outcome. I think that screensharing use cases are meant to work even for PulseAudio users (that is, without enabling PipeWire's PulseAudio \"emulation\" layer).\nIt seems that part of the problem is that had those extra knobs for turning various emulation layers on and off (such as and ), but those are no longer respected when using WirePlumber. This might lead to some weird conflicts for PulseAudio users (for example, if WirePlumber decides to start pipewire-pulse).\nBut does this mean that the WirePlumber switch broke audio for every GNOME user that does not use PipeWire? This seems like the default config!\nThose knobs are for Pipewire itself, not pipewire-media-session. Wireplumber should still be functional in a mixed PW/Pulse setup, so it's probably worth reopening the upstream issue.\nI'd ask someone else to open that issue though because my knowledge on this topic is close to zero, and I'll probably not be able to accurately describe the technical details here. I've also tested this on my laptop, and it has the same issue, so that actually seems like a realistic possibility. I'm not really sure when it comes to user statistics, but from what I've seen, GNOME tends to be mostly used by less experienced users, while KDE and WMs like i3 are a lot more popular among advanced users, and the unstable repositories of an operating system are generally almost exclusively used by advanced users. So after considering the small user base of NixOS, it's really possible that there are just not that many GNOME/nixos-unstable users who are likely to encounter that bug.\nIdeally yes, but there are two places in the NixOS module that look at : uses it to decide if the systemd service should be installed, and looks at it to create the file. Therefore, is given a hint that pipewire-pulse shouldn't be started, while upstream WirePlumber does not seem to have such an option. If I had to bet, I'd say WirePlumber does not currently support this use case.\npipewire-media-session doesn't start pipewire-pulse, systemd does. Wireplumber should detect the right setup automatically.\nYeah the file dates back to before the systemd service for pipewire-pulse existed. It is probably redundant nowadays.\nI agree with what you two are saying, but I think I didn't express myself clearly, so I will try again; apologies if this is annoying. My point is that we have no indication that WirePlumber is designed to support mixed setups. The release notes for WirePlumber 0.4.0 say \"It is a fundamental goal of this series to maintain compatibility with pipewire-media-session, making WirePlumber suitable for a desktop PulseAudio & JACK replacement setup, while supporting other setups as well (ex. automotive) by making use of its brand new Lua scripting engine, which allows making customizations easily.\" and it's unclear if they care about our use case (IMO it would be perfectly reasonable to not support PulseAudio). I'll ask upstream.\nPipewire and Pulseaudio are (should be) mutually exclusive. They both try to multiplex the underlying hardware, so if you're running both of them, that may or may not work depending on whether the physical sound hardware is capable of handling multiple streams. Some is, some isn't. It's not as common to have that functionality as it used to be, since modern CPUs are more than capable of handling it in software. How it should work is that: If is set, that brings up the default audio manager. Whatever that is. Pipewire would be a good choice. If is also set, that enables pulseaudio and disables pipewire. If both pulseaudio and pipewire are explicitly enabled, that should hit an assert. There are very few cases in which enabling both makes any sense. Pipewire has a pulseaudio compat layer; this is not the same thing as running pulseaudio itself. But I haven't checked the actual behaviour of nixos, so I have no idea what's happening. Try ; this should tell us what is running on your system.\nThat is not true. Pipewire can be used for video and audio, and running Pulse for audio and Pipewire for video is a supported configuration.\nIs that in fact how we're using it? Or is it opening the sound devices regardless? ...I should just check. Will report back later.\nWith Pulseaudio enabled (by ), Pipewire started when I opened a video on youtube, and no sooner — there was, however, no sound. Chrome connected to pulseaudio regardless. Pavucontrol reported no sound output devices available; only sound input was possible. A bit of searching in /proc indicated that pipewire was holding the audio devices open, and my sound card indeed only supports one stream at a time... but I'm unable to confirm that this is the issue. Regardless, I was only able to make sound work by either completely disabling pipewire, or completely disabling pulseaudio. I concur the Pipewire should be usable without making it grab the sound devices. However, how does one tell it to avoid that?\nHmm without looking deeply into it, I'm guessing the problem is that wireplumber and we don't have any way of configuring that via nix yet. Can you try if it works with pipewire-media-session if you disable all the audio stuff? This should cause alsa-monitor to never get loaded in pipewire:\nThat works flawlessly, as mentioned in my initial message in this thread.\nRight, so this is the point where we have to make wireplumber configurable via nix. There has been towards that but the lua system maps rather badly to nix so nothing in nixpkgs yet. As an interim solution I think we should add an assert that tells people to manually switch back to media-session if they have both wireplumber and pulseaudio enabled?\nFixed problem for me\nI must say that this problem has been driving me crazy and you guys just prevented hours of debugging. I will try your fix and report any problems.\nOK, I think we need to add an assertion here. I'll do it tomorrow unless someone snipes me.\nSo, while switching to Pipewire is nice, I really don't think that justifies breaking people's mixed PW/Pulse setups. Please test the fix in if you're affected (and make sure you disable Pipewire beforehand).\nwhere do you put this?\nPlease don't bump one-year-old threads. File a new issue.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-090aa222ce4d45f922f725e2ff4ea9d52c8ef668ab8a1f48e975cddec8e43eef","text":"I don't use at all, but I'm getting this error when I'm trying to rebuild my flakes (same happens when I run ) and also I don't really use it or install it Steps to reproduce the behavior: or Expect or to work fine I ran before running <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nBy grepping through nixpkgs, depends on by default. (see ). However, abook is only lists linux as supported (Im not sure if this this is just nixpkgs or abook in general) Up until , lbdb has not depended on abook by default. Perhaps a mistake was made while refactoring. In the meantime, you can just override the package."}
{"_id":"q-en-nixpkgs-bb8ff13fbf61b073dbbe04d13bab9999815462c22045f0a9747a16f28483609a","text":"are you on top of this?\nThanks for the heads up, yeah, just waited for the GitHub releases to appear. I’ll create a PR soon."}
{"_id":"q-en-nixpkgs-087bb5bb2c54a2ad330beb990fdb8a9e240d0e958e409b797ca051e469da2f60","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: lapce <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 0.0.10 desired version: 0.0.12 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Maybe it will fix the crash:\ncan you create a pull request?\nI am using changes in local build. it is working fine"}
{"_id":"q-en-nixpkgs-f5509acc8008d2a00495c3295ff82c2bdbbda43ad10b4d651450db18c73cd83f","text":"After the update, with an seemingly related build error on . This blocks the channel at the moment. Possibly related: . <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --cc\nNote: this affects both unstable and 21.11.\nIs there any plan to make a different branch? unstable has been blocked for every platform for half a month because of darwin issues (after just a few days of being working after another blockage of more than a week also because of darwin).\nFrom the perspective of a dual darwin/linux user, the issue seems to be that insufficiently tested PRs keep getting merged to master, which repeatedly regress behavior on darwin and break unstable. I'm not sure how downgrading darwin to not even deserve a place on unstable will do anything to alleviate the deeper issue that the PRs aren't being tested enough in the first place.\nIf you use a Linux, why don't you use NixOS channels instead? Even if you don't use a NixOS, the packages get extra testing through the NixOS tests.\nFrom the perspective of a person without any Apple devices, testing darwin stuff is quite hard, even though builds work nowadays, at least for small rebuilds. Some people get even annoyed by having to deal with darwin-specific issues, and I have sympathy for that. The zlib update in particular fixes quite a severe security problem. I certainly would not want to delay that on Linux (or even on Darwins) because of this qtbase regression.\nThat’s fair of course - the status quo has lots of underlying factors! I know it’s outside the scope of this issue, but should we all think about how to use some of the recently Mac build machines to, say, be able to trigger something like a more full nixpkgs-review on packages that could cascade a bunch of failures like this? Barring that, I know that the Darwin maintainers group has grown a lot. If we could find a way to get looped in on some of these changes before merge, that might help avoid the breakage churn. There’s gotta be a way to help Linux-only maintainers and committers prevent regressions more proactively, rather than frantically fixing things after the fact. Sorry I don’t have an obvious solution!\nWell, it depends. This regression in particular was known and reported (to a thread with quite many participants) before merging to master.\nIs it possible to use nixos-unstable with home-manager in a non nixOS distro? If it is, I had no idea.\nI don't think there can be anything preventing it. Though I don't really know home-manager in particular.\nI use (along with ) on macOS, so presumably yes. The branch isn’t NixOS-specific. It’s just what criteria are used to advance the branch (so sometimes things are broken on Darwin IME).\nNon-NixOS Linuxes use the same binaries as NixOS (contrary to darwin). And you always have the whole binary cache available, not just what's associated with the particular channel.\nThanks a lot, I'm changing to that branch right now, you made my day.\nFwiw, for Darwin in particular, nixpkgs-unstable does run some tests that nixos-unstable doesn't.\nYes, NixOS channels have no dependency on darwin at all.\nI did the change and it's working nice, but some packages are not getting downloaded from the cache (till now zathura and llpp) and painfully they also fail while compiling locally no idea why, any pointer to what can I check??. I know this is not the place to ask this kind of questions but I don't know which is the correct one either.\nThere's no guarantee that everything in nixpkgs is working all the time, for zathura see ."}
{"_id":"q-en-nixpkgs-64480cfcff75f24a72da7208d34e7581e533d09acdcaab0ae0c324017e5fc370","text":"Steps to reproduce the behavior: Please run and paste the result.\nThank you for reporting this issue; I can reproduce it. The problem is not in the warning displayed in the last 10 log lines, but in an earlier error: Unfortunately I don't think that I will be able to contribute further to this issue.\nI'm aware of this failure, but fixing it will require a bit of patching. Upstream didn't update the package to ghc 9 yet. Right now I don't have much time to do that, so in the meantime you could install it from stable, where it does build. Please change the title to something like \"jl: build failure due to ghc 9\"\nThis has been correctly marked broken in . See also where there was a small amount of discussion about marking this broken.\nI've created an upstream .\nThe PR is now merged, and the package is now on Hackage for the first time. However, since it's not been to Stackage Nightly yet, it won't make its way into for quite some time. I'll update the derivation to use the new version directly from Hackage."}
{"_id":"q-en-nixpkgs-37afe9f1598f6cfa173e6bc708a2da9cf073e35912bbb2f618b338216fed9aaa","text":"While checking why we have two subversion releases I noticed that both releases are LTS versions. They receive 4 years of support and 1.10 was first released on 2018-04-13, so it's probably wise to remove it before the 22.05 release. Originally posted by in\nI don't see anything that depends on in nixpkgs (anymore), so IMO we can just remove it now, instead of waiting for some kind of EOL announcement."}
{"_id":"q-en-nixpkgs-88503907ec7575a978a98d47e39860c9d0fcdccc4f2811ec5b558c10be23edeb","text":"Try to build for . Note, I'm trying staging because is necessary. fails: staging is at :\nThis was caused by setuptools 60, which now uses bundled distutils. All Python packages with extension modules fail to cross-compile currently, but this will be fixed by . I also noticed that the libbtrfsutil Python bindings are never actually installed, so I also created ."}
{"_id":"q-en-nixpkgs-2c5a2b91eada13488c51abe6ace656e8b588db8dbf2498836eb6c1a5e379da82","text":"For example: fails similarly. The line that causes the failure according to my testing is: A list of some packages that didn't break: Could be this is an upstream issue.\nSo digging into it, this is an issue raised by Octave's package management system itself. Namely, this -statement from Line 80 in that file is the statement. It expects to remove a file, which is not existing. The reason the other packages did not break was because they either do not have a directory () or they actually do have a generated file (netcdf & control). You can technically solve this by adding an empty script to the directory. I manually did this, and the package built correctly again. As for why this change happened, I cannot tell. The particular statement shown did not change between 6.4.0 and 7.1.0 in the file.\nLooks like\nThat's the exact issue. Seems like we will need to wait for upstream to fix it. How does this usually get handled\nSeems like we should apply to the patch:\nThere are still 5 packages failing to build, but that is a lot fewer than before."}
{"_id":"q-en-nixpkgs-412a1383b962dde1299fb7f5d5d605e69e292a0b2e5a8050dbf905b0d575dc15","text":"Traceback (most recent call last): File \"/nix/store/z4sd700l05shq1z72g21w0fav2xs955d-pipenv-2022.1.8/lib/python3.9/site-\", line 83, in androidfolder from jnius import autoclass # noqa: SC200 ModuleNotFoundError: No module named 'jnius' The error is belong in pipenv using jnius on Android. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --- system: host os: multi-user?: sandbox: version: channels(nix-on-droid): nixpkgs:"}
{"_id":"q-en-nixpkgs-e107465c359de3cfbe87f7f920763f822d91cdb3981203df562a12db68021f11","text":"Build of failed on x86_64-linux as of . This is currently breaking python3Packages.tensorflow-bin. python3Packages.tensorflow-bin maintainers cc: python39Packages.tensorflow maintainers cc: Other cc: n/a nixpkgs at commit This issue was automatically generated by . that created this issue. Internal tag:\nI'm not quite sure why this doesn't have the same internal tag as they seem to be the same AFAICT.\noh I see... the hashes in /nix/store/[hash]-pname-version for each of the dependencies changed. hmmmm"}
{"_id":"q-en-nixpkgs-9e77945f4975f4a663e0de5a0c586156fb9eab17ff08368976c04b97e59178a4","text":"Build of failed on x86_64-linux as of . This is currently breaking maintainers cc: google-cloud-cpp maintainers cc: Other cc: n/a nixpkgs at commit This issue was automatically generated by . that created this issue. Internal tag:\nThis has also broken tensorflow-datasets.\nFixed in"}
{"_id":"q-en-nixpkgs-451cab574cc9604bc25a466d5ad6197d77eaf3bdd7635fc351b5f0063a70ef0f","text":"I am new to nixos... When I try to use printScr Key on keyboard. it not responding. non of the printScreen optins working.. Steps to reproduce the behavior: ... press Key PrintScr ... ... Should take a screen shot ask for location to save to. If applicable, add screenshots to help explain your problem. Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nI am not too sure but according to the , cinnamon does not have a \"native\" screenshot tool but gnome-screenshot could be used. I am not too sure if enabling cinnamon auto includes gnome-screenshot. If it doesn't you can just add it and see if it solves your problem."}
{"_id":"q-en-nixpkgs-6e814bef95cc01d0e68eb055cdc2ca86f258985f95d1666ac9b67b5845836e40","text":"Build of failed on x86_64-linux as of . This is currently breaking . python3Packages.pytorchWithCuda maintainers cc: magma maintainers cc: Other cc: n/a nixpkgs at commit This issue was automatically generated by . that created this issue. Internal tag:\nThis looks like another gcc11 failure. could you take a look?\nSure thing\nGoing to attempt to bump from 2.5.4 (2020) to 2.6.2 (2022)"}
{"_id":"q-en-nixpkgs-01c127fa5757ecb6d9543bfa8f46e2b5c0b0b4af777410b459a7a35ce0a32638","text":"Somewhere between and there is a commit that changes the default fonts on NixOS. Most text is just different (I'd say, a bit less readable). But some text in the terminal (gnome-terminal) is completely borked (in regards to positioning): ! Look at how the 's' is drawn on top of the 'M'. And generally the positioning of glyphs is... not right! Compare how fonts look in a few apps on nixpkgs vs . System: x8664-linux NixOS at master branch (). No fonts.* settings in my file (never had that). GNOME3 desktop and gnome-terminal.\nMy guess is . cc I'll try to switch to master and debug what happens.\nWild guess, is there a and a folder? Could you post the result of ?\nHm, I upgraded another machine to the same commit. Fonts look generally a bit different than they used to (for instance, in chromium + gmail). (I think the glyphs are just thinner and there is a lot of \"air\" around each letter.) But the horribly broken stuff in gnome-terminal doesn't happen on this machine. What's going on?\nCan you try to remove and and do a to see if that improve the situation?\nYes it does! Now all text look like it used to. Thanks! After this, is no more, only /etc/fonts:\nIs this something we can fix in the NixOS module that sets up /etc/fonts, so that users upgrading from 16.03 to 16.09 won't hit this?\nA heuristic idea for this -- a function : if path is a file, return True iff it's a symlink to . if path is a directory, return True iff for all items in it is True. On any conflicts, if old path is static, it's safe to replace and/or delete stale. Otherwise make a backup and notify the user via a journal entry and console output. What are the possible caveats? This would certainly be nice to implement. I can look into this if it seems a good idea, but I suspect corresponding code is in Perl, which I'm illiterate with. On July 15, 2016 1:25:53 PM GMT+03:00, \"Bjørn Forsman\" wrote:\nYes, the cause seems to be in the function of .\nTruth be told, I like better an idea of making fully static, with several files (like when ) explicitly set as symlinks to somewhere (?). This simplifies the whole bookkeeping a lot, but however is a radical move. I'm eager to try implementing this but we need to survey possible cases when people manually change something in . Maybe I'll get a chance asking in the ML...\nin principle I'd like that, but in practice FHS is pretty unclear about whether /etc is supposed to be mutable. Would definitely be interested in what you came up with though. Personally, I'd rather fix more programs to not use /etc and rather to link directly to configuration in the store that they need (since we get more granularity there and know exactly what needs to be mutable), but that's probably a lot of work in some cases.\nFor input method packages, I used dedicated packages for the and the folders under via . Could a similar approach for be a way to solve this problem? (edit: and accessorily a way to too?)\nwhy would we care what FHS says? :-D\nI get this line when switching (and then fonts look ugly): Where 81 line is Next fixes the fonts:\nOpened a PR with a fix:\nsee for read-only /etc\nNice find! I haven't yet done any research, so this thread will come handy."}
{"_id":"q-en-nixpkgs-691db861dfd6227e4bb96f6fc28d08c47cb01ea78cfa5f7f50d305ca9904ecf8","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: haruna <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 0.7.3 desired version: 0.8.0 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: I tried to update it myself, but when trying to open the result I get either: or: Some dependencies could be moved to , since they are build time only. The project supports other than . Since uses the former, maybe switching could be considered. The homepage is outdated. It's now \"\". Please tag this issue in your PR.\nCan you (re)produce after this update?\nThis is an issue requesting an update, not a pull request (edited the title to be more clear). Still, with the update I tried to do by myself I couldn't open Haruna in the first place."}
{"_id":"q-en-nixpkgs-e6d29004b9ae4ef54fe6c9471a2ed00e858f19c4fd1e60ac3075386a3d20c3cc","text":"Enabling services.hardware. (as a systemd service) breaks. Steps to reproduce the behavior: error about and several other errors like \"has not been declared\", \"does not name a type\", etc. To rebuild xow and everything else without any issues. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nThe git not found problem is not fatal and the derivation overrides the version info anyway. Looks like another victim of our stdenv compiler upgrade. Looks like upstream master has a fix for that. Upstream is also orphaned so upgrading the derivation to an unstable revision is going to be needed in the short term and in the long term a replacement should probably be found. I don't have a working controller to test xow with any more so if someone wants to adopt the package, feel free to."}
{"_id":"q-en-nixpkgs-51d245dbd6bc119eb68f30fb419af589f321780888e879900bda0928bfd1cffb","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: rtl-sdr <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 0.6.0 desired version: 0.8.0 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: This package is currenty using but it should probably switch to Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-ae7f95bec28599e2bb9edd6e7559b0f11c41e3af38da217bb26375001b80d063","text":"For example This test has always been failing: From the logs: /cc\nWhoopsie! Maybe instead of keeping hard-coded URL to x86-64's Alpine we could build something dynamically? (this way the test would automatically work on all architectures) I'll try to take a look over the weekend, but feel free to take a shot at it, if you want\nHmm that is a cool idea. Though OTOH it is currently nice that when this test fails it is clear it is a problem with running the image, and not with creating it. Did you get that from somewhere or did you create it? Is it the 'mini root fs' from ?\nYeah, that's the one;"}
{"_id":"q-en-nixpkgs-7ade68d8c2fd64aec2802727697e92e626054336ef52b948c770e432424df472","text":"fails to build. Looking at hydra when it started to fail () it could be due to GCC 11 update. Just adding to the only leads to the next error. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --"}
{"_id":"q-en-nixpkgs-432665378abf6fad489c6ea4adc81c25b41f18d1ffd627eed882de6c399371ca","text":"On nixos-unstable it appears that was renamed to . However the config line for applying the option still references the obsolete option name: This means that enabling produces a warning like Relatedly, using this without defining also throws an error ' is used but not defined.confignullOr submodule` with no default value). Steps to reproduce the behavior: should not produce any warnings or errors. Uncommenting the line removes the error but leaves the warning. (though the flake output in question is for )\nYou used the wrong GitHub handles for some of the maintainers. cc\nOops. I copied the names out of the file as there's no GitHub team for it, but I forgot to indirect through . Is there any simple way to look up the github handles for a given module? I'm not even sure how to get at the module programmatically.\nIt works, but it isn't pretty."}
{"_id":"q-en-nixpkgs-35332c20e5b627ae6be1c9dcf06be92a5fca7bcfa6fd35fe26f18ba7ff711ec5","text":"See Hydra failure: I could not find anything directly keepassxc-related in the changes on hydra from last successful to first failing build. CC Also relevant: Version bump to 2.7: A fix for this issue might give a merge conflict there in the worst case, or alleviate the build problems there in the best case.\nDoes keepassxc depend on Qt? We're stuck on an older Qt because of the SDK version for x86_64-darwin. I think keepassxc may require a newer SDK.\nmentions QT = 5.9.5.\nCurrent version builds just fine. There's one test failure instead: I'll see if that's fixable...\nAnother issue though is that somehow became wrapped so can't be loaded anymore. Which may be related to the test failure... Either way, after trying to run the app, the gui just spins with 100% CPU. The cli part works just fine though.\nI came across this issue because I was investigating something else, this is because of the wrap qt apps hook which wraps all the executable files, even if it's a dynamic library. This can be fixed by removing the executable flag from the library. You might be able to resolve it with something like: In the derivation. This should be fixed in the fixup script however"}
{"_id":"q-en-nixpkgs-88cb8556346a6e8a390a3172d534bb0cd65e40f5b0e16ee7410d08902d841170","text":"nixpkgs-review stopped working on recent branches, I didn't update nixpkgs-review before it break so I assume something internal in nixpkgs has been broken. Steps to reproduce the behavior: or there is a build failure related to hydra <details<summaryLong output of the failure</summary</detailsit shouldn't fail\nThis is certainly related to , ping :)\nSorry about that, it's a bit annoying that aliases are allowed locally and not in nixpkgs-review :/."}
{"_id":"q-en-nixpkgs-fae25a679823e56a7abf99349fea6204dbbdf1cd776f659824aa147a2b1351d4","text":", CVSSv3=9.8 (nixos-unstable) CVSSv3=9.8 (nixos-unstable) Vyper is a pythonic Smart Contract Language for the ethereum virtual machine. Versions of vyper prior to 0.3.2 suffer from a potential buffer overrun. Importing a function from a JSON interface which returns generates bytecode which does not clamp bytes length, potentially resulting in a buffer overrun. Users are advised to upgrade. There are no known workarounds for this issue. Vyper is a pythonic Smart Contract Language for the ethereum virtual machine. In affected versions, the return of is not validated to fall within the bounds of . This issue can result in a misinterpretation of the integer value and lead to incorrect behavior. As of v0.3.0, is validated in simple expressions, but not complex expressions. Users are advised to upgrade. There is no known workaround for this issue. Scanned versions: nixos-unstable: . Cc\nSee also:"}
{"_id":"q-en-nixpkgs-b08cd3e918cccfcf29fecdfea6a3dc58f0c3afc0a0c3d8de39cf83420ba18bfb","text":", CVSSv3=7.5 (nixos-unstable) CVSSv3=7.5 (nixos-unstable) A lack of password length restriction in Zammad v5.1.0 allows for the creation of extremely long passwords which can cause a Denial of Service (DoS) during password verification. A lack of rate limiting in the 'forgot password' feature of Zammad v5.1.0 allows attackers to send an excessive amount of reset requests for a legitimate user, leading to a possible Denial of Service (DoS) via a large amount of generated e-mail messages. Scanned versions: nixos-unstable: . Cc Cc Cc"}
{"_id":"q-en-nixpkgs-43ce9a71ac2ae9a28a0063a564e6ee7680729ab26935e25ef56e454e409ce458","text":"Project description KDiskMark is an HDD and SSD benchmark tool with a very friendly graphical user interface. KDiskMark with its presets and powerful GUI calls and handles the output to provide an easy to view and interpret comprehensive benchmark result. The application is written in C++ with Qt and doesn't have any KDE dependencies. Metadata homepage URL: source URL: license: GPL 3.0 platforms: Linux\nI am currently trying to work on this. It builds and runs, sometimes I run into a Kauth error\nsee , please review"}
{"_id":"q-en-nixpkgs-a7079e198aca988098951b4fa76cc9c83dc618ebe0d3503c23f6bfa32d8f7208","text":"Since May 10, .NET 5.0 is EOL (End Of Life), and should not be used. Should it be marked as EOF similar to dotnet 2.2 and 3.0? See All packages using dotnet 5.0 should be updated to use a newer, supported version. These packages still need updating: [x] - [x] - [x] - [x] pkgs/development/tools/build- - [x] -\nThere is at least one package, which currently uses dotnet 5 sdk - . So we will need to update it beforehand. Btw, I'm not sure if we have any policy regarding packages removal, for nixpkgs in general or specifically for it's dotnet part. , what do you think?\nThe package is also using .NET 5: . But this is because I made the PR upstream updating it from .NET 4 to 5. I suspect that the transition to 6 will be similarly easy.\nI think we should just mark it as depricated, as many projects do still use it upstream. Dotnet version 3 is also still commonly used, so I dont expect it would be realistic to remove version 5 any time soon. also uses dotnet 5 by default, I'll go ahead and switch that over now. See . That doesnt mark dotnet 5 as deprecated or change anything from a SDK perspective, just some preperation work.\nSee as well, it adds a warning whenever dotnet version 5 is used. Not using a throw for now since a few packages depend on it and I dont have the time to clean that all up at the moment.\nCan you provide a list of packages that need updating?\nBased on a quick grep: pkgs/development/tools/build-\nI believe that after all the above PRs are merged, we should be able to safely remove .NET 5 from nixpkgs :)"}
{"_id":"q-en-nixpkgs-d54db5eab51dbf1c00801656215374098c238bfa9f0592c848836aed5763c9c2","text":"Shotwell can't build thumbnails for video files. package has an additional executable , which is broken. When run manually, it fails with this message:"}
{"_id":"q-en-nixpkgs-15bc34de874989a461287283e008176cdc40a4547f6dec970d768d05e947bf1a","text":"Project description  Selectdefaultapplication is a very simple GUI application that lets you define default applications on Linux in a sane way. Basically it just loads all installed applications by reading their .desktop files, reads the MimeType fields to see what it supports, and updates with what the user wants. Not all desktop environments provide such a tool, prompting for an agnostic tool. Metadata homepage URL: source URL: license: gpl2 platforms: linux Additional info: AUR package:\nYou can try using this in an overlay, use .\nSeems to work!"}
{"_id":"q-en-nixpkgs-54c03bcf57ccdd85a2e9b5e1558e217789ba2b9db1c52fc4b96d2a1a4b02829f","text":"I'm bringing up new RPis and their clocks are often wrong. Even though upstream systemd has given exceptions to be able to skip dnssec check for ntp sync, it doesn't... appear to be working. If I set the date manually, often by time I check timesyncd will have already taken care of kicking in and it reports that it's synced. Steps to reproduce the behavior: the clock badly. . and see if your DNS works\nIf I read it right, only works for systemd-resolved. One general workaround is to use hard-coded IP addresses for NTP (our routers actually do that IIRC).\nThat commit seems to indicate that it should bypass DNSSEC for timesyncd. This would get my clock back in sync, thus resolving DNSSEC errors for the rest of the system. But timesyncd isn't doing its job. I assume that's because: date wrong -dnssec failure. But maybe also fails if the clock is just too askew? I'll try to look more later, maybe I have to opt-in large corrections? Seems a bit weird to me. Or do we maybe use timesyncd in a non-standard way?\nI ran into this on a Raspberry Pi 3B. I'm using systemd-networkd including systemd-resolved. At some point DNS started failing, and I noticed the system clock was out of sync by about 4 hours and not synchronizing anymore. The log of systemd-resolved was full of DNSSEC s: I'm not sure what changed, because I was able to install NixOS just fine. It also was offline before for some time without this problem happening (as far as I noticed at least), and it started happening after it was online for some time without break. Also, the systemd-native workaround should be working on NixOS as well, I can't find anything that indicates a non-standard setup that might overwrite tha workaround. Checking is the standard original systemd unit file that includes the relevant line and as far as I can see it isn't overridden. I also checked and it includes . For now, I worked around the issue by specifying my NTP servers by IP (I used the ones by the ).\nClosing as this should be fixed by Please reopen if this is still an issue."}
{"_id":"q-en-nixpkgs-49083fac66991291239ddadc3c6dd4dcacc8378c187fffb755d73abbcfbf44ee","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: syncthingtray <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 1.1.3 desired version: 1.1.20 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR.\nThere's a new dependency now needed for syncthingtray - but there's a problem in the logic upstream here: Here's my packaging attempt:"}
{"_id":"q-en-nixpkgs-ff0561f181b3eeadca0ccd48b4b6648fe8926f9af686a08ed9b712e86188d52c","text":"The file generated for nixos is not compatible with the R programming language's interpreter (). On startup, the R interpreter prints the following warning (though it does continue as normal after printing the warning): This could be resolved by adding a trailing newline to . Steps to reproduce the behavior: the R interpreter by running , or if you don't have R installed. No error message should be printed. I'm not entirely sure whether this should be resolved adding the trailing newline to , or whether R should be modified so it doesn't assume a trailing newline. The man page for describes the format of the file as \"a newline-separated list of environment-like shell-compatible variable assignments\". The existing format (no trailing newline) is indeed newline-separated and shell-compatible (bash and zsh both have no problems with it), so maybe I should bring this up with the R folks instead? On the other hand, all the other distros I've checked (ubuntu, archlinux, and alpine) include a trailing newline, so maybe we should just stick to what they're doing? The fix is trivial: Let me know what you think, and I'd be happy to open a PR wherever we decide the fix should be made. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nThis! It shouldn’t be needed and yet here we are. No harm in doing so that though, so I’m all for this. Sounds great!"}
{"_id":"q-en-nixpkgs-c6fa64b4692f8c0fd8c8b6824b784b31013a036748cdeeed27b74f349df61da1","text":"The Zen kernel is failing to build for me since the update to 5.18, and so I thought I'd open an issue since I can't seem to find any existing info about it. Bisected to system: host os: multi-user?: sandbox: version: channels(root): channels(alex): nixpkgs:\nLooks like it didn't build even in the CI of : But, we already have a new version: Cc (since you're always the one to bump it)\nUnfortunately 5.18.1 doesn't seem to build either for me. :( Given that the file in question hasn't been touched in months, I'm wondering if this has something to do with the move to C11 for the 5.18 kernel. But that's just a guess. I can think of two possible quick solutions: Include a patch that adds a comment Ignore that warning everywhere in the kernel\nStaging recently got change to avoid Werror-style failures.\nGiven that that change will take a bit to get to master, do you think it would make sense to set for just the zen kernel in the meantime? Or even just revert and move it to staging so that it gets merged when the change does? I'm asking because—while it's not too much trouble to add patches to my personal NixOS configuration—right now the zen kernel is broken for everyone, and so it might be a good idea to get it back to a working state in the meantime. I'm not sure how many people use the zen kernel though, it seems to be not very many, and so if it's better to just leave it as-is for a bit that's fine too; I don't really know how such things are usually handled in nixpkgs and am trying to learn. :)\nI'd say setting for zen kernel is perfectly fine. Should be a matter of adding it to Do you want to give it a try and propose a PR? AFAIU you already run that kernel version and have a chance to validate it if still boots. You could also add a TODO to remove it once staging is merged.\nshould've gone to master and/or be merged with the next kernel bump or something if the rebuilds were that critical which I don't think they are. We could cherry-pick it if we need it now.\nI made , but it's late here so I don't have time to test it properly and am waiting to open a PR until I can do so. It at least builds though. I figured while I was at it I should also bump the version to 5.18.1-zen1, but in hindsight maybe I should wait until the mainline kernel is updated to 5.18.1 in nixpkgs? Anyway, I'll test it when I have time tomorrow evening, make any necessary changes, and open a PR, unless we decide to cherry-pick ? That seems like a nicer solution to me personally, but I don't really have a strong preference.\nI finally got around to testing this and opened a PR: :). I'm personally in favor of cherry-picking , though."}
{"_id":"q-en-nixpkgs-1b183ecf7f714be75c3801a1cc8175e8bb802f6c479c4df2bc6af4efd092480a","text":"I setup the server quite a while ago. I recognized that the service is not running. As I normally do auto update I saw that the process failed as the configuration changed. After rebuild + switch the service fails to start: Steps to reproduce the behavior: Service be able to start. I tried to 1) chmod +w -still fails and permissions are somehow adjusted 2) delete -file is created, startup still fails\nDigged a little bit deeper: The error is thrown by the pre-start script. It seems that my installation is missing the which causes some gitea command to generate it and write it into . Which fails as the file is set to non writable by the pre-start script."}
{"_id":"q-en-nixpkgs-bd73ea19dd03def4ab43efb96b03fea461ce8729f24d01baafc4ad9e6ab7e035","text":"I've been using PipeWire+JACK with until now. I'd like to switch to , but the service crashes at startup. The key error in the logs seems to be: . The result is that : (1) I have no sound, (2) I see no volume control knob in GNOME, (3) I see no devices appear in or any other JACK patchbay tools. Here's (hopefully) all relevant build configuration: Here are some logs: It seems others haven't run into this. What could I be doing wrong? Steps to reproduce the behavior: the above build configuration in the computer the above-written commands to view relevant logs I expect to the service to start, and to have sound.\nDo you need a systemwide setup? That seems to be the issue here.\nMy use case is: I use virtual machines that run as another user (for security), but I want them to have sound. Using allows this, and it's the only (secure) solution I'm aware of.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nThanks for this patch. I was finally able to test it. Unfortunately, I still have the issue. Your patch removed the \"\" error, but the wireplumber service still fails to start. Here's some evidence that your patch is applied: Here's the failure output/logs: According to the \"\" shouldn't be a problem because wireplumber shouldn't require X11. Looking more broadly at logs, there's this:\nThere have been some dbus related fixes in Wireplumber master, can you try that?\n: I don't know how to cherry-pick from wireplumber; I only recently learned to cherry pick your pull request. I suppose I can wait until the next release. Maybe I should open this as an issue against wireplumber? Also, this crash only occurs with . Can you offer insight into what this means, so I can best phrase my question? I think the pipewire socket moves to a different what else? Does pipewire and/or wireplumber run as a different user? Anything else?\nshould get this fixed for good.\nAlso, if you don't know what does, you should probably disable it, as it's not a well supported configuration."}
{"_id":"q-en-nixpkgs-2276ccccfee7ad41a711d09c0864f9b4088fafb2b553eccda3089c83f165d7e5","text":"Fonts are missing, square characters are displayed ! Steps to reproduce the behavior: vmware-workstation vmware workstation or player I started a discussion on <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --/cc\nI can reproduce this, probably related to some missing font in freshly installed systems\nI launched vmware with FC_DEBUG=8, but I'm not able to interpret these errors messages, hope this can help\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nI found reason for missing fonts, it just loads from instead of system path and vmware's config points to wrong FHS paths, removing file fixes it, i will make a PR with fix, thanks.\nI confirm that deleting fixes the issue. Thanks a lot !"}
{"_id":"q-en-nixpkgs-e84a0f388220133bf028249b3e445ff0ee54159aa64ecaa66c1f28be89dd88e3","text":"Building fails when rebuilding system configuration. System configuration should build successfully with on kernel 5.18 There are no current maintainers. This appears to be a repetition of\nArch Linux appears to have a new patch here: Attempting to apply it has failed for me. I tried this block: Maybe someone who helped solve this the last time can comment?\nAlready have a PR up .\nLooks like an issue with the patch for 5.17. Try replacing it with this one:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-b1e4186ea44fbdfff31adf04b92c794519b5edabd5fab5728b902c46d0b39bad","text":"Steps to reproduce the behavior:\nCan you check if has an effect?\nI've got this error:\nPlease check the full log of that build and look for ( catching the actual error) and any preceding lines with (actual error from the compiler). What you posted are just harmless warnings and (an error has occurred somewhere else so needs to quit).\nUpstream issue:"}
{"_id":"q-en-nixpkgs-f53a76f942c0f3c5155241d74e01b1b2cba209b6e7f538d72f99368e3abbc67a","text":"nicely describes the issue and the fix Steps to reproduce the behavior: salt on to run it errors salt working and may be of interest to ;) <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result."}
{"_id":"q-en-nixpkgs-367808b5573242b149397fb3a3ce49ebbff48b7b9a4cff3c1699c39322f7f2e1","text":"tig crash on macOS (m1) Steps to reproduce the behavior: -shell -p tig ANYGITREPO type \"/\" to enter search mode Don't crash ! Exception Type: EXCBADACCESS (SIGSEGV) Exception Codes: KERNINVALIDADDRESS at 0x082a0d5c00ab0a2b -0x00000d5c00ab0a2b (possible pointer authentication failure) Exception Codes: 0x0000000000000001, 0x082a0d5c00ab0a2b Exception Note: EXCCORPSENOTIFY Termination Reason: Namespace SIGNAL, Code 11 Segmentation fault: 11 Terminating Process: exc handler [] VM Region Info: 0xd5c00ab0a2b is not in any region. Bytes after previous region: Bytes before following region: REGION TYPE START - END [ VSIZE] PRT/MAX SHRMOD REGION DETAIL commpage (reserved) - [384.0G] ---/--- SM=NUL ...(unallocated) ---GAP OF 0x5f9000000000 BYTES MALLOCNANO - [128.0M] rw-/rwx SM=PRV Thread 0 Crashed:: Dispatch queue: 0 ??? 0x100b84c6c rlinitterminalio + 272 1 ??? 0x100b84c60 rlinitterminalio + 260 2 ??? 0x100b6fc2c readlineinitializeeverything + 164 3 ??? 0x100b6edd0 rlinitialize + 44 4 ??? 0x100b6ed48 readline + 36 5 ??? 0x100981ce4 readprompt + 80 6 ??? 0x10098aaa0 searchview + 56 7 ??? 0x100971680 main + 2608 8 dyld 0x100d61088 start + 516 <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result."}
{"_id":"q-en-nixpkgs-af7d6fe2c80b59906bd696fb356e33c272c2874da25c182fda7e3a76254f0d8f","text":"Noticed while debugging hash errors for . Since got merged via (bisect log further down), doesn't produce fully deterministic results anymore. Sometimes I get an identical hash, several times in a row even? But often it's different. It is worth noting that the fetcher output has changed after that commit. Using as an example, the final line has been replaced by for the main repository + every fetched submodule. Steps to reproduce the behavior: (current master at the time of writing) repl: multiple times and you should get a different hash eventually. \"Best\" combo I got so far is 3 identical hashes before a different one. Hashes shouldn't randomly be different. It would be understandable if the not-yet-deterministic was used, but this happens even with . Changing hashes: ! ! Changed output: (last identical lines are highlighted) ! Bisect log: Bisection done with . uses with , which uses under the hood, but the hash change should be observable by directly using as well. CC (commit author) CC (PR merger) Please run and paste the result.\nI did a quick investigation. First of all: I was able to produce the same results as the OP on my machine. My initial investigation led me to believe that this behavior is caused by 2 bugs. One introduced by commit and the other one being unrelated to this commit. The bug introduced by the commit in question is that seems to use in situations where it shouldn't. The second bug is that specifying produces non deterministic results. Evidence for this is provided by the following experiment: I checked out a revision of that was before No I ran the same experiment as the OP suggested but instead of using I used . As you can see I got two different results for the same call. I suggest that should be reverted to mitigate the damage by the bug it introduced. I don't know well enough to do anything about \"the other bug\".\nThis is a .\nPinning this until the fix has made its way into and staging.\nThanks a lot to everyone involved in investigating and fixing this! Also, I'm sorry for breaking this in the first place!\nDon't take all the credit for yourself. We broke this together ^^\nI'm seeing behavior that I think may be related to this issue. Please consider the derivation contained within the following flake: Changing the dummy value of results in different suggested replacement values. It seems to me that any value except for the corrected one should result in the same suggestion. Furthermore, all suggestions do build successfully (so the allowed values of are not unique) but the only one that actually respects is the empty string case. This may indicate a separate problem with ."}
{"_id":"q-en-nixpkgs-437897a83c2c26dbddc72bd7c2b886b50c0f03914c562ecb313684b3c740d5aa","text":"The reason why I didn't open a PR, is described in . I'm aware of four platforms for Qt's themes right now: Using GTK themes through + ; Using GNOME themes through + ; Using Qt themes, with the non-KDE solution , which uses ; Using Plasma-targeted Qt themes, with as , as done natively in a plasma session, using the theme's colorscheme. Steps to reproduce the behavior:  and it does not provide ; equivalent to  providing ; , then add some Pasma-target themes like , then adjust envvars , set the theme in or with , you'll get the right theme with the proper color scheme. should provide (or ) as an option, it should add only the required packages to the system, and set the proper envvars. (or maybe rename it, since qt6 also uses the same variables and envvars) qbittorrent + breeze + + dark: ! qbittorrent + breeze + + dark: ! qbittorrent + (ignored my theme): ! qbittorrent + (also ignored my theme): ! I tried to discover what package that is being when evaluating provides this platform-theme, but I failed miserably. stated as maintainers of plasma-integration: stated as maintainer of qt5ct: qt5 module contributors: cc Please run and paste the result.\nNot exactly your bug, but we should remove and integrate it with configs instead.\nis another QT platform theme that can be integrated into config."}
{"_id":"q-en-nixpkgs-4c78ffbd9b70e252a29ed85aeed5726d21db0313542e48eab630db487b7a499b","text":"Latest version uses Electron 18 instead of 16, which fixes issues with Wayland + Nvidia graphics. <!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: mullvad-vpn <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 2022.1 desired version: 2022.2 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-e4a5d2e448e13b08d64220f5b2613c1cd1c13cf5abc26914d8a66c2ce6d04cbe","text":"Steps to reproduce the behavior:\nI don't have an M1 to test\nOh, it suffices to use . I'll open a pull request."}
{"_id":"q-en-nixpkgs-8295a707525233f5f5a5a70d4fa2579b133d3574926de60949154af5928fcbaa","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: winePackages.{unstable,staging} <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 7.10 desired version: 7.12 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR.\ncc\nPR at , needs to be tested still. My last attempts at updating have been held back by the fact that there are so many variants that building everything is prohibitively time consuming on my machine, and PRs that were ready have been lingering so long that I've had to update them with one or two even newer wine versions before they were merged, so I was not particularly motivated."}
{"_id":"q-en-nixpkgs-c03c73c94e66846c7ff0d80c6f63d7a297694d7eecd5e181ff916cf48dd05c4d","text":"The onlykey package crashes when a file dialog should open (\"Save as\" / \"Choose file\") Steps to reproduce the behavior: onlykey to the Firmware / Backup Tab \"Choose file\" or \"Save as\" A select file window should open, not crash. ! ! I have nautilus and LF installed, which is successfully used by other programs (firefox etc). Already tried installing gsettings-desktop-schemas and setting XDGDATADIRS as described . I use dwm and no Gnome and no Gnome Packages except nautilus. I would guess it needs to use wrapGAppsHook as nativeBuildInputs but I have no idea how/where to add that. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- [username nix-shell -p nix-info --run \"nix-info -m\" system: host os: multi-user?: sandbox: version: channels(root): channels(username): nixpkgs:\nHave you cloned nixpkgs? Did you find the nix file where the package recipe is defined?\nYes, the uses node2nix which I do not really understand. It doesn't look like other .nix files. According to this , wrapGAppsHook can be like in . But I can't find , the hook or even just the part where it is supposed to be .\nthe important parts of are and runCommand so you could replace the call with the usual"}
{"_id":"q-en-nixpkgs-6782afab1da039f8bd69b2e6dc7fdb170b53de554603f2860a28a42b9c96d2f2","text":"Project description music streaming server / subsonic server API implementation Metadata homepage URL: source URL: license: gpl3 platforms: linux package: The package is already there, maintained by It'd be nice to have a module too!\nI can submit a module PR once is merged, as it adds some new options (filesystem watcher)."}
{"_id":"q-en-nixpkgs-22f864b66178f6d838fe9c282b5930c95a4957f66a17085f911c215c2f451a82","text":"Dive now fails to build on darwin. This was working up until maybe a week ago. I poked at the vendor code that is failing. Looks like there some failed symbol references that look like: I saw some ramblings about libSystem updates, iirc. Is this related? Please run and paste the result.\nLooks like dive doesn't build with go 1.18 yet. You can fix it with:\ndive needs to update its vendoring, not reverting back to golang 1.17. Please don't pin programs needlessly to go 1.17 just because of this.\nThat is only true for aarch64-darwin, not any other platform."}
{"_id":"q-en-nixpkgs-72a9ea0e1f6977e60dc56fc385c6d0fff69953cad18bf8e355b2bd430c5a42fc","text":"Steps to reproduce the behavior: build /nix/store/1ikzh2ax90ji2156hh921gp4iambgwsm-kubectx- Add any other context about the problem here. Please run and paste the result.\nsry I dont use darwin, my reflex would be to mark darwin as not supported except if someone can step up\nit looks like this may be solved by upgrading x/sys but does anyone here know how to do that in nixpkgs? I have no idea about the go ecosystem here.\nPlease test from master.\nI was able to get this to build with the following: rev: sha256: vendorSha256:\nI can confirm it's working for me, thanks very much!\nThe approach to pin go 1.17 here is wrong. The package just requires a bump of"}
{"_id":"q-en-nixpkgs-783ea8a796fb0bc1a6261d06cb56934a08ce6d8afc2ffe41e49009084ca41021","text":"Steps to reproduce the behavior: The regression is directly caused by , and the root cause is the library is not properly linked, upstream issue: <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result."}
{"_id":"q-en-nixpkgs-8b482861578cc10dd04725b717c9996a307357f5625eccd27c13a70b04e68424","text":"Project description Hatch is a modern, extensible Python project manager. Metadata homepage URL: source URL: license: MIT platforms: unix, linux, darwin\nIsn't this already packaged as ? The projects look similar\nAccording to"}
{"_id":"q-en-nixpkgs-303f4fe4fe621ed6ad214e852ff87b464305258eb971dcd9a07f78126ea02c55","text":"When u type grapejuice app or grapejuice player you get the following error: Steps to reproduce the behavior: Grapejuice opens as normal\ngtk3 has to be to buildInputs\nAdding gtk3 to buildInputs results in the app opening but results in the following error\ncheck if that Wine home path exists and it contains a bin directory (very weird if it doesn't as I have the same path) for me after adding gtk3 to buildInputs the program starts successfully\nater doing: fixes te issue with the wine path, gonna send the pull request for the gtk3"}
{"_id":"q-en-nixpkgs-777b90614d299b153c2c978f22ba8f5ce2f36bc84d1fce0bc195fc9fa79c360b","text":"Project description Chromium fork with performance & privacy benefits. Metadata homepage URL: source URL: license: GPL-3.0 platforms: Linux, Darwin\nI wanted to try Thorium out of curiosity and built this: <details</summary</detailsBut I decided I wasn't going to use it, so I'll leave this here in case it's helpful to anyone. It seems to work just fine, but I haven't tested it very much.\nWhat is with this?\nthat is not it, we need thorium in nixos, i suppose this is browser of choice for many users specially, it is not filled with unnecessary bullshitery like brave and has chome store support and widevine cdm (unlike ungoogled chromium)\nWould definitely also like to give the thorium browser a whirl, as there is currently no easy way to use Google Chrome Sync for ARM based Linux.\nPR\nany update on this PR? popular browser. seems to have stalled\nI have been using definition for a few months without any issue but since he has archived it and the version is out of date here is the same but\nThorium was removed again due to the mentioned comment. Unless things change upstream, we won't consider packaging Thorium in nixpkgs."}
{"_id":"q-en-nixpkgs-5519824d217095c326cd9d69b66af1a8f5bd81902722a6e82a1a9c38fc5255e3","text":"There are a few more flags we should enable on : - (I had to add these flags for the sake of something that depends on secp256k1 - I don't remember for sure, but I think it's bitcoinj.)\nThe schnorr flag appears to be gone in the latest version."}
{"_id":"q-en-nixpkgs-cb7172d3471dada477951849b90b67c78cd38095ba0d3e5c9191a8b0abe5cab0","text":"Project description vt-cli is the official command-line tool for interacting with VirusTotal. It is written in Go. Metadata homepage URL: source URL: license: asl2 platforms: unix"}
{"_id":"q-en-nixpkgs-b2021ddda225dac0619c5c57efda502aca4af0488b14a703d451dacd4d61df7b","text":"Steps to reproduce the behavior: purple-signal this as plugin for pidgin. Pidgin will not find it, as the output of the derivation are not as expected. First hydra build: Last hydra build: The generated output is within the wrong folder. It is expected to have folders and , but actually everything is within a subfolder . The expected structure can be compared with Got via: Please run and paste the result."}
{"_id":"q-en-nixpkgs-72f7eaa7a9591a79c9553972296253852a2f48828926b2ab2d43de73f64b2d94","text":"Project description Natural Language Search Engine for your Org-Mode and Markdown notes, Beancount transactions and Photos Metadata homepage URL: source URL: license: gpl3 , ..."}
{"_id":"q-en-nixpkgs-feef0a0da1217a6e1a8bbcfff060b4c360f0284a53a4009f78546f97ed0601b2","text":"Project description A program that allows you to set the terminal theme from the command line and scripts. ! Metadata homepage URL: source URL: license: MIT platforms: linux, macos"}
{"_id":"q-en-nixpkgs-821eeda98efb56bdc3dc6681d88f6843e75cc419a0a7261c20019676df61254e","text":"There are in-the-wild tarballs, especially in NPM registry, whose directory entries has permissions instead of . GNU Tar is unable to unpack them, while some other unpackers (WinZip, 7-Zip, ...) are We may need to patch or have a fall back unpacker to handle gnutar failures. Error about sha256, but it fails earlier on unpack phase.\nThe error caused by , , , .. directories are created with permissions which prevents creating files inside. That is clearly a bug - it should set relaxed permission for unpacking time. But taking into account the age of that software it is unlikely to be fixed upstream, so a modern replacement or fallback would be desired.\nPerhaps bsdtar is an option. It also supports additional archive formats.\nreplicated. this should be fixable by passing to ensure all files are marked as writable.\nwas merged, tracks when it hits channels. This issue can be closed."}
{"_id":"q-en-nixpkgs-8a3b4f0557eb19851ba04682abe1f1f31430fcbe3edb5942c0819233fd4030ef","text":"It's all about client libs (mis)trusting the server, so no security problem for most setups, but still &ndash; we should fix them. I propose to wait for actual releases (assumed within days).\nStaged."}
{"_id":"q-en-nixpkgs-5738173aaed70b03aa1fac0c2e8bdf7c49928523b877480c8f61363ab1017571","text":"Project description Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multi-task model that can perform multilingual speech recognition as well as speech translation and language identification. Metadata homepage URL: source URL: license: mit platforms: all\nPTAL."}
{"_id":"q-en-nixpkgs-778020f56ff9c85d7927ab44c674f9736461f9490058a66e2265b422044f42a0","text":"After installing thunderbird via , no entries are to Gnome Default Applications for Mail and Calendar. Steps to reproduce the behavior: thunderbird Gnome Settings to Default Applications that Thunderbird is not in Mail or Calendar lists A clear and concise description of what you expected to happen. Thunderbird should be in Mail and Calendar in Gnome Default Applications. If applicable, add screenshots to help explain your problem. Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nMore specifically, the issue is that the desktop file lacks and in the key.\nUh, we generate the ourselves, by the same code as for Firefox, so it's a launcher for a browser\nMy proposal desktop entry: We cannot use since is hardcoded to in . References: Desktop Entry Specification: Desktop Menu Specification: Current wrong desktop entry: Template used for Gentoo: Debian:\nI made PR ."}
{"_id":"q-en-nixpkgs-3a3de5e53284de5b28ac755e924b11bf62460797e2e32c8f3cc7821d527a6e1b","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: Jetbrains versions <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 2022.2.1 desired version: 2022.2.3 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR.\nI am not a JetBrains maintainer. Also, its probably easier to run and open a PR tagging me (would gladly review the MR)\nGot it. Sorry I think I just saw your username and some previous PRs and you based on that. I created a PR from the update script. No idea how to test the updates."}
{"_id":"q-en-nixpkgs-fb7f7f21fdcfabb4ca232726230cdd8dfc9c5eb592dfdc22132e784756cb31ae","text":"User creation is broken: Invidious creates misnamed material views in the database leading to a crash. More precisely, creating a first user works, but a crash happens when creating a second user: Steps to reproduce the behavior: in your configuration your invidious instance in your browser a first account , and create a second account --It crashes with the following error message No crash happen Two material views have been properly created, named and instead of a single one named . ! I opened an issue upstream: The maintainer suggested to try with a newer version of invidious.\nIt seems that this can be fixed by passing the to . Source:"}
{"_id":"q-en-nixpkgs-3e3881b137825a89bad099c7c652a6c8bb13e8c411a57d9e624da5e18dfa912a","text":"Same on nixpkgs-unstable:\nsee\nalso see\nI tried cherry-picking both PR's but still unable to build on . I'm on 5.19.16 kernel.\nFYI, your kernel doesn't matter. The headers used here are a separate thing."}
{"_id":"q-en-nixpkgs-3cddee3582fd8097d3b37018f623cdb8082b087b4c4ed806937e39ebe243e0c2","text":"The symlinking between the multiple outputs of seems to prevent it from installing. I was looking to install it to have the dig command available. - System: 17.03pre94050. (Gorilla) Nix version: nix-env (Nix) 1.11.4 Nixpkgs version: \"17.03pre94050.\"\n11.4 still doesn't contain :-( How could we have that as default in 16.09? I consider it quite an important feature for + multiple outputs. would you be against porting that PR to 1.11.5 and making that default in 16.09 and master?\n+1 for a 1.11.5 Nix release in general :)\nSince this seems to be an issue for multiple packages, I am raising the severity on here.\nLooks like the same kind of issue as So +1 for Nix 1.11.5 on NixOS 16.09, if that's what it takes to fix this.\nping again about nix-1.11.5 (my comment above)."}
{"_id":"q-en-nixpkgs-98ddf749866dfa59cbc23646494a6a52e6d01094a365206b6442f7ee3bccc907","text":"Steps to reproduce the behavior: streamdeck-ui I'm unsure why this happens and how to fix it so I would need help on this Please run and paste the result. (unstable - )\nLooks like broke with That PR is full of mines that make it challenging to bisect fully."}
{"_id":"q-en-nixpkgs-9fdeaf523d707deb88371c21349d89749cb6754de3bcf5a7f2ac4a9c8851bfd2","text":"The xen-bridge.service serves an invalid netmask over DHCP to VMs. Start with a NixOS host with Xen dom0 enabled using the following option in . By default, the VMs’ network controllers will be connected to a bridge set up by the xen-bridge.service. The network on is set to , but the VMs set up their networks as . The VMs get their networks configured using DHCP, through dnsmasq running on dom0. Looking at the DHCP exchange using reveals that the netmask sent to the clients (VMs) is instead of , as DHCP option 1 (shown below). The DHCP client on the VM side is likely ignoring the invalid netmask, and guessing a netmask instead. The problem can be fixed by removing the wrong netmask from the dnsmasq configuration file. Then, dnsmasq will infer the correct netmask by looking at how xenbr0 is configured. Will submit a PR. System: Nix version: Nixpkgs version:\nThe issue no longer appears\nDont know if related to this (not sure if it is a bug really) but if you set the mask for the aforementioned bridge, it set it incorrectly. I tried to set it as 24, get 255.0.0.0 as a result... Both 16.04 and this new 17.03 have this behaviour. I understand that this is not the best place to ask, but if what i am doing is producing the expected behaviour, I will apreciate any guidance on using it correctly.\nCould you explain your issue in more detail, for example how to reproduce it? I am not sure how you are setting the net mask.\nHi, i am just using the configurations on section of configuration options. so: Setting to some valid ip adress works but Setting to anything other than the default gives a 8 bit mask. I must say that i used xen on Ubuntu, but I am very new to nixos, so can't say for sure that i'm not doing something else wrong.\nCan't see what could be wrong. Anyway, this looks like a separate issue that should be reported. There might be something wrong with the scripts under in .\nOpened at ."}
{"_id":"q-en-nixpkgs-f952db0e6ba8d40e6422af16e480538ea118dafca7ff51341ae17def5e824cc0","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- : yes, <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 2.7.2 desired version: (cf. PyPI) <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-ff75693cfb67c6744dedb731ee627b9043fbf51331843b4cfa3fcb2671824b73","text":"When multiple ZFS filesystems are present and configured to import, each individual attempts to load the keys for ALL pools/datasets, instead of just its own. A clear and concise description of what the bug is. Steps to reproduce the behavior: encrypted zfs filesystems in your NixOS configuration. & Reboot All targets complete successfully The below snippet from the services's script. It's 'ing, and just looping over all pools/datasets. Instead, it should just specify the name directly in the command. (Generated in ) Please run and paste the result. system: host os: multi-user?: sandbox: version: nixpkgs:\nCan you post the relevant Nix config you used? Do you have set to anything?\nYep, here's my zfs config. uses (unlocked at boot) is unencrypted and have on-disk\nOk, I think this is fundamentally the same issue as . Basically, when there are multiple pools and is a list, we accidentally run with no dataset argument in the import service for pools that have no datasets in the list. So those services end up iterating over all datasets on all pools instead of none."}
{"_id":"q-en-nixpkgs-689d33199430430a7c00d1a57dc47f9b833e2271fab9766610b85946636329fb","text":"twmnd does not launch because qt5 library path is missing. Install the twmn package and run twmnd. twmnd works when I manually set the LDLIBRARYPATH to <qtbase/lib. System: NixOS 16.09.931. Nix version: 1.11.4 Nixpkgs version:16.09.931.\nI noticed the same or similar thing when trying to run twmnd as a systemd user service. It just SIGABRTs: I could reproduce with Although works fine... So clearly there is some impurity in my local environment which is making twmnd work \"by accident?\" Possibly related?:"}
{"_id":"q-en-nixpkgs-2f9b7f03224630d23f8f547948777774391289d20e83cc9b5fd05026c436308d","text":"BlueZ is currently setup to read auxiliary config files (i.e. , ) from its package directory, which makes it hard to write to them properly on a NixOS system. These configuration files are the only way to configure certain behaviours of , such as in , which is an important setting for device power saving. It's not possible to specify these options in . And it's not possible to specify the path of at runtime. Steps to reproduce the behavior: a file, either manually or from a nix configuration. with . hasn't loaded any settings from . (It's easy to test if you fill with trash since it should log: ). It should be possible to configure all aspects of in the same way. BlueZ loads from a hardcoded path: The same is try for : The define is set to the package directory by default. For a symlink is setup, however no such link is setup for or : I wonder if a fix for this might be adding to the of BlueZ? There's no maintainer listed in , so I'm pinging the people who last touched it :) , Please run and paste the result.\nI feel like this could be easily fixed rather in the by adding 2 more options for and under (?) Let me see if I can carve out some time later this week..\nFixed in"}
{"_id":"q-en-nixpkgs-005898a1451932e4836484b7b3fd2a63690e2c45b214d38fab78d252b4711033","text":"Build PHP on non-NixOS. Both PHP 5.6 and PHP 7 fail to build for the same reason. System: Debian Nixpkgs version: 16.09\nApparently, LDAP_xxx variables are ignored: This is output by Near\nDamn, this is because I have LDAP on the host system.\nOk? This also helps PHP forget host and probably other stuff.\nAre you building with sandboxing disabled? If so, I think this kind of stuff is to be expected.\nI know this is kind of old, but it looks like the Travis build for failed due to this issue. Any reason that patch hasn't been applied? https://travis-"}
{"_id":"q-en-nixpkgs-0f1e466c134f6f389618a3c75803dd9e77a4837df0397ad8360d6f440d4e9618","text":"I would like to use the dolibarr module without nginx. I plan to use caddy instead. In my NixOS configuration, I have . All the nginx-related settings in the dolibarr module are ignored if which is good. However, this line is not: I get the following output: This happens with simply adding to my config.\nOpened for the fix, as I do not use any other web server beyond NGINX, I have no plan to maintain advanced integrations of other web servers, if you are interested into this, I will need another co-maintainer :)."}
{"_id":"q-en-nixpkgs-be083a6eed9749a5ae5adc0584763bc9d4c015baffef4c30e9dea57d89093652","text":"Upstream issue: Since the issue seems to be specific to NixOS, let's track it here. The two commands above should only print \"True\". If applicable, add screenshots to help explain your problem. Tested on xonsh 0.13.3 and 0.13.4 system: host os: multi-user?: sandbox: version: nixpkgs:\nUpstream fix:"}
{"_id":"q-en-nixpkgs-84d465d0df59255f22e09ecbfa9b74b981c10588cddfa6328fe93b3fb3edc934","text":"The build fails in the test phase when CUDA is enabled with errors like this: Note that the did not have this problem. Steps to reproduce the behavior:\nI've played around a little bit and couldn't get it working. The problems is that the CUDA drivers are available at runtime in but not in the cudatoolkit. I've symlinked and then the library is found but CUDA is not running properly in the sandbox (). I've looked around somewhat and neither Torch, nor Tensorflow, nor Numba, ... have CUDA tests running in their respective expressions. I would assume for exactly this reasons. I suggest to just disable the CUDA tests and have them running for CPU only.\nThat sounds reasonable."}
{"_id":"q-en-nixpkgs-d6d949422e9a21a7a3a3bab756cbc64ac67a4253b83b1f2306ea9d5e4d105e3b","text":"Configuring airscan by writing to via silently does nothing, i. e. the file still has the default content. Steps to reproduce the behavior: airscan to sane backends: airscan, e. g. specific devices: After (or & reboot, etc) should contain This stopped working when upgrading from 22.05 to 22.11. I suspect the change to the module in is the root cause. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --I didn't find a entry in the sane module, so excuse me pinging the last few people who touched the module / were involved in the discussion of : Please run and paste the result.\nOh, sorry about that! I guess we can either point to somewhere else but , or figure out how to override using instead of ? As to why your config is ignored, does perhaps get ignored when is set? If so, that should probably be fixed.\nSince this worked before, doesn't that mean airscan isn't using , and instead hardcodes ?\nUsing both and results in an error/warning at build time, but it doesn't prevent successful activation: Btw, you wrote (wrong order) in the first post :-)\nYou can use this workaround for now:\nAccording to it's using the enviroment variable first, then falling back to . I guess that means in nixos 22.05 it was reading both the found in $SANECONFIGDIR (/nix/store/...) and then also I guess we can bring back the old behaviour by just pointing to or something instead of . Thoughts?\nI made\nOops, I should start to copy paste this option, I always mess up the order That is a lot cleaner than what I came up with. Thanks. Wow, that was fast. Again, many thanks to you :)"}
{"_id":"q-en-nixpkgs-9434b5542a9e8fce4542dcbf3bddb0f1590a2033af518d30b6f646f94a076a4c","text":"Building this package twice does not produce the bit-by-bit identical result each time, making it harder to detect CI breaches. You can read more about this at https://reproducible- . Fixing bit-by-bit reproducibility also has additional advantages, such as avoiding hard-to-reproduce bugs, making content-addressed storage more effective and reducing rebuilds in such systems. You can use to analyze the differences in the output of the two builds. To view the build log of the build that produced the artifact in the binary cache: Looks like symbol ordering differences:"}
{"_id":"q-en-nixpkgs-1b8352de98e93b99aad7764ff7a8f6bb4f2aa3094a6015fe0143fe7f339ca1e7","text":"Since which updated from 1.9.2 to 1.10.2, the scheduler service can't be started, the error reported in the logs is: Steps to reproduce the behavior: or to see error as above in the logs The service starts succesfully <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Module maintainers: Package maintainers: Please run and paste the result.\nProbably fixed by"}
{"_id":"q-en-nixpkgs-026ea8ca8a162def863ee871975c11bfe5a612f41ed0c3f078858bf62f5bc3f7","text":"A clear and concise description of what the bug is. Prism Launcher doesn't allow enabling Feral GameMode or MangoHud under the \"Performance\" tab because it can't find it. Steps to reproduce the behavior: GameMode ( ) MangoHud ( ) Prism Launcher ( ) new instance in Prism Launcher instance and go to \"Performance\" tab A clear and concise description of what you expected to happen. GameMode and MangoHud should be able to be enabled when installed. If applicable, add screenshots to help explain your problem. ! ! Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nAddressed here -"}
{"_id":"q-en-nixpkgs-ca93aa37f4c8d6a59b39f8bbcdcf53cbac25f1ef150fadd57b162e22986f0a4f","text":"does not running when you edit . Steps to reproduce the behavior: JSON value of Running service with my setting. ! Left is generated by nixos, Right is generated by supergfxd and edit me. ! I can run without and can edit conf generated by . However, I cannot run with same setting. (edit ) I use zephyrus G14. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nPlease test\nComplete ! Thank you !"}
{"_id":"q-en-nixpkgs-901b880b0b79d592085fa9d191e8205a888d94aaddf43ab28835483e7ae729b0","text":"$ nixops --version Traceback (most recent call last): File \"/nix/store/ni55r1wq3wp907bii23wzxnji0myd296-nixops-\", line 5, in <modulefrom nixops import deployment File \"/nix/store/ni55r1wq3wp907bii23wzxnji0myd296-nixops-1.4/lib/python2.7/site-\", line 15, in <moduleimport nixops.statefile File \"/nix/store/ni55r1wq3wp907bii23wzxnji0myd296-nixops-1.4/lib/python2.7/site-\", line 6, in <moduleimport sqlite3 File \"/nix/store/rawxj3af8l2sbxv6kpclwx5f1gdxi631-python-\", line 24, in <modulefrom dbapi2 import File \"/nix/store/rawxj3af8l2sbxv6kpclwx5f1gdxi631-python-\", line 28, in <modulefrom sqlite3 import ImportError: No module named sqlite3 System: (NixOS: 16.09.1149. (Flounder)) Nix version: (1.11.4) Nixpkgs version: (16.09.1149.)\ncopy\nLooks like it cannot find python C extensions for sqlite3. Looking at the nix expression for the package it should actually use pysqlite instead of sqlite from python stdlib.\nIt seems pysqlite is broken: But pythonPackage.sqlite3 is ok. If you replace the pysqlite dep by sqlite3 in nixops expression, nixops works again. Sorry, I don't really know why pysqlite is borken...\nFixed\nThanks!"}
{"_id":"q-en-nixpkgs-23e798068f393fb1e95199f5cc1a6529377c8eca9202a602a671608f22ab32a2","text":"This means nearly all qt6 packages depend on gdb which is 641M on its own. Why is that the case? I can't imagine needing Valgrind for \"production\" qt6 builds but I have no knowledge in this domain. cc\nvalgrind was not in my qt6 draft in valgrind was in but valgrind or gdb dont show up in the closure of <details </summaryto compare, </detailsso ... no problem? edit: <details </summary</details\nThe inputs are only propagated in the output and that's what other packages use. I only came across this because I found gdb in my system closure. showed <the package-quazip -qtbase -valgrind -gdb why did you add valgrind to qt6's propagatedBuildInputs?\nIt should have been a build time dependency for qtbase. And as it is only referenced in dev output, it should normally not be referenced in system closures, except for issues like Edit, quazip wasn't fixed by that, let me look into it tomorrow.\n... so should not be there &rarr; /nix/store/y1m61ayisqmp29ckzi0qq4fp640qfl3l-qtbase-6.4.1-dev/include/QtTest/qtestmouse.h &rarr; /nix/store/y1m61ayisqmp29ckzi0qq4fp640qfl3l-qtbase-6.4.1-dev/include/QtTest/qtestmouse.h &rarr; propagatedBuildInputs &rarr; propagatedBuildInputs also may be relevant: in , are copied to\nFor the , they should have been in dev output, but and do not have dev output, adding it back should be enough. For , it should be of the same problem as in , maybe we should add that in the hooks of qtbase, or we can fix this ad-hoc."}
{"_id":"q-en-nixpkgs-96317392179cedd7509522722c8a561edb7c845c46121fc072993830fcc7d0da","text":"Steps to reproduce the behavior: sure that the system is set up for cross-compiling (extra-platforms). The following patch fixes the cross-build I am looking for (but it might break other builds, I don't know): Issue seems to be that the configure script of libvpx does not know the target (as a result, we need to force it), which results in a mess. By using instead of , the target is understood and everything works for me. Please run and paste the result.\nI'm experiencing this issue as well, I'll test your patch in ExpidusOS and I'll reply back with whether or not it works.\nAfter applying your patch, I wasn't able to build it. I'm going to look into what wants it to be so it could compile.\nI was able to successfully cross compile , I'll work on bundling into a PR later.\nstill an issue, and it seems like the fix for armv7l really is as simple as described in this bug report (i.e. change the target flag from \"armv7l-linux-gcc\" to \"armv7-linux-gcc\"). if you get a chance to upstream that patch, i'd be happy to test it for cross aarch64-multiplatform and armv7l-hf-multiplatform.\nDone. Fixed by ."}
{"_id":"q-en-nixpkgs-79f875b50dacc0288989a2ae7f86d71a8d0cda3b16798d46aaed6d2b41c26db2","text":"Steps to reproduce the behavior: , set the following: Then navigate to a git repo (e.g. nixpkgs) and call Git logs and a success message Nushell prints the following error: <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nI have proposed a fix in"}
{"_id":"q-en-nixpkgs-961bfc40141fd09a2ab771ea6436f9895a2ca6ce3dc0f98ddd43d8b1b3842a22","text":"This issue was described by kvtb at Quoting: Extende the ExecStartPre with something like: Currently the unit depends on , maybe there is something smarter to do.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nmight be more suitable?"}
{"_id":"q-en-nixpkgs-1a19491626eae84e9db7787f519264ccd6b2d088571aeae34fb78f641cfda51e","text":"I'm trying to install ncdu on darwin, which depends on zig; it has started failing to build in recent nixpkgs on the 22.11 track, though. Steps to reproduce the behavior: zig installs cleanly. <details<summaryoutput of nix log</summary<p</p</details Please run and paste the result."}
{"_id":"q-en-nixpkgs-2934d96dbaf26a7acc069d03738354a71287ab05dd7369ab0037cac481299e93","text":"NixOS 16.09:\nMerged to 16.09 in , hope it works for you. I haven't exactly tested this extensively. It builds and the test programs it produces pop up windows :-)\nThanks! For my edification, would you mind pointing out how you knew that needed to change to ? I was trying to PR this myself but never figured that out.\nWell, I first tried just updating the version, but as you also noticed it wouldn't build. In the build output it wrote however, and this certainly makes multiple-outputs a prime suspect. So I did and modified to also have with the include path in the output. Note the one is still there. After that I wasted some time trying to get Tk to use the system libjpeg but gave up on that after a while :-)"}
{"_id":"q-en-nixpkgs-f713cf59899f7ff782797f8307ae97bc5b4855f1628bd9dbac04cdb01d8d4f37","text":"The icu541 package fails to compile at the patching step. Some of the patches issue warnings, others fail completely. Install icu541: . The build process will fail. System: NixOS: 16.09 Nix version: 1.11.4 Nixpkgs version: 16.09.1202."}
{"_id":"q-en-nixpkgs-e55c56cb7361d875d52b4cc0e808a897654e2f5aba5157fe978f79ba110542f2","text":"Once in several days, my X session crashes. It always happens at the time systemd-tmpfiles- runs. Here is the corresponding part of [ .525] (WW) Open ACPI failed () (Connection refused) [ .994] (II) config/udev: removing device Power Button [ .006] (II) evdev: Power Button: Close [ .006] (II) UnloadModule: \"evdev\" [ .006] (II) config/udev: Adding input device Power Button (/dev/input/event4) [ .006] () Power Button: Applying InputClass \"evdev keyboard catchall\" [ .006] () Power Button: Applying InputClass \"Keyboard catchall\" [ .006] (II) Using input driver 'evdev' for 'Power Button' [ .006] () Power Button: always reports core events [ .006] () evdev: Power Button: Device: \"/dev/input/event4\" [ .006] (--) evdev: Power Button: Vendor 0 Product 0x1 [ .006] (--) evdev: Power Button: Found keys [ .006] (II) evdev: Power Button: Configuring as keyboard [ .006] () Option \"configinfo\" \"udev:/sys/devices/LNXSYSTM:00/LNXPWRBN:00/input/input11/event4\" [ .006] (II) XINPUT: Adding extended input device \"Power Button\" (type: KEYBOARD, id 6) [ .006] () Option \"xkbrules\" \"evdev\" [ .006] () Option \"xkbmodel\" \"pc104\" [ .006] () Option \"xkblayout\" \"us\" [ .006] () Option \"xkbvariant\" \"intl\" [ .006] () Option \"xkboptions\" \"ctrl:nocaps\" [ .770] (II) Open ACPI successful () [ .770] (II) config/udev: removing device Video Bus [ .775] (II) evdev: Video Bus: Close [ .775] (II) UnloadModule: \"evdev\" [ .775] (II) config/udev: Adding input device Video Bus (/dev/input/event15) [ .775] () Video Bus: Applying InputClass \"evdev keyboard catchall\" [ .775] () Video Bus: Applying InputClass \"Keyboard catchall\" [ .775] (II) Using input driver 'evdev' for 'Video Bus' [ .775] () Video Bus: always reports core events [ .775] () evdev: Video Bus: Device: \"/dev/input/event15\" [ .775] (--) evdev: Video Bus: Vendor 0 Product 0x6 [ .775] (--) evdev: Video Bus: Found keys [ .775] (II) evdev: Video Bus: Configuring as keyboard [ .775] () Option \"configinfo\" \"udev:/sys/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/device:09/LNXVIDEO:01/input/input23/event15\" [ .775] (II) XINPUT: Adding extended input device \"Video Bus\" (type: KEYBOARD, id 8) [ .775] () Option \"xkbrules\" \"evdev\" [ .775] () Option \"xkbmodel\" \"pc104\" [ .775] () Option \"xkblayout\" \"us\" [ .775] () Option \"xkbvariant\" \"intl\" [ .775] () Option \"xkboptions\" \"ctrl:nocaps\" [ .839] (II) config/udev: removing GPU device /sys/devices/pci0000:00/0000:00:02.0/drm/card0 /dev/dri/card0 [ .839] xf86: remove device 1 /sys/devices/pci0000:00/0000:00:02.0/drm/card0 [ .839] failed to find screen to remove [ .935] (II) config/udev: removing device Video Bus [ .941] (II) evdev: Video Bus: Close [ .941] (II) UnloadModule: \"evdev\" [ .941] (II) config/udev: Adding input device Video Bus (/dev/input/event14) [ .941] () Video Bus: Applying InputClass \"evdev keyboard catchall\" [ .941] () Video Bus: Applying InputClass \"Keyboard catchall\" [ .941] (II) Using input driver 'evdev' for 'Video Bus' [ .941] () Video Bus: always reports core events [ .941] () evdev: Video Bus: Device: \"/dev/input/event14\" [ .941] (--) evdev: Video Bus: Vendor 0 Product 0x6 [ .941] (--) evdev: Video Bus: Found keys [ .941] (II) evdev: Video Bus: Configuring as keyboard [ .941] () Option \"configinfo\" \"udev:/sys/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/LNXVIDEO:00/input/input22/event14\" [ .941] (II) XINPUT: Adding extended input device \"Video Bus\" (type: KEYBOARD, id 7) [ .941] () Option \"xkbrules\" \"evdev\" [ .941] () Option \"xkbmodel\" \"pc104\" [ .941] () Option \"xkblayout\" \"us\" [ .941] () Option \"xkbvariant\" \"intl\" [ .941] () Option \"xkboptions\" \"ctrl:nocaps\" [ .942] (II) config/udev: Adding input device Lid Switch (/dev/input/event2) [ .942] (II) No input driver specified, ignoring this device. [ .942] (II) This device may have been with another device file. [ .942] (II) config/udev: removing GPU device /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card1 /dev/dri/card1 [ .942] xf86: remove device 0 /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card1 [ .963] (EE) [ .963] (EE) Backtrace: [ .972] (EE) 0: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (OsSigHandler+0x29) [0x5954c9] [ .972] (EE) 1: /nix/store/jm1n87rp8vr90j9ahcrfzr57nc2r8vgf-glibc- (restorert+0x0) [0x7f27cfc91e8f] [ .990] (EE) 2: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server- (FreeScreen+0x23) [0x7f27c4df9413] [ .993] (EE) 3: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (xf86DeleteScreen+0x5a) [0x481d4a] [ .993] (EE) 4: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (xf86platformRemoveDevice+0x11e) [0x49612e] [ .994] (EE) 5: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (DeleteGPUDeviceRequest+0x9b) [0x49bb4b] [ .994] (EE) 6: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (deviceremoved+0xdf) [0x49785f] [ .994] (EE) 7: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (wakeuphandler+0x90) [0x498460] [ .994] (EE) 8: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (WakeupHandler+0x5c) [0x43c07c] [ .994] (EE) 9: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (WaitForSomething+0x1f6) [0x58dd86] [ .994] (EE) 10: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (Dispatch+0x8e) [0x43734e] [ .994] (EE) 11: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (dixmain+0x3c3) [0x43b543] [ .995] (EE) 12: /nix/store/jm1n87rp8vr90j9ahcrfzr57nc2r8vgf-glibc- (libcstartmain+0xf0) [0x7f27cf0a92e0] [ .995] (EE) 13: /nix/store/0195lcgvy4g6vihbfm4qqfa9cckkpcfa-xorg-server-1.18.3/bin/X (start+0x2a) [0x4258ba] [ .014] (EE) 14: ? (?+0x2a) [0x2a] [ .014] (EE) [ .014] (EE) Segmentation fault at address 0x8 [ .014] (EE) Fatal server error: [ .014] (EE) Caught signal 11 (Segmentation fault). Server aborting [ .014] (EE) [ .014] (EE) Please consult the The X.Org Foundation support at for help. [ .014] (EE) Please also check the log file at \"\" for additional information. [ .014] (EE) I can't reproduce the crash, it happens randomly. System: 16.09.1229. (Flounder) Nix version: nix-env (Nix) 1.11.4 Nixpkgs version: \"16.09.1229.\"\nI wonder if we'd better pick Xserver bump to 16.09: .\nHmm, I picked it, whether it fixes this problem you experience or not. I expect it to hit the channel within a couple days.\nThese lines seem suspicious: .\nI have a computer in my living room with nixOS 16.09 on it. Since about a month every time I close the screen (the TV), the X server close itself. It does not segfault. But I have the same suspicious line in X11's logs. Any idea what cause the bit? Obviously, in my case, the \"shotdowning\" of the screen triggers that, but why? It might be related to what have.\nMaybe its X11 trying to reconfigure itself. Maybe providing a configuration file would prevent it.\nOn one Intel-based desktop I was occasionally getting into an irrecoverable black-screen state on some similar conditions (all but the screen seemed OK), but I don't have any logs of that anymore.\nSorry for the spam, but I forgot to say that I didn't have this problem when I first installed nixOS 16.09 on this computer (fresh install), I installed it right before the official release of 16.09. The problem started to occurred after an upgrade, but I can't tell exactly when.\nMy computer too uses Intel graphic.\nMaybe a different kernel? Does this occur on all the kernel versions currently available in release-16.09?\nI've got 4.4.x all the time on that machine.\n(triage) does this still happen?\nI don't know, I don't use nixOS anymore.\nI am closing then. If anyone experiences this issue, I will be glad to reopen."}
{"_id":"q-en-nixpkgs-f4c29418f375f61ccac9ad4fda23e410a54b1f25c62db9357a1ee524259d0ad4","text":"Project description From the : Metadata homepage URL: source URL: license: MIT platforms: linux"}
{"_id":"q-en-nixpkgs-5ae364f57be645b640dfd8bebc52088d5a0e6670bee0c2d453a36ba805ca5cfe","text":"Seems to be related to PR Steps to reproduce the behavior: from master branch () the result share directory () the state of the man link () This will result into build error if used on nixos system with enabled man pages last line of trying to rebuild the system Either working link to man page or at non at all."}
{"_id":"q-en-nixpkgs-1e47499a45409bc7fca01d69930e24a6d1f4d3352a4e48901ff3264e65d56beb","text":"First failure in Likely due to\nThe referenced file is indeed missing from source: Our fetcher should include submodules.\nChanged in , where is converted into a submodule.\nThese qt modules currently include submodules:"}
{"_id":"q-en-nixpkgs-fc632ad4fa217be68f5d8a2add39cd87994f69400c3939cefa93b036eea4aa21","text":"Steps to reproduce the behavior: i noticed the break after the nheko 0.10.2 -0.11.0 upgrade landed in nixos-unstable: hydra failure (click for the build log): build is currently broken in both master () and has been fixed in nixos-22.11, see for having initiated the backport. tagging as maintainer of both mtxclient and nheko. Please run and paste the result.\ntested against the outstanding mtxclient 0.9.0 -0.9.1 PR: still broken.\nI have no idea why a standard library would depend, let alone be missing, on a different architecture.\nLikely due to us using outdated gcc (gcc9) on aarch64-linux. The story is complicated, but is being addressed.\nForcing it to use gcc11Stdenv may fix the issue.\nIt seems so: I tested by cross-compiling mtxclient (I don't have an aarch64 system).\nOpened\nGreat... now coeurl fails to build. EDIT: Fixed, but nheko is still broken.\nhope nobody minds this extra notification that is me thanking you all for jumping in so readily to fix things, even for someone you’ve never met. anyone reading this: what you’re doing here has more impact than you know."}
{"_id":"q-en-nixpkgs-5da2a99ea684e0c8911a432b7e824e5393b435e84c2d18e81866e627a236a3fb","text":"After enabling with (CLI and GUI) I am unable to use any other protocol except wireguard Steps to reproduce the behavior: and related service from channel , switch to openvpn, try connecting All protocols should work Clearly lacks some files: Probably needs something similar to whatever is done in (CLI) package:\nworks as expected, but is broken. We can wrap MULLVADRESOURCEDIR for in like in . I''ll make a quick fix. Maybe I don't understand something, but why do we have two mullvad packages? cli and gui are both from the same repository, maybe we can merge this two packages and make some aliases for those, who wants to install cli and gui separately?"}
{"_id":"q-en-nixpkgs-0be168a41b7ab5b36bab0deda39f40dfce509bea936c536244880c2079f19e81","text":"Steps to reproduce the behavior: libreoffice-fresh off master Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nIs this issue closed because it is solved? It's still happening to me: If I include in my config, fails with the message . If I exclude from my config, rebuilding succeeds. I'm using the channel.\nYes, it's just that the commit hasn't arrived yet at nixos-unstable, see ."}
{"_id":"q-en-nixpkgs-0f35545f2e1df14464d30c3f360bc412135f5a89a15a9eb0ba435dac7c08467a","text":"Recently, we fixed : However, on , still fails to run on my machine throughing the dreaded: Steps to reproduce the behavior: run nixpkgs#puddletag The application seems to be correctly wrapped: but still, it does not work. and myself. Please run and paste the result.\nThere’s definitely an impurity somewhere - it’s working perfectly fine here. I hadn’t used it in a while, but it runs with nixos-unstable` currently on this machine.\nThis is interesting. The command definitely fails for me. (I was incorrectly saying , but I am also using ). I have trouble debugging this. I had a look at the wrapper shell script, and all looks fine. The paths are ok and do exist.\nThat seems to indicate an impurity then.\nThis is also happening on nixos-22.11, not just unstable. A friend is having this issue, so I tested it out for myself and got the same thing. For what it's worth, we both use i3, so the impurity may be related to things a DE would normally set up.\nruns fine for me.\nSo definitely a purity thing then. Care to do a PR?\nSure, I suppose. Just a minute.\nHi! Thanks for looking into this! I could not test the PR directly because the review process starts to build Python Pandas an whatnot. However, I could start using the fix of Some remarks: The wrapper without the fix is: The file with the fix (adding the Qt wrapper line): So there is a substantial difference, and wrapping doesn't seem to work properly without the extra line, even though it should work. Further, when opening on using the fix, the program crashes when clicking with the following error: I do not think this is related to the wrapping, but it should definitely also be fixed. Somehow, is not found.\nMy version wraps twice, so the original wrapper and all its settings are still there, in , while the real puddletag script is now in . There's a note at the bottom of which says that doesn't wrap scripts, only ELF binaries. I don't know what the reasoning for that is, but apparently invoking the wrapping manually is \"what you're supposed to do\". At least as far as I could tell.\nOk I see. So the first wrapper handles the python dependencies, and the second one the qt dependencies. Is that so?\nCounting the layers outward from the core (rather than, say, in execution order), yes, that's correct.\nThen we should modify the wrapper args instead to avoid double-wrapping.\nWell, to do that, I believe you'd need to reimplement either the python wrapping or the qt wrapping systems, which seems prone to breakage."}
{"_id":"q-en-nixpkgs-2e2b554f2e4b83ba1efa4106cba4dcad3af88f9e66bf663b7bbff6e81a0e257d","text":"Project description RiscPC Emulator Metadata homepage URL: source URL: license: gpl2 platforms: linux, darwin This also includes ROM images which could also be downloaded, as ROOL maintains these on an open source licence. Not sure the best idea for these... /var/lib? Thanks very much for all your hard work.\nBuilds and compiles. I will upload it tomorrow\nThe roms are mostly optional, in the sense the software will seek them in . There is no such a need for include them now.\nThanks very much.\nhow did you get it to work? I can't get it to recognize the existence of the hostfs folder."}
{"_id":"q-en-nixpkgs-6a40c61dc97cd45b19ce1f0f1c2b026e93d21fdacccb88f4deff70985ff880b2","text":"Project description spotify-player is a fast, easy to use, and configurable terminal music player. Metadata homepage URL: source URL: license: MIT platforms: linux, windows, darwin"}
{"_id":"q-en-nixpkgs-c7293d0ca6d565294718596b13907975047009c687af9662a36b176519286d83","text":"On or cross-compiled from , breaks: From : You can either run this on another architecture using . or should not crash: This was responsible for failing to replicate datasets to my ARM NAS. The issue does not_ occur for natively compiled builds. It appears to be a cross-compilation issue only.\nIt is a bug in / derivations. In non-cross cases it gets copied from used to bid into output: But in cross-case used to build does not yet provide as it needs bits of for it (and the library path is slightly incorrect). One of possible workarounds is to try to add to to get it pulled via from expected location.\nWe don't do the copying to glibc anymore, but this error still occurs as described on current nixpkgs master.\nYup. cross-compilation never propagated to in any form (and does not today).\nCould you please try this PR? It should fix the problem.\nWith that PR, I'm seeing this on both (a Raspberry Pi CM4) and with QEMU on : Log from aarch64 machine: Log from machine with QEMU:\nWell it's finding now, so there's clearly some other problem with addition to that. When I run it with I even get line numbers from inside : That line () is the last line in the following snippet of code: ... so is deliberately ing here due to . Grepping through the docs turns up: This is happening because uses . Once you do that, glibc needs to unwind the stack frame of any thread that calls . And unwinders are a giant headache. I'm afraid that's all I've got here. I've removed the closed-by from since it was only part of the problem here.\nAh, I think I know what it is. We can compile a without , but it won't have an unwinder. We need to pass the glibc headers to the build in order to get one of those.\nwould you please try ? I believe it fixes one of the two problems which caused this issue, and it includes which fixes the other one. It works for me: BTW, this issue (cross-compile , then run it under qemu) is a really valuable test case. We really ought to add it to so Hydra will catch any regressions automatically. Especially since the failure mode gives the user no error message to help them troubleshoot the problem! is a great candidate for this since it's one of those very rare non-C++ programs that uses advanced stack-unwinding features from pthreads.\nConfirming fixes the issue. Tested with both and on real hardware and with QEMU on x86_64.\n~I think this is fixed, and more cleanly,~ by Edit: it is not. :frowning_face: I will keep trying. Test is in Currently building.\nAdded this to , but commented out since it is not fixed yet."}
{"_id":"q-en-nixpkgs-83b52cc01533a67f391856357cbb00682dfbfc1e86607f18b8c6090ba32ff564","text":"<!-- Note that these are hard requirements --<!-- You can use the \"Go to file\" functionality on GitHub to find the package Then you can go to the history for this package Find the latest \"packagename: oldversion -newversion\" commit The \"newversion\" is the current version of the package --- <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- name: msmtp <!-- The current version can be found easily with the same process as above for checking the master branch If an open PR is present for the package, take this version as the current one and link to the PR --current version: 1.8.22 desired version: 1.8.23 <!-- Search your package here: If no maintainer is listed for your package, tag the person that last updated the package --maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-67c70330c259926f5baffc1716a870a07b6286aba920287f2e6d1b86e51a43e7","text":"build Full logs: Errors happen when building for musl, dynamic. Alpine package: Nixpkgs package:\nThis can be fixed by a , but I'm not sure if we want this?"}
{"_id":"q-en-nixpkgs-056bc6df11253d1f80563edb92c156cf5310fc83afc58d4827906a7d5a41630b","text":"I want to change a configuration variable to onlyoffice. Currently, the NixOS module adapts on the fly in the upstream configuration with a call. I'd like to make my own adaptation. For now it's not possible due to ExecStartPre being not declared as a list, but as a direct value."}
{"_id":"q-en-nixpkgs-6dde6867bf71d1c0628cd8d8cd7b69f4a291b5f3bad9099cff017e6f4ee8d4f8","text":"SpamAssassin generates and compiles C code from anti-spam rules. This currently fails because the generated code makes use of the function, which is not available during the rule compilation process. Steps to reproduce the behavior: the SpamAssassin service that sa-update.service fails with the error message: that, consequently, spamd.service will not function. Alternatively, the error can also be observed by invoking directly. sa-update.service succeeds <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nA smoke-test that serves as a reproducer is available in .\nIn theory, this could be addressed by making libxcrypt available when compiling rules, but care must probably also be taken so that the dynamic linker can find it at rule run-time.\nRule compilation seems to complete without errors for the rule if we set . However, of course the resulting binary is not linked against libxcrypt then. I’m not sure when Spamassassin will generate code that actually calls functions from that library."}
{"_id":"q-en-nixpkgs-a3862f47b58f1ea32fa8709f9c242743f0a82da095fbf0b05d5b864f0c975b62","text":"Building uses more () threads than it should (i.e. 2 if / ). Steps to reproduce the behavior: as the build will use way more than 2 cores. Build should honour properly. This was with . This does not in fact use just 1 thread. :smilingfacewith_tear:. ! Happens on master () and 22.11.\nI don't have any time planned to fix this but if you figure it out please send a PR."}
{"_id":"q-en-nixpkgs-7878269a2a9c76e4731585fd39fd1205274af45e7b8845c3858cbd3ab33e2fd7","text":"According to documentation of the app: there's a dozen dependencies we don't explicitly have a dependency on, so there's multitude of bugs that might appear depending on what's on $PATH.\nDo we need a new maintainer of to move this issue forward?\nWe should pull the package from NixOS core...absolutely no reason to have such a bloated less ship by default."}
{"_id":"q-en-nixpkgs-999d508c6133dcea5eaa50875e54db8dcd135f6a3d42e90c9fd6967559d11283","text":"LibreSSL 3.4 is no longer supported upstream. It became unsupported on October 14th 2022, one year after . At the time of writing, 3.5 and 3.6 are both supported. that fixes a serious vulnerability: I suspect (but haven’t verified) that this vulnerability is present in LibreSSL 3.4. It is time to remove it from Nixpkgs. However, there are some packages that still depend on it. all depend on . I tried building against and , but both fail. I don’t see any packages that depend on , so I think we could remove the older versions, if we can first update FoundationDB to a version that is compatible with a more recent LibreSSL release. Maybe will be sufficient for this. Alternatively, we might patch FoundationDB. The compilation errors are about an incomplete type for which only a forward declaration is known. I haven’t looked into it deeply but it may be easy to fix. as the maintainers of , what are your thoughts on this? This has been addressed by . <details<summaryOriginal comment</summaryThe other package that explicitly depends on LibreSSL 3.4 is . This is a proprietary package so I couldn’t test if it is compatible with a more recent LibreSSL release, it might actually be. as the maintainers of , could you try and see if it is compatible with a newer LibreSSL release? </details\nThe x400Address issue that was patched recently is present in LibreSSL 3.4. The fix is very small and applies to all older versions that have it, so it might be an option [patch it][1] while you sort out the other issues. If you could share build logs for the foundationdb issues I can take a look if you want. [1]:\nYou can drop this. uses their bundled OpenSSL and drops the dependency since it's incompatible with even this LibreSSL version. It's about to merge.\nall of darwin depends on 34 at the moment. You've pinned to 3.4, do you know if by any chance we can upgrade to 3.5 ? maybe ?\nI'm quite sure we can't. My history with that package is and . happened, because someone did . iirc happened, because this package depends on ancient long unsupported openssl APIs that weren't part of anymore.\nSmall rectification, after a bit more testing, it doesn't appear to be all of darwin that is impacted by that change. So perhaps we don't need to put too much effort into this. I'm still trying to figure out the impact.\nI’m working on a fix for , which drops the dependency on from by unconditionally dropping Unbound from the build (instead of only on aarch64-darwin). Apple doesn’t even include Unbound in macOS anymore, and it’s not present in newer source releases. That should remove as a blocker for dropping LibreSSL 3.4."}
{"_id":"q-en-nixpkgs-cb476cd482dd65cde02be0fffbb2c2be79c7380493e40890a1b6b52947b63938","text":"I wanted to setup unbound on a server. I did have this configuration: I then manually created the file using in the folder as it did not automagically create one. The folder was set to 0777 permissions and the key to 0666. I checked then the systemd status: All looked fine. Port 5335 also opened on the firewall. Basically running PiHole container there that should call unbound for recursive resolution. As systemd showed unbound running, I tried to do a query but it errored out: Interestingly I run unbound on another server. On the other server I did not set the as it does not wipe everything that's not explicitely saved (see ). I thought I would need the on this setup, as everything gets wiped on reboot that is not explicitely set to save. After removing the option from the configuration, unbound works as expected. So there are two problem with the option: does not get automagically created don't work system: host os: multi-user?: sandbox: version: channels(root): nixpkgs:"}
{"_id":"q-en-nixpkgs-02d1025b112e254345efaa5735b6d3cdedb2779fb93cf7e88e5c6af7563c8d4f","text":"p7zip source is gone try to build without substituters cc\nWhat commit are you having the issue from? it works for me at\nCurrent master, that is sets to . There is no such tag at\nMaybe the fork cleared the tags by force pushing?\nHmm, I wonder if all the changes got merged into with the fork then rebasing and force-pushing…\nYeah, looks like it. A bunch of jinfeihan57's commits are now in the commit history I'll change it to point back to p7zip-project's copy now"}
{"_id":"q-en-nixpkgs-477d3936661b8240fd1eafd1599c585b2b65e1c5d7a7ef4f6ed241bab5403d44","text":"Project description CryptPad is a collaborative office suite that is end-to-end encrypted and open-source. Enable this a service for NixOS. Metadata homepage URL: source URL: license: agpl2 platforms: unix, linux, darwin, ...\nCryptPad used to be included in nixpkgs, but it was removed because there weren't enough maintainers. The removal occurred with the commits and .\nNot quite true: it was removed because they had a strong dependency on out-of-date software. . Now, there is an ongoing effort to re-introduce it in : ."}
{"_id":"q-en-nixpkgs-e4f212193185a99bf795feb6e0527e6e6d39f21ca0fa6a385a491540eba66e94","text":"Package name: mathematica Latest released version: 13.2.0 <!-- Search your package here: --- Current version on the unstable channel: 13.2.0 Current version on the stable/release channel: 9.0.0 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR.\nNo need to open an issue for every package update PR that you're about to file."}
{"_id":"q-en-nixpkgs-d9ac6db4b43da5d75135f31fe988dd50a04cc826a089039409e6bd2ad77d06b4","text":"Steps to reproduce the behavior: -env -i gotop Happy to provide more info if needed. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nMaybe we can apply as a patch.\ncan you try latest master ? gotop 4.2.0 should aleady contain patch mentioned by"}
{"_id":"q-en-nixpkgs-fbf7d7cdca90dc4d99ecbff43d45f035d214dbb4ec6fa82100269c3951c6c952","text":"Linux app for streaming music from YouTube. <!-- Monophony is a free and open source Linux app for streaming music from YouTube. It has no ads and does not require an account. --Metadata homepage URL: source URL: license: mit platforms: linux\nI'd be interested in doing this if its not too hard, I'll look into it.\nThe package is almost done, had some complications including unpackaged dependencies, but it should be good to go within the next few days.\nJust waiting for a merge now."}
{"_id":"q-en-nixpkgs-17390bfb44b6bd5288e597ed528fc7763c172ee020f233e7f063ae40c8eb4680","text":"I'm trying to do something similar to the following: Apparently, however, does not translate the Nix list into a Bash array as I'd expect. I see the following error: Add to the attributes which contains a argument with whitespace. For example: This results in a bash array with the following items (one per line): Use the following patch to print the array members: Please run and paste the result.\nUsing flags with spaces as a derivation attribute can . I'd append to in , as explained in the link:\nThanks for your comment. That's frustrating. I wonder if should have an argument at all, if it works sometimes. Also, I think manipulating in does not work as the defined by run first.\nSpaces can be supported if some quoting is to : This is also how it's done in Python's build-support:"}
{"_id":"q-en-nixpkgs-4c9cea3e5a84372fa4be493abfec6145bb94ce0a81e3b3dc79ffcebfff337415","text":"While rebuilding on this system that requires the package a failure while building has been occurring since NixOS 22.11. One of the tests fails.\nI am not the maintainer of this package. Maybe can help? also have a look at:\nSilly me for not checking and just the current release, disabling tests is a hell of a workaround but definitely works!\nI'm also not a maintainer of the package, but I'm happy to fix it nonetheless :) There's an issue in that we don't have a good way of setting HDF5PLUGINPATH appropriately. Maybe I'll come back and fix this at some point, but for now this will do..."}
{"_id":"q-en-nixpkgs-ff90696f9af111ddf69aa3a3a37be6b0b319ac884e4569743b17f4b7df269aa8","text":"Neovim frequently coredump It is hard to reproduce the behavior, it always breaks down in unexpected situations no core dump ! <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nIt might be caused by neo-tree. I haven't seen any crashes since I moved from neo-tree to nvimtree.\nAlso pretty sure it's neo-tree, I have the same issue . I've already built debug symbols for and am waiting for the next crash to get a detailed report.\nnot nixpkgs relevant (report to neovim with stacktrace)."}
{"_id":"q-en-nixpkgs-172fb2ea3dc0f1211498575edb61bcb911e81c3b0ab4e7110fe3cc911e53b863","text":"grub fails with nixos-rebuild switch -I nixpkgs=/home/lethal/nixpkgs\nThe issue title here is highly uninformative, which is why I didn't find this when checking for duplicates. :P"}
{"_id":"q-en-nixpkgs-af760972c467b17bbab1ff27508a6866cd68d654e6b095a8a2bc79b54fbaf42c","text":"When running , sometimes fails to start with an error message: Run System: NixOS: Nix version: Nixpkgs version:\nI can confirm this issue. It also delays restarting of other units (probably dependent services) after upgrade. I do: as a workaround when this happens, so that the other services can restart faster. It only happens on my physical machine. My v-server never had this problem.\nCan confirm this is still happens on 17.09 (just happened right now, and has been happening on every single ).\nMy workaround is to use:\nhow do you specify that? I get:\nthere was a typo: UPDATE 2: One also need to make it a list to override!\nThe solution posted above no longer seems to work after upgrading to 18.09 - has anyone else solved this in a different way?\nRan into this issue when I didn't have the EFI boot partition mounted while rebuilding. I think I was starting up an old nixos configuration. Seems to work fine now.\nI forgot to make it a list. The empty string makes systemd first resets the option so it can be overriden afterwards.\nA potential fix is also here: I want to test this option a bit for myself first."}
{"_id":"q-en-nixpkgs-558cd5cb2fecb24e67d4a530b7db2b2dc1041d0cfc25536a90132233e93a8c9e","text":"See also Steps to reproduce the behavior: This is such as basic cross-compilation that I wonder whether something is wrong in my setup. However, the errors are surprising, because I thought most packages were cross-compilation aware enough to not run host binaries on the build platform. Please run and paste the result.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nwas mentioned as a possible fix.\nI can confirm that fixes this issue."}
{"_id":"q-en-nixpkgs-a91d1dfcb3296a3dae2d311b91cfc030c254498b081224434efa8b60ac292989","text":"From a brief look at the updated cuda section in the manual, it would be logical to take as an argument, rather than the compiled string. And to expose that in via gets even further beyond the scope, so let's skip it Originally posted by in\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-301137099609987318f752585b90e1b2441ebc0469666e1bda8ef2c060411272","text":"The program Jabref will build, but give a JFX related error upon trying to run the program. Steps to reproduce the behavior: The program launches as it used to. Please run and paste the result.\nIt's fixed in but hasn't been merged.\nHi. Sorry for digging this up, but I'm having the same error for starting a JFX application inside Intellij IDEA, where is unable to find (or ). My hack to resolve this involves: Make sure that there is within your ( for example) Add the path to :Start Intellij IDEA from the shell. I felt like this is worth a separate issue (Intellij and JFX related), but wanted to ask it here first. Many thanks!\nI can't understand what you want to ask... This issue is caused by the reason said in\nI meant I had the same issue (missing for ), but in another context, and was wondering instead of that hacky workaround, is there any way to make my JFX application starts with no error."}
{"_id":"q-en-nixpkgs-5cc7162bb502acccc8450750e012e730299908137d4c525fbd60ff737e244541","text":"lbreakouthd fails to launch with \"libGL error: MESA-LOADER: failed to open iris\" Steps to reproduce the behavior: lbreaouthd with lbreakouthd launches Framework laptop with intel graphics. Please run and paste the result.\nkeep an eye on the and tell me if anything goes wrong"}
{"_id":"q-en-nixpkgs-4fed3c1cfd30f7f86278ca1ea8a3756f1b7a70989e7e3ec05be5691fa6c9393a","text":"Since the introduction of in (), derivations should use instead of to further reduce the usage of the attribute.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-8c83b31c119b875274edb78ba54871d156911496212bf3dad8c2ecc7b9c66c6c","text":"NixOS unstable is affected by on Wayland. Steps to reproduce the behavior: signal-desktop on Wayland. No crash. Add any other context about the problem here. Please run and paste the result.\nI am also getting a crash every once in a while on startup. Starting it again works so far and it seems stable after that. Don't know what to do about that.\nI've fixed this locally by using the following overlay:\nApparently, Signal beta works again on Wayland.\nJust adding a data point. Having Wayland disabled actually caused more bugs for me. Re-enabling Wayland via 's overlay helped fix my compositor crashing whenever I clicked on certain UI like \"preferences\" in the \"file\" menu. Currently running: with set to and have the following env variables set:"}
{"_id":"q-en-nixpkgs-d1773559d8744b289617c2e327ac45bd786fb63a60f6b4503d93faebc762ea8a","text":"Steps to reproduce the behavior: Bisected: Introduced in\nFrom the build log: What the heck is that meant to do? This test would pass if the reported size of the TIFF starts with a 7, a 9, or a pipe character, and fails otherwise? Surely this is an upstream gscan2pdf issue.\nIssue reported to the gscan2pdf project: Since this should not be impacting the other features of the program, I think it's safe to temporarily disable that one test for that package on our side.\nDisabling the test temporarily in:\nWith merged, let's close this. There's nothing we can do until upstream reacts.\nApparently upstream did update their test according to , but we need to wait for a new release containing the fix."}
{"_id":"q-en-nixpkgs-525713e4e06448fb4b19c9906fc9ad7c36bb1b4005b994bf2a063a797d2f9281","text":"If gdm is enabled with Gnome enabled, gdm will crash and not load into the login screen after logout or reboot. This is due to gdm using wayland as the default. Once wayland is disabled in gdm, the login works. Steps to reproduce the behavior: channel nixos-unstable, \"sudo nixos-rebuild switch upgrade-all\" with Gnome enabled and gdm enabled in will hang at the gdm check point. gdm wayland in and the reboot works but with Wayland not working. Reboot into gdm with Wayland <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nI just hit the same issue. After updating to latest nixos-unstable, the boot hangs just before login screen I think this is related to , at least the workaround posted there worked for me also\nWould be relevant with a machine without AMD graphics and without KDE? As, I am running Intel's build in graphics and using Gnome. Also, I did find a \"workaround\" by disabling wayland in gdm then nixos-unstable will boot.\nI managed to work around this issue with on GNOME Wayland. Graphic card: Intel Xe Graphics (TGL GT2)\nTotally worked! Thanks and ! Should I leave this open as this is a bug or a new config that needs to be applied to 23.05?"}
{"_id":"q-en-nixpkgs-6da340f0d963257b12e8960afb321234319767d746c189e44de4fd7aaa5e0fbc","text":"Project description VMware OVF Tool is a command-line utility that helps you import and export OVF packages to and from many VMware products. A package already exists for Linux ARM and x86-64. VMWare ships package for macOS but, but support for it is missing from the package. Metadata homepage URL: source URL: license: unfree platforms: darwin\nThe main problem in my case is that I don't have an Apple machine nor somewhere to emulate one, so I might not be able to support this ovftool type.\nWork in progress, doesn't function. I have a few questions on that PR about how to do this and not make it a complete mess."}
{"_id":"q-en-nixpkgs-4f68c3d733b72dbbfa8e2dd5d672264ccbe9e8fed30b9861339d689a3d9e56ff","text":"Package name: organicmaps Latest released version: 2023.04.02-7Current version on the unstable channel: 2023.03.05-5 Current version on the stable/release channel: 2023.01.25-3 Notify maintainers Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-f2434346d5117d991b7e82bde7c1b29f87747267551735a5165b22ae4f0fb380","text":"Building powerpc kernels (I only tested pkgsCross.powernv, but examination of the kernel source code suggests to me that 32-bit powerpc, and well as big-endian 64-bit will be the same) is broken since This happens because on powerpc, the kernel target we install is vmlinux, rather than any derivative of it like zImage. The kernel build system will strip when creating such a derivative image, but because we install vmlinux directly, we get an unstripped image. This means the kernel contains references to its build directory, which since that PR will be the $dev output. AIUI, 64-bit POWER boots via kexec, and a zImage is a self-extracting kernel ELF image, so hopefully the fix would just be to switch the kernel target for powerpc to zImage, and things will work again, and POWER users will get smaller kernel images as a bonus. I'm not sure about 32-bit PowerPC, but this might be helpful. If we can't use an *Image target for any powerpc targets, there is also a target called that we could fall back to, but let's take the opportunity to evaluate whether anything else is more suitable first, when we're changing this stuff anyway. Apart from powerpc, the only we have that uses vmlinux as its kernel target is a MIPS machine called the Fuloong 2.0. I don't know how it boots.\nI'll handle this. I don't have big-endian POWER hardware, but pointed out that qemu can probably boot a BE kernel under emulation at full hardware speeds. So if that works I can take care of the 32-bit POWER (which is mostly BE) too.\nFor reference, I managed to get a working kernel derivation for Power9 with Just setting building a zImage without fully stripping it afterwards wasn't enough. There still was a reference to in the debug info somewhere.\nYep, very plausible that there are other references as well. We had for x86 and RISC-V, although that specific issue didn't apply to other architectures.\nI'll dig into this this evening. thank you for the tip! Also putting this link here so I don't lose it:\nVerified (after fixing a bunch of other things) that is affected too. Thanks again for the heads up Working on a fix.\nEdit: I meant to post this .\nFixed by"}
{"_id":"q-en-nixpkgs-8ca0a30a270a15b2e84a1a4d8ee313d4bd3c36d68a88274a0a71d683ae16604b","text":"The pkg-config file in pkg looks like this: The and variables are empty. The keyword is also empty. As a result, the flags returned by pkg-config are wrong and checking the version does not work. Note the empty and and that the package version is actually 3.0.1, i.e. = 3.0.0. and should point to the correct directories in the and outputs of the pkg."}
{"_id":"q-en-nixpkgs-e50570af4f134a26793d6320b6bab771e8a63ec383f72f45b466b1a30479d3dc","text":"Project description A cross platform mod loader for The Legend of Zelda: Breath of the Wild in CEMU. Metadata homepage URL: source URL: license: GPL3+ platforms: Windows, Linux\nI actually have this packaged in my nur repo: , but I let it go out of date for a bit. (should get back to updating that ). You can test it out with The version I packaged works really well (or at least when I was still using it ) - it was just frustrating to maintain with all the patches & new dependencies, and I decided to hold of on merging to nixpkgs for now.\nMaybe should be made instead?\nPlease test / ."}
{"_id":"q-en-nixpkgs-c1f6a395b55bf3c2ed735b47cde4ec519644bbe9b85ce115905ba6ce53dcb337","text":"Project description Metadata homepage URL: source URL: license: MIT and Apache 2.0 platforms: linux and darwin (might also work on other platforms) Additional Notes It's worth noting that appears to be updating off the branch with a re-created tag upon each \"release\". It may be the case that doesn't make sense to be in right now, but perhaps folks have ideas on what it could look like to update automatically. EDIT (17 Apr 2023): I suspect the package will go in . It's also worth adding the following note from the source URL's ...\nDidn't try it out myself, but this may work\nI'm getting close to publishing a PR actually :) Should be up soon.\nUpdate: adding packaging to the upstream flake may be the best idea for now. Either nixpkgs will need to support Rust-based packages with nightly toolchains (unlikely and for good reasons!) or will need to be buildable, and intentionally so, not as a hack, with stable Rust (linked issue: ). The good news is that caching solutions exist and work in should provide a good bluepoint for where to start with an upstream contribution. Full Context: Fortunately, likely will not have the same issue being included in nixpkgs, so I am authoring a PR to add reindeer (issue: ).\nAs of yesterday, I have a flake available on my fork that you can use to consume the binary in your downstream projects easily. For now, consider this a simple workaround. I plan on semi-regular-kinda-ish updating it and rebasing on trunk, and eventually contributing it upstream; see also Flake usage is easy. Something like the following should work where you can use or inside your flake outputs to refer to it, and you can put it in your environment, etc: You can also get a binary in your with\nAs of , there are now buck2 binaries, thanks to the Meta team doing releases! They can be updated every two weeks easily. So that means you don't need a Flake, cache, or any expensive build steps. Please let me know if the binary doesn't work for you; I've done some light testing on NixOS, and it seems to work fine in a normal environment; if things are broken somehow, we can pull it or figure something out.\nTwo side notes: I will probably migrate the flake to a new repository, so that I don't have to keep my main development fork clean at the moment. We can probably close this now, I think.\nThanks for the help I believe does close this issue since is now in nixpkgs. Great work and thank you again :)"}
{"_id":"q-en-nixpkgs-a1ccb13aab78638cec78c76b97ee38bfa3b07c881f4dc07bd41288a8c7e78516","text":"Project description is a cargo subcommand, which shows crates info from See their for an output example. Metadata homepage URL: source URL: license: mit OR apache platforms: unix, linux, and darwin"}
{"_id":"q-en-nixpkgs-6ee16623ff1d530b52bca5bf22217b241aed1970a76aaacd56391a3736647f0c","text":"Package name: gnome-obfuscate Latest released version: 0.0.9 <!-- Search your package here: --- Current version on the unstable channel: 0.0.7 Current version on the stable/release channel: 0.0.7 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-83414eed7f913030053f8c34e49bf4f94a1dc296d8e95e96dfa4cff782ea0760","text":"Project description Perhaps it would also require a module to set the correct permission for /dev/uinput? dotool reads commands from stdin and simulates keyboard and pointer events. It works everywhere on Linux, including in X11, Wayland and TTYs. <!-- Describe the project a little: --Metadata homepage URL: source URL: license: GPLv3 platforms: unix, linux dotool requires permission to to create the virtual input devices, and a udev rule grants this to users in group input. You could try: echo type hello | dotool and if need be, you can run: sudo groupadd -f input sudo usermod -a -G input $USER and re-login and trigger the udev rule or just reboot. See , but this greets the world: echo 'type Sup, Lads!' | dotool and this screams for three seconds: { echo keydown A; sleep 3; echo key H shift+1; } | dotool There is an initial delay registering the virtual devices, but you can keep writing commands to the same instance or use the daemon and client, and . dotoold & echo type super | dotoolc echo type speedy | dotoolc dotool will produce gobbledygook if your environment has assigned it a different keyboard layout than it's simulating keycodes for. You can match them up with the environment variables and . echo type azerty | DOTOOLXKBLAYOUT=fr dotool dotool was written for , which has a you're welcome to join. You can also send questions, thoughts or patches by composing an email to ."}
{"_id":"q-en-nixpkgs-2a822f5ef8034e3c3731bd8d7c707bcbdf51f1b8647ed3b20f2e882e87516ed1","text":"The script defined in contains a if it does not yet exist. The file name of the disk image is stored in the bash variable which is when creating the disk image. Thus, if contains a space, the disk image creation fails. The following is an easy way to reproduce the problem under macOS with a recent nixpkgs; not sure about a Linux example, but anything that triggers a VM creation via in the current directory should do the trick. and enter a folder with a space in its name: a VM via nix, e.g., (a recent builder VM for macOS). (will ask for sudo password for key setup) image creation is triggered and fails with the following output: The builder disk image for the VM should be created and the VM should be started. (Also seems like there is a minor typo in the error message: \"Disk image does not exist, …\".)\nGood point, can you send a PR?\nSure, here it is. As far as I could see this seems the online place in the script where the variable was not properly escaped. A quick test via from a folder containing spaces worked as expected."}
{"_id":"q-en-nixpkgs-ebff16805922fde9747228a2c26cff81b753e319b3a5fcfdaef643b60150f46b","text":"Regression caused by : on non-NixOS (in my case, Ubuntu 22.04), building the formats causes the following error Steps to reproduce the behavior: nix and nixpkgs on a Ubuntu 22.04 on some custom collection of packages. : call (after adjusting , because is not currently wrapped). Whereas calling seems to work fine, thus it's really the call to that messes up things. Formats should build, of course, and there should not be any mention of the impure ! also I don't know how to fix this one! Please run and paste the result. `\nYeah, I don't think we can fix this. I suggest a revert and patch properly.\nIn the meanwhile, I discovered my nix install (via the Debian package) couldn't sandbox properly. After removing the Debian package and installing nixpkgs proper, I could sandbox the build and avoid the issue. So this is in part a bug of the Debian nix package (because it does not sandbox properly). My suspicions are also on the function in perl, which calls (hardcoded!). That's my current best guess as to why the impure linker is involved. Maybe the correct solution to compile perl with the correct , if that's even possible... ? In the meanwhile, if non-sandboxed builds must be supported, we should definitely revert before 23.05 branches off.\nQuick note: I am fairly convinced now that is causing problem because preloading libfaketime causes a glibc conflict. We should probably patch , or , in the perl Configure script so that perl calls a nix shell. if you know the appropriate person to cc for perl stuff, please do, I don't think I can patch perl correctly on my own.\nThe fact that the system function executes /bin/sh is governed by glibc. I remember this was discussed before, and there were issues with hardcoding in there.\nPresumably a cycling dependency bash -glibc -bash, at a minimum. The lesson: is fragile and must be handled with care. Better patch fmtutil locally, like we do for mtxrun, to add faketime in front of the format calls. (Changing the perl configuration might still be doable, though, regardless of whether it changes the behaviour of . I'll give it a try.)\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-d146c43e62866820affabfcd23b632ed96e0b9b4303044c8ffb00d7e40197a53","text":"With the urxvtd service enabled on NixOS, the first instance of does not display anything, but keeps running. All following instances work as expected. This differs from the behaviour of not enabling the urxvtd service and just running manually, where everything works as expected. Set . Reboot and log in. Open a non-urxvt terminal and run from it. I'm running on XFCE with XMonad, and the problem has been reproduced on plain XMonad. System: (Gorilla) Nix version: nix-env (Nix) 1.11.6 Nixpkgs version: \"\" /cc\nUhm, I can't reproduce it with either bspwm or no window manager. I will try XMonad later. update: Just tested with XMonad and negative. Are you sure you don't have anything else interfering? Can you test it in a VM?\nI just encountered this issue on a fresh install of . I'm also using xmonad, though it might be unrelated. On boot, shows nothing. Upon opening xterm and running urxvtc, no new window opens but nor does the command return anything; it continues to run the job. Now if I open another xterm and run again I see both a .urxvtd-wrapped and a urxvtc instance, and running urxvtc now spawns a new terminal window as expected. Earlier when I ran I got a torrent of warnings for collisions, I think between with-unicode3 and with-unicode3-with-plugins, even though I've never put the latter in my config. Don't know if it's related. I've since GC'd and cleared old profiles, and the warnings were not reproduced on my most recent . Edit: actually I got the collision warnings again, so that's still a thing.\nI have had this too for a while now but I had forgotten about this issue. I couldn't figure out what's the cause. I'll try investigating some more.\nIs the problem still there with 18.09?\nYes, unfortunately it's still here.\nI looked at this on NixOS 18.09 under XMonad. I think the issue is that urxvtd is not compatible with socket activation. When urxvtd starts, instead of listening on the socket provided by systemd, it creates its own fresh socket file at the same path (), replacing the old one. So when you start urxvtc the first time, it hangs because it's connecting to systemd's original socket whereas urxvtd has created and bound to a different one. When you start urxvtc a second time, the issue does not occur since urxvtd is already running (urxvtd creates a new socket only once, when it starts up). To verify I used these shell commands: systemctl --user stop urxvtd.service systemctl --user start ps -ef | grep urxvtd # urxvtd is not running stat --format %i \"$RXVTSOCKET\" # prints the inode of the socket file urxvtc # hangs, kill it with Ctrl-C ps -ef | grep urxvtd # now urxvtd has started stat --format %i \"$RXVTSOCKET\" # socket inode has changed urxvtc # works as normal stat --format %i \"$RXVTSOCKET\" # socket inode has not changed Checking [rxvt-unicode source code][1] confirms that it intends to delete the existing file at . [1]:\nI have submitted a to fix this by removing the urxvtd socket activation.\nthanks for figuring this out! tiny nit-pick the link to source code is wrong.\nwhoops! Fixed now."}
{"_id":"q-en-nixpkgs-f5e65dc0b483d3f23b2a210db5e4426ed7aa4629c7d703c80aabe98f2a2ff380","text":"Package name: swayfx Latest released version: 0.3 <!-- Search your package here: --- Current version on the unstable channel: 0.2 Current version on the stable/release channel: not packaged <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Note for maintainers: Please tag this issue in your PR.\nc.f."}
{"_id":"q-en-nixpkgs-e8f3c7c580a42ef952cfa8920f991a7616a7171f5ffdf80a6eddf0608782902c","text":"NetworkMiner NetworkMiner is an network forensics tool that extracts artifacts, such as files, images, emails and passwords, from captured network traffic in PCAP files. NetworkMiner can also be used to capture live network traffic by sniffing a network interface. Detailed information about each IP address in the analyzed network traffic is aggregated to a network host inventory, which can be used for passive asset discovery as well as to get an overview of which devices that are communicating. NetworkMiner is primarily designed to run in Windows, but can also be used in . Metadata homepage URL: source URL: platforms: linux"}
{"_id":"q-en-nixpkgs-cc6862e1c6cb7c7e053df1b78cf2ce0ca500fa080c12d4dd6796cb98675229f9","text":"Builds ok on x86_64, but on aarch64-darwin it hangs every time at  Steps to reproduce the behavior: <details</detailsI'm surprised not to see other issues about this as it seems to be logic contained in a major package (pytest), so I wonder if it's a problem local to my system. I tried rerunning after and and outcome was the same. Please run and paste the result.\nI can reproduce this and, in my case, it is caused by the test in pylsp-mypy. I do not yet know why.\nI believe this will be fixed by I'm planning to wait a few days for a response from upstream before patching this in case the code has to change, but please feel free to pull it in earlier if you have a need.\nYes, that sounds suspicious -- I'll try to build the PR and report back.\nSure enough, that patch does the trick."}
{"_id":"q-en-nixpkgs-2173b17f9835e3fac524fa0f91291503e0589e81a2702e2ebead5c3276a29db4","text":"So I today I spawned a machine with and satisfyingly the directory was generated with all of the certificates etc that taskwarrior needs, including the clients' certificates. I noticed that the users' keys located at are owned by root with read only permissions for the root user alone. No matter what are and . I'd like to enable the group read those client keys, as well as the file so that I can configure a client easily. Thankfully the group can read the file, but not the client keys. The manual indeed explains that the command: Could be used to get those keys, but I'd like to use syncthing for that to automate the process even further. I got a bit lost while reading the . Specifically, I'm not sure why does always works in the following umask: I mean, ~why not~ - a solution is suggested in . A word of comment will be appreciated. cc ."}
{"_id":"q-en-nixpkgs-a863339e4d69b124b6134169ef15732db79d781e89353e07942cbf47465c26c9","text":"Package name: electrum Latest released version: 4.4.4 <!-- Search your package here: --- Current version on the unstable channel: 4.4.0 Current version on the stable/release channel: 4.4.0 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Reporters Note: Tried the update locally and got the error Looking at the diff on the electrum github I see the substituteInPlace pattern will no longer match: In:\nPR in"}
{"_id":"q-en-nixpkgs-4c1552c27d6076cc7f34c61aec6fe53ae57f32f84d890eac839e0ba08f511d9f","text":"Hello I was trying to review just now the source code that 'rstudioWrapper' uses. When I hit source from nixos-search I was redirected to this link However after checking in , I noticed that the expression that runs 'rstudioWrapper' is this one I'm not sure why this is happening, nor if the report should be here or in nixpkgs. In case I made a mistake, I would appreciate if you let me know so I can open another report. Thank you very much for nixos-search, I can't imagine the Nix experience without this amazing search engine.\nThis is indeed a nixpkgs issue as nixos-search just links to . The default position is determined by , so things like usually get their position reported at the definition of rather than at its call site. Maybe that could be fixed by making use the position of one of its arguments, but I don't think there's anything to be done in the case where the second argument is empty ().\nOh, that was fast. Thanks for your help! I stay tuned if I can/should do anything.\nHello again! I'm just seeing that your PR has already been merged. Do you agree if we close this issue, or is there anything left? nix-search unstable already perfectly references rStudioWrapper! Thank you very much for your help.\nThis is probably not completely solved yet (uses of without arguments would still have incorrect position information) but there's nothing actionable left to do so we might as well close it."}
{"_id":"q-en-nixpkgs-4ac9564ec460b52958aac849e418811690f2ebe5341b4d687d3c12917424d909","text":"As a newcomer to nixpkgs and MacOS user I've followed manual how to install . Unfortunately second command in tutorial which is supposed to list available Haskell packages fails to execute: This was broken by refactoring done by in the commit: Run command on MacOS. System: OS X 10.11.5 / Darwin 15.5.0 Nix version: 1.11.7 Nixpkgs version: 17.09pre102737."}
{"_id":"q-en-nixpkgs-8645ae15e015949b424929f11f2de22dc25ba28cde36b6ffe60f033853622449","text":"The nextcloud package set offers apps, that have licensing information attached, but not exposed as part of their meta attribute. Also, there are some apps, like news, where the licensing information says , which requires clarification. This needs to be fixed.\nThank you for opening the draft PR in order to fix my mistakes and broken packages. I'm sorry for the bad communication and the inability to resolve this serious issue. Fixing this for Wordpress packages and themes (wp4nix) and Nextcloud apps (nc4nix) in a foreseeable time exceeds my skills and competence. Therefore I would like to offer to add a flag to these package sets and remove them for the next release 23.11 if it's not resolved by then."}
{"_id":"q-en-nixpkgs-18aedd42ea6dd2b7a08ac21ca6465ba0bc6ab3ac48302daa9fd6f6a13e48582d","text":"Project description GoatCounter is an open source web analytics platform. Metadata homepage URL: source URL: license: EUPL platforms: unix, linux\nI searched issued but not pull requests so I missed , should this stay open until that's closed?"}
{"_id":"q-en-nixpkgs-ce7e5c2658929b6520a36feaa57a19ce368cb050f553edc548ce95022b75251b","text":"defaults to the tmpfs defaults: 50% of RAM. What do you think about making it configurable? tmpfs has no real overhead when not used and just like normal memory, it can be swapped out if memory gets tight. I have a hp microserver n40l with 8GB RAM and i can't rebuild the linux kernel with , as my /tmp fills up even though i have a bit of RAM left.\nI know i can do this instead (copied from one of my machines): but I don't like having a nice option for this, and still using the \"manual\" way.\nsince you have already found a solution for this, have you considered opening a PR?\ni don't find that conversation atm (discussion was probably on irc?). anyway, my \"fix\" was too liberal in it accepts as input (bool and string) and didn't reuse the original systemd template/unit file. feel free to reuse this as a basis for your own pr or give me pointers how i could do that in a few lines and without messing around with the option. maybe i'll try a pr with the solution, but thats more of a hack. thanks for the reminder, that this is still an issue.\nWhy is a hack?\nbecause it doesn't automatically get new features from the upstream systemd unit file. that was the main part of point two. whenever systemd gets a new feature, the unit files in every distro get automatic upgrades - just nixos would need manual intervention.\nOh I wasn't aware how this is currently done. Is it realistic to expect new \"features\" for a filesystem mount? Maybe my imagination is lacking, but I have no idea what that might be.\nsee\nThank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as \"still important to me\". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly: for maintainers and people that previously touched the related code and @ mention them in a comment. on the . on the on .\nThis still seems like a desirable feature, especially since nix builds happen in /tmp.\nI marked this as stale due to inactivity. &rarr;\nAny activity on this? I've a build failing, even though boot.tmpOnTmpfs is set to false\nhave you tried increasing size of /tmp ? eg\nWhere is the RuntimeDirectory located? I thought that should be , which on my system is pretty big: I'll try to add that option to my config, not sure I fully understand what it does though, is it documented in more detail somewhere?_\nsorry i made mistake in my post when i mentioned because its really not that one on my system it's You can look up systemd documentation for RuntimeDirectory\nRebuild nixos and reboot.\nThere is now an option for this:\nFor those who end up here, my problem was that for some reason was set to . Tip: just change the value and inspect the result with ."}
{"_id":"q-en-nixpkgs-b93d36fa84a54426753cf324406014a9bf099a57d2b31c10debba12e7a9daf52","text":"Some cuda packages come with vendored copies of libraries that are already available in nixpkgs. These vendored libraries still require patching their Runpaths, their presence reduces determinism, and breaks \"deep replacement\" (i.e. nixpkgs' feature allowing to ensure that everything in a package set uses the same revision of a dependency). For instance, ships , which otherwise is available as: Cf. Tasks [ ] Look into deleting these vendored libraries and re-linking their reverse dependencies against nixpkgs alternatives"}
{"_id":"q-en-nixpkgs-7250940a382d93ce3f8b03ef2387010e35ecd76f783c89788a2fe7a9c1577279","text":"Project description A video game ROM collection manager to help filter, sort, patch, archive, and report on collections on any OS. Metadata homepage URL: source URL: license: gpl3 platforms: linux"}
{"_id":"q-en-nixpkgs-d1368aa371a327dc0e11499482e79a1a7f9ec7fd4d15c76a1447d42820f08c88","text":"There are syntax elements, like which are used in the manual but undocumented in the . the generic syntax is document , but a few more non-admonition variations have been implemented recently that aren't yet documented in the manual. Originally posted by in"}
{"_id":"q-en-nixpkgs-ca08b5b8bfb273df21e640cb89dca37a4a406385ff1831f4351071fe10210422","text":"the ruby gem tzinfo exists in two different versions/branches which are both maintained. the branch does not contain the file referenced by nixpkgs, only 1.x does. Ref: Extract from pkgs/development/ruby-modules/gem- I fixed it by overwriting tzinfo (again!) but this is not a good solution :/ cc System: (NixOS: , Ubuntu/Fedora: , ...) Nix version: (run ) Nixpkgs version: (run )\nHmm. I don't remember by heart, is available? If yes we could switch based on the version. Pseudo code:"}
{"_id":"q-en-nixpkgs-d61694aa81f4a3b056e934f4c12155499244d943dd08c5262dfb8aa8590d0f77","text":"webUISupport is needed for the command. However, enabling this flag fails the tree-sitter build. Steps to reproduce the behavior: to your favorite checkout of nixpkgs A successful build. I'll give you a section of the build log instead: Please run and paste the result.\nI still get this error with tree-sitter 0.20.9 in a Nix flake. Did setting the emcc cache outside the nix store fix the build for you?\nIf I try to reproduce it, I get this error: Which looks similar, but I think is a distinct issue. I do not think this is related to the emcc cache. I think it's best if you open a new issue for this. The emcc cache is not set in the Nix store btw, it is set in a nix build directory, wherever that may be. You can generally not set the emcc cache somewhere outside of the nix sandbox.\nSounds good. Thanks for the clarification. Wasn't sure if there was a setting I was missing\nI've got a PR out (see above) to fix this new issue.\nAwesome. Thank you"}
{"_id":"q-en-nixpkgs-a33da509c3b6146d265d8ea6d5c18ea59d49722f9ccb10657339b49568d714b3","text":"Project description python docstring generator <!-- Describe the project a little: --Metadata homepage URL: source URL: license: bsd platforms: unix, linux, darwin, ..."}
{"_id":"q-en-nixpkgs-c2b997fa53697853633d3b8e9f9e163efee82c573d1aba0751a40f28a9ea2db0","text":"I see that the custom CPU scheduler, zen's main feature is disabled: This raises a question: does nixpkgs' zen actually differs from the default kernel or nixpkgs disables all the customizations by applying the same config as to the default kernel? Moreover, nixpkgs provides the -lqx variant with the same config but the main difference between -zen and -lqx variants is the config: liquorix provides a pre-defined config with its Debian/Arch packages and when I see lqx in the repos, I expect nixpkgs to use it as well as there's no source code difference. The -lqx variant seem to be poorly maintained in nixpkgs in general, it lags behind -zen variant (6.4.1 vs 6.3.11 in current nixos-unstable). Steps to reproduce the behavior: -zen should have its features enabled more or less, the -lqx variant MUST use (and probably be available only for x86_64). See the for example. Would be nice to have variants with different schedulers like the AUR package allows, preferably pre-built on Hydra (well, I don't have a powerful enough hardware to build kernels).\nIt might be a good idea to also have a xanmod variant built with .\nI created a PR that should fix this: I took a look at the Arch config for -zen and it looks like it is basically the vanilla kernel with and some tweaks to the RCU/Preempt/tick. So this is what I did above. For the -lqx variation, I activated the PDS process scheduler, BBR2 and other tricks that seems to make Liquorix kernel (according to the site: ). This is not exactly true, the release of the -lqx variant generally lags between major kernel releases, so while -zen gets to X.Y.Z, -lqx will stay in X.Y-1.Z for some weeks (I imagine for stability purposes, but I am not familiar with upstream). This will be difficult, because we can't simply take its configuration as-is. If the current approach in the PR above is not acceptable, I would say it is better to just remove the -lqx from nixpkgs. We already don't seem to have that many users of the current variants considering you're the first person to note that we are compiling the packages without the customizations. So unless there is really a strong demand I would say let's not do so and just keep 2 variations of this kernel that should be enough for most users.\nIt's acceptable as a first step but why can't we?\nI guess only the minority checks actually? I periodically see in Arch groups how people try -zen thinking they get all the features for better responsiveness but don't see a difference and don't even think about checking what scheduler they run actually.\nIt would be really difficult to use upstream configuration as-is. I mean, I don't think it is impossible, but I definitely don't want to be the person to do so. What I mean is mostly that I don't think most people care anyway. Even users of those custom kernels doesn't seems to care too much if they're actually using a custom scheduler it seems, considering again you're the first person to note this issue and this kernel is in nixpkgs for a long time already. So considering that nobody is that interested in the first place I think adding another variation of those kernels is too much. We definitely can do in the future if there is enough people that wants those though.\nWhat I mean is they do care, it's just they don't know how to check. They see -zen and believe they get all the features out of the box."}
{"_id":"q-en-nixpkgs-96e3ea347c0cf73b0a2ad52eaee67ebd3b5613a2e4bedcea998ae1b4d25444c3","text":"Project description Virter is a command line tool for simple creation and cloning of virtual machines based on libvirt Metadata homepage URL: source URL: Blogs: , license: Apache v2 platforms: Linux only written in: go Thanks a ton!"}
{"_id":"q-en-nixpkgs-03d0318aa045183550707991fd1831aeae19f4cf24f2e19413ab83b7f5d83841","text":"When attempting to configure the integration I get an error in Home Assistant Steps to reproduce the behavior: to to configure integration from Home Assistant UI Able to configure <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nThe package does not eval on master right now, because of a dependency being flagged as disabled for python 3.11. Though I'd be surprised if that is your issue. nix-replhome- error: … while evaluating the attribute 'drvPath' at 221| in commonAttrs // { 222| drvPath = assert condition; drv.drvPath; | ^ 223| outPath = assert condition; drv.outPath; … while evaluating the attribute 'drvPath' at 221| in commonAttrs // { 222| drvPath = assert condition; drv.drvPath; | ^ 223| outPath = assert condition; drv.outPath; (stack trace truncated; use '--show-trace' to show the full trace) error: cherrypy-18.8.0 not supported for interpreter python3.11\nRegardless, at least makes evaluate."}
{"_id":"q-en-nixpkgs-38b1b1256938e37ba2eadeac10da05ea0aa2050adea4293c8cf359828c8712e6","text":"The Haskell Hackage database allows distros to report which packages they ship, and include a link. Right now, it seems that nixpkgs reports all haskell packags, including broken ones. This makes for a bad user experience – I might see on hackage that nixos comes with a certain package, and then happily go ahead trying to use it, only later to get the dreaded error message. Same actually for repology, reporting packages as available that we know do not build is not great.\nWe intend to support all package. We just need haskell maintainer\nOh, absolutely! Maybe I shouldn't say “supported” but “available”.\nNo, we don't. We only intend to support packages which have value to our users. Hackage is append only and contains a lot of dead or niche packages.\nI agree this should have been fixed long ago."}
{"_id":"q-en-nixpkgs-af491a7034ebf84eb45956291e0ff134cafe4cb3224e8730f9e06f636c5d0975","text":"VSCode/VSCodium's desktop icon (specified in the file) is set to . On the icon theme (the default dark icon theme that comes with Budgie desktop on NixOS), this makes it look like a generic text editor, and not the respective application. The issue appears to arise from how both apps use the same desktop files that were inherited from :   NixOS w/ Budgie desktop Icon theme to Qogir/Qogir-dark VSCodium/VSCode app list VSCode should use the icon VSCodium should use the icon ! My NixOS install is running in a VMWare VM."}
{"_id":"q-en-nixpkgs-38088af3bac66e52a4aef168eb6d92781f2ce563d1bc17c6f3f3aae84a700332","text":"I use the following listing in to install a custom keyboard layout: This refers to files in the following structure: This worked until recently (last updated a few months ago), but now fails to build: <details<summaryFull trace</summary</detailsCommenting out the line causes the same error but for . Adding in plausible files in the suggested places also doesn't fix the issue. Steps to reproduce the behavior: up the configuration as above. . The system successfully builds, and I have access to the custom layout. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nSorry, I forgot to notify maintainers, but I think it's just\nThis breakage happened with with the update of as they have switched to the meson build system.\nThanks, that's really helpful! I suppose I'll revert to before that merge for now (though I'm finding it hard to find the list of good commit hashes). Do you know any other workarounds?\nYou can use . that does not have those changes. For a list you can look at in the \"Input changes\" column. That is a log of the \"nixos-unstable\" branch, for which the binary cache has been filled after a few tests have been run.\nThe script that patches the makefile to include the extra layouts need to be updated.\nCan you test if your layouts work correctly with ? I a test for a basic file but I haven't tried other types.\nWorks fine for me but I also only use symbols\nFix merged, you can track the progress to nixos-unstable ."}
{"_id":"q-en-nixpkgs-6efb3442f5a57f53498d246f0f73207c820cda4911dfb9c39d3e42717ccde753","text":"Steps to reproduce the behavior: X Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result."}
{"_id":"q-en-nixpkgs-47c88565e24a100c833a572f59e24eb5bb578bb5db80d06298916d5e9073d5af","text":"Project description EteSync is a \"secure, end-to-end encrypted, and privacy respecting sync for your contacts, calendars, tasks and notes.\" Since GSoC 2020 KDE supports it. automatically detects if it is installed and builds the appropriate modules. Metadata homepage URL: (\"build from source\" link points to a site that is gone) build instructions: source URL: license: BSD-3-Clause platforms: unix, linux, darwin, ..."}
{"_id":"q-en-nixpkgs-b3d0f3967e2853882c6f8a0338a5bc7df1e109b0d61097f8de84748718c4812f","text":"JetBrains IDEs are currently at which is kind of confusing because the packages here aren't all under name. For example WebStorm, PhpStorm, PyCharm and so on - these are properties of JetBrains. Therefore I think it will be good to move all them under and rename to .\nDo you want to make a pull request?"}
{"_id":"q-en-nixpkgs-cd3ecebcb79d243a99cbb3c0b0e8c1751048b6e07f4e99719ccf2cbcd093cd34","text":"tries to send (empty) md5 hashes to fetchurl, which doesn't accept md5 hashes it seems Steps to reproduce the behavior: enter the following command: The firefox plugin builds successfully <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --git revision tags of both NUR and nixpkgs are in the command given above\nCan you try , I tested with your repro cmd changing to in nixpkgs src with the patch.\nThis works correctly, thank you very much!"}
{"_id":"q-en-nixpkgs-b82978e60eecc691ae933eecb8185668502db649d652503be51f9fc63919a6c2","text":"Contributors working on improving the nixpkgs manual would be redirected to the . This section is not up-to-date with all the syntax features supported by , forcing contributors to look for use examples over the whole manual. The same markdown syntax is used to write and generate the NixOS manual. However, this manual does not make any mention of the supported syntax. This makes it hard to contribute to the NixOS manual with relative isolation from the nixpkgs manual. Update in Nixpkgs manual to include reference material and examples on all the syntax features supported by . Link in Nixos manual for completion. [x] remove references to in the . : [x] remove references to in the . [x] add a redirecting to the syntax section in . [x] add reference to lack of support for [x] add reference to support [x] add reference to support [x] add reference to support [x] add reference to support [x] add reference to support [x] add reference to support [x] add reference to support [x] add reference to support () () )"}
{"_id":"q-en-nixpkgs-7e64332a5555de292062af9d2d908e1ce83095b4ba252849399ebb85205c9659","text":"The wrapper only has , so the corresponding where would check for manpages is absent. don't have regular vim installed install something like 09pre104936.\nYou can find a fix in /"}
{"_id":"q-en-nixpkgs-41436d08a5b44b3f5a9a956aadcf42b7ce39e17ebe176460441378f23dec6f2e","text":"Hi TLP shows the following in output: Anything to do with it? Thanks. System: (NixOS: , Ubuntu/Fedora: , ...) Nix version: (run ) Nixpkgs version: (run )\nI believe the tlp module could be made to set for those systemd units\nFixed, thanks for your report!"}
{"_id":"q-en-nixpkgs-f0089ad87b65aa7abe9d6886bb90316c566760867dcd9458411531e510274572","text":"NixOS specialisations are supposed to have their own entries in the Grub menu for previous generations according to the . However in practice that is not the case. The older generations do not hold any boot entry for the specialisations, but only the default entry. nixpkgs: I am not sure, whether to update the wiki to correctly mention current behavior, or whether to fix the grub module to generate entries for specialisations for older generations. <!-- make sure this issue is not redundant or obsolete --- [x] checked [latest Nixpkgs manual] ([source][nixpkgs-source]) and [latest NixOS manual] ([source][nixos-source]) [x] checked [open documentation issues] for possible duplicates [x] checked [open documentation pull requests] for possible solutions\nFixed by ."}
{"_id":"q-en-nixpkgs-e91de60cb14060d3e176a0f536c2ca8dc32648f7dba68a648c373ba95c481e6d","text":"Project description From the project page: A tool to sign PDF files with Linux support. Why There appears to be a lack of applications that run on Linux and allow for attaching free-hand signatures to PDF files in a good way. (Non-)alternatives include: Xournal and Xournal++: Allows free-hand annotations but not saving those for later use. Allows inserting images and .pdf files but will rasterize them rather than retain the vector graphics, resulting in lower quality. LibreOffice Draw: Allows inserting images and .pdf files but will rasterize them rather than retain the vector graphics, resulting in lower quality. Okular: Allows free-hand annotations but not saving those for later use. Allows inserting custom saved stamps, but these will not be visible in other PDF readers. Evince: Allows certain kinds of annotations, but nothing that can look like a free-hand signature. pdftk, Ghostscript, and other command-line tools: Allows superimposing .pdf files, but not interactively selecting where. Metadata homepage URL: source URL: license: MIT platforms: linux\nWorking on it"}
{"_id":"q-en-nixpkgs-e5d46b2e2dc49947ed892ca97d7e724b780d65a8c0af006a5704f81aa3e74452","text":"Trying to set up an OpenVPN connection using the NetworkManager gui (in plasma5, but in theory it should be the same as in gnome, right?), and connections are failing immediately after a successful connection with the following error messages in syslog: Then the connection starts again and it repeats forever. System: 17.03 at Nix version: 1.11.8 Nixpkgs version: 17.03pre\nDoes the nm-openvpn user need to be to the unix group for this to work?\nThe connection reset in my log snippet may actually be caused by but I suspect that the dbus rejected sends are still a problem.\nAny news on this issue?\nI don't think I have encountered this bug for quite a while, but I also haven't been using OpenVPN recently. Are you still seeing the same error messages?\nI did not, just doing triage. I'm going to close this, but feel free to reopen it if it surfaces again."}
{"_id":"q-en-nixpkgs-39d54561538b9575b77241f7e6d700103a4762b3dce987b22dd37a09045c4f1c","text":"is failing to remove requirements as requested by the setting of . Attempt to build Upon building: Tested on . to remove the dependencies as specifies. n/a n/a Latest committers to python-relax-deps- <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nSome debugging: I'm fairly certain this is due to a naming issue with the whl. pkg_name () is set to based on some print-statement debugging. But the actual wheel that is produced is . Why are these names different? Why isn't the wheel named ?\nProbably because the wheel name is based on the package metadata. Maybe we should just ignore the version during the matching?\nYes, we should.\nCreated a PR:\nClosing now that has been merged."}
{"_id":"q-en-nixpkgs-b3137e51e58053f863cae67a2bd683c6a438f2d62e2332e48d7bcb9457a896fd","text":"Package name: muse-sequencer Latest released version: 4.1.0 Current version on the unstable channel: 3.1.1 Current version on the stable/release channel: 3.1.1 Notify maintainers No idea who the maintainer is:"}
{"_id":"q-en-nixpkgs-9dfcaf65b849b34e2c2f2d1dd6048b315cd0be7a2832b30e18269287263c74dc","text":"System: 17.03.917. (Gorilla) Nix version: nix-env (Nix) 1.11.8 Nixpkgs version: \"17.09pre105314.\"\ncc"}
{"_id":"q-en-nixpkgs-f4458b432be145c6ad08ac5c73c1b09d4db8fa139f17c7bef72bcfc7c271bc69","text":"We got theses reports of Kubo panicing on startup due to probably an empty field This is a bug in our software however I belive this is also a bug in: Because a Kubo daemon without an API listener is borderline useless, your choice to change to a unix socket () makes sense however I think some default got lost and now it is set to . Given I am not a Nix user myself I got help from to fill in the template, thx :heart:. Steps to reproduce the behavior: to notice it's failing with a panic. After installation Kubo daemon should start and the RPC should be usable by a permissioned user. Please run and paste the result.\n/cc (I forgot to fill it in the template and I don't belive GH pings peoples from edits)\nWeird that it works on my machine and in the NixOS tests. One of the NixOS tests only has and options set. does quite a lot but I didn't see anything obvious why this should affect the API settings. But maybe it does make a difference and maybe I specified some additional option on my machine makes it also work there. I'll try to reproduce it with just the enable option set in the next couple days. On one machine shows that the field is set to and yet Kubo starts without a problem and can be seen in the log. According to a comment at the top of, , \"Enabling this [socket file] will completely override any API listeners configured in your config.\". NixOS enables this socket file and overrides with (which is ) plus whatever the user has set in . What I don't know is how systemd communicates this information to Kubo but this seems to work. So in conclusion, I would think that setting the field to in the Kubo config doesn't matter because is enabled, which should override this and specify a socket for the API.\nThx I wasn't aware we even had this feature (or that you could use systemd to configure application sockets in the first place). We have a simple incorrect assumption which is panicking since kubo is configured without listeners. However this is fed by which tries to load as you linked. So if you have a but use something like it should work (AFAIT). I can't test experimentally that feeding the API through systemd works right now, I'll report later, but it seems to me that the like config is not applied somehow.\nI was able to reproduce the problem. Maybe I just need to order the unit after the two socket units. I'll take a closer look in the following days when I have some more time.\nThis happened to me too on first installation, the reason was that systemd couldn’t bind port 8080 because it is already in use by another service, then started Kubo anyways. So my theory is that Kubo checks if there was a socket made available by systemd and if there was none falls back to trying to read the listen address from the config (which is then null) and then panics in the observed way. For anyone else affected by this, run and check if both sockets were correctly bound when Kubo crashes in this way.\nI created , which should fix this."}
{"_id":"q-en-nixpkgs-13b2eeeb3bedf032fba4ee3298515bc79b94ed74e887805ce510bcd573e25190","text":"Steps to reproduce the behavior: on nixos-unstable I believe this is related to <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result.\nIt'll need a similar fix like the ones in Wasn't caught in staging because it's marked insecure"}
{"_id":"q-en-nixpkgs-38c3f01079e92395c93013dedf8628f770331b1867d3140ab4756e8a923efa7c","text":"Starting this week, we happens anyways. [^1]: will enable it, with support by Unless this causes major problems, we intend to keep this enabled permanently. Please use this issue to report any problems you encounter because of this. The and branches that are consumed by Hydra. Please contact the release managers as they will be to the exception list. In general, we expect that the PR workflow will be enough for the vast majority of cases. This change was originally proposed almost 3 years ago with closed when direct pushes were way more frequent at . Very recently a new attempt was started with , for which computed that nowadays of commits are pushed directly. The RFC was struggling to get enough shepherds, when rightly that such simple decisions shouldn't be blocked, following up with the proposal to just enable it and try it out.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nnixos-release-managers already existed\nNote that releases should also work without direct pushes, which is being discussed in , so this exception should be removed in the future\nLink is 404.\nThanks, fixed now by restoring the RFC's branch\nthat it looks like PR's can't be merged anymore when a committer requested changes (though it seems like those reviews can be dismissed? Edit: Confirmed, can be dismissed), additionally auto-merge seems to be available now: ! It doesn't seem inherently bad, but I did not expect that to happen.\n\"Require approvals\" can be disabled. And indeed, depending on the configuration, committers can also dismiss a review.\nI'm pretty sure that setting is not enabled. Indeed as an example, was merged 1 hour ago (when the setting was already in effect), and it was merged without any reviews. The good thing is that dismissing reviews does allow merging again, even if it's your own PR. And even better: To dismiss reviews you have to give a reason, I tested this here:\nFyi, we do have the option to turn off allowing auto-merges, see .\nI think we should disable the auto-merge feature because it works only based on the reviews and completely ignores the OfBorg status. And until we have more solid CI we cannot also require CI checks to pass before merging.\nI turned off auto-merges.\nMeh. I don't oppose this policy, but I don't like how this decision sidesteps the RFC process. Does this imply proposals can be tacitly approved if they don't get enough attention? It doesn't seem right.\nfor the record, I'm fine with that and I withdraw my skepticism :) The problem I had with the previous behavior was that there was no way to override it (and from my experience almost every rule tends to have an exception ;-) ). I haven't checked it so far, but if it's possible to dismiss now while having to provide a reason for that seems good to me, thanks :+1: OK with this one being pinned globally, my other issue is also resolved, this provides reasonable visibility for each committer. IIRC we haven't had notable config changes in the repo for a while and I'm not sure if there are follow-ups planned to this, but for the future I'd like to suggest to always communicate it via a globally pinned discourse topic (and perhaps even cc all of the committers in $github_issue).\nI know it looks a bit like we pulled a Flake again, but: The change does not have any long-lasting API or user implications, it can be reverted at any point without problems The change has a very low negative impact, of commits were direct pushes Meanwhile, the change has a significant positive impact, especially for security Considering this, I don't think it's necessary to have an RFC for this, because it should be an uncontroversial change, and indeed, we haven't seen any good reason in to not do this. But also a big argument is that going through an RFC for something so minor distracts people from more important problems to fix. Not having to maintain this RFC for months frees up time for me to e.g. focus on implementing accepted NixOS/rfcs, which is considerably more impactful.\nClosing as this was up long enough"}
{"_id":"q-en-nixpkgs-dda36e2f6d308124ca80d16956612fb595dfe55fc66023b2c6b9320e2b0bcb14","text":"php is unable to find datadog- (and it's normal) Steps to reproduce the behavior: a shell with php and datadogtrace extension php.withExtensions ({ enabled, all }: with all; [ xdebug curl opcache openssl datadogtrace]) nix-shell and run php -v display \"Warning: PHP Startup: Unable to load dynamic library '/nix/store/qi72nhxwvbysg35nbvf89q3ywad46xyc-php-datadogtrace-' (tried: /nix/store/........;\" and the extension isnt' loaded the extension is well installed but php configuration is looking for a wrong one (the original should be named and the package is looking for\nThanks for creating the issue :) Fixed in\nYou ask me for that i think it's normal, it's a minimum :) NOw we just have to wait to make it work so ? have a nive evening (i'm in france) Le mer. 16 août 2023 à 21:29, Pol Dellaiera écrit :\nYou can monitor the progress of this PR by using the following tool: . If you are using Nix tools from the branch, they should already be available to you, but they will need to be compiled on your machine. This means that you'll have to wait for the compilation process to finish before you can use them. However, if you can wait for one or two days until this PR reaches , the binaries will be hosted on . At that point, they will become immediately available, and you won't have to deal with the hassle of recompilation. Greetings from Belgium.\nno probleme to compile, l try !!! yes ! [image: ] [image: ] Le mer. 16 août 2023 à 21:43, Pol Dellaiera écrit :"}
{"_id":"q-en-nixpkgs-72571cbbda2cae3b4e0285798173be424c45a54319f64822de055b072f3abf4a","text":"outputs something like \"187183230175148230\" instead of Chinese when it should output Chinese. Steps to reproduce the behavior: of the terminal use a command similar to the following Can output Chinese normally <img width=\"803\" alt=\"image\" src=\"\"The terminal is the one that comes with the gnome desktop. <img width=\"803\" alt=\"image\" src=\"\"<!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --"}
{"_id":"q-en-nixpkgs-c61638a6fd2175aff4a3f90649528094da1d5022bba77427a76b3f875c372ae3","text":"Project description Enable to show workspaces. Metadata homepage URL: source URL: license: mit apache platforms: linux\nI'm using Yambar on the unstable version with the + It's displaying workspaces correctly.\nThank you, didn't think to check unstable :~)\nAh I think I forgot what this issue was about, I'm looking specifically for packaging of a rust binary that listens for workspace info from hyprland's IPC.\nAlright, I'm currently working on it. !\nTYVM!"}
{"_id":"q-en-nixpkgs-e7838e1b1559e2eb56889936d0be238c2321dd85c11d3d4f9e6c449d8b8b0148","text":"Dopewars is a game simulating the life of a drug dealer in New York. The aim of the game is to make lots and lots of money... unfortunately, you start the game with a hefty debt, accumulating interest, and the cops take a rather dim view of drug dealing... <!-- Describe the project a little: --Metadata homepage URL: source URL: license: GPL-2.0 platforms: Linux, Windows, MAC"}
{"_id":"q-en-nixpkgs-be1dfc843eec01a09fa9b483d920f907317e286bf0f42506a8c8f53bafc01e6b","text":"Steps to reproduce the behavior: package 'grapejuice' Fixed by adding to propagatedBuildInputs Please run and paste the result."}
{"_id":"q-en-nixpkgs-da1b2f21ba4e2462c9aa308b29a87278e1e870e01a97b31982d01e99031c54ed","text":"Salmon is a bioinformatics tool to produce a highly accurate, transcript level quantification estimates from short-read RNA-seq data. It's mostly written in C++. There's a similar tool already in nixpkgs (Kallisto), but they both have a few unique features, so packaging Salmon is still useful. It's also quite a popular tool. homepage URL: source URL: license: GLP v3.0 platforms: linux, others?\nI started packaging it, and made good progress. It's a way for me to learn Nix, but may need a few pointers to get it done\nOpen a draft PR, then."}
{"_id":"q-en-nixpkgs-cab22eab6d5ebd2f72a049678b0b7201699b0cf1596f7564185f9741752a2818","text":"calibre-web failed to start Steps to reproduce the behavior: -rebuild switch -u calibre-web The service start successfully, and no error occurs. If applicable, add screenshots to help explain your problem. Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result."}
{"_id":"q-en-nixpkgs-a2f471874ee38ceea1c6ca006c8fd2f3713289077b4b31a40b768bda1413ccdf","text":"fresh/clean install. edi the config file to add a package (in this case an editor) - but no editor is easy to find in the initial config to edit later the config (this is called the chicken and egg problem?) by the way - the manual, when mentioning the config file, point out its full name (including the location in /etc/nixos). this is not obvious when you run nix the first time! (This happens with all reasonable NixOS versions — )\nis an editor, . , you can or something without editing config, then, if desired, remove the editor from profile when it is installed system-wide.\nthank you - i did not think of nano. i was looking through the menus and did not find any; of course, nano would not be listed there. perhaps you add a comment to avoid stupid questions like this from newbies trying to use your interesting OS.\nLet's hijack the issue to describe the real problem then…\nWhile we are on it we could also think about adding an option in nix that installs your chosen default editor and sets the env accordingly.\ngood idea! for newbies, give some hints (e.g. gedit or similar)\nPerhaps add to the template? But I saw that we also have and .\nI suggest we deprecate.\nadd one of the simple editors as well - gedit, mousepad or one of the other graphical easy ones. It is important to have a very low entry threshold to capture new potential users. Something like the 2 second rules on web pages: if the new OS does not work out of the box for the newbie, it is dumped...\nI'm all for making editing text files easy, but I'm not sure what good it is to set EDITOR to something with a gui. I think a user that might need help with editing text files will install a desktop environment that presents it's default editor to the user by itself. On the other hand, a user might try to do something like ` because they pasted it from some web page and unknowingly run gedit with root, making their system insecure.\nWhat makes it more insecure and using an terminal-based one with root privileges?\nis expected to point to an editor that works without X, whether run as root or not. uses and is expected to also work in the system terminals, independent of whether any X session is present.\nI'm all for setting EDITOR to something sane like by default and for generalizing it similarly to X11. I don't think this argument really applies to NixOS. Any OS that requires you to edit some configuration files before you even install it is not for a newbie unless that newbie has some interested IT personnel around. I imagine it could be possible to make a usable GUI for with text lenses (in Haskell), though. But, in general, good installers (like Debian's) take a lot of effort to produce and yet most people never ever install their OS and most people that do install their OS could have just learned the shell commands.\nThank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as \"still important to me\". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly: for maintainers and people that previously touched the related code and @ mention them in a comment. on the . on the on .\nI think we should include nano in our that we generate.\nWould be the right place to change that?\nI marked this as stale due to inactivity. &rarr;\nyes."}
{"_id":"q-en-nixpkgs-31411509938fcb29bc2b536f7fdc40bf5d949ac2af0ee12ffbb43b5e9134a030","text":"Want the latest 1.3.0 dev release of pants (to get a particular bug fix). Just running pants w/in nix-shell w/ trying to use the latest version. System: OSX Nix version: nix-env (Nix) 1.11.9 Nixpkgs version: \"17.09pre106045.\"\nFixed in"}
{"_id":"q-en-nixpkgs-62bc58c7a8876f500bfb7c71c35261ea18e68b4191827c82019db6d7af83790e","text":"Package name: wavebox Latest released version: Current version on the unstable channel: Current version on the stable/release channel: (doesn't seem like the has anything to do with Wavebox 114 to 116?) Notify maintainers Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-df65d4d79b1619033fbe4af2f458da6bec77811622c5d64047188acd4362bb73","text":"Project description <!-- Describe the project a little: --official sostitute of nimlsp Metadata homepage URL: source URL: license: mit platforms: unix, linux, darwin, ..."}
{"_id":"q-en-nixpkgs-13ad95f163fbcb3717e4ecdcf485a350415e81d1003555e0cd5a359cb46cb990","text":"Project description The missing backlight auto adjusting program for Linux. Metadata homepage URL: source URL: (zip file) license: unknown platforms: linux virustotal:\nI feel like shaving a yak. I'll take this.\nActually.... among other things the source here is a pain and basically impossible to make reproducible. Pretty annoying so I shall email the author.\nhere is a package in the interim.\nfyi, I found the GitHub repository\nOh, nice. Still needs a license though before we can include it.\nNot a blocker from my point of view. Package can be till the licensing situation is clear.\nAuthor is on it re both of these things, btw! I would just give him a day or two.\nIt's now MIT licensed.\nPR filed."}
{"_id":"q-en-nixpkgs-e0c4426a0773fbbc0d8d7bf3e1d0c754c00f6b2a591a91df363e14b72250f26a","text":"CMake seems to use ps command internally, but looks like CMake in nix on darwin can not find ps command, so it complains with message \"sh: ps: command not found.\" (very many times) Although this does not seem to harm build, it is desirable if this can be solved quite easily by some simple patch.\nI found a similar problem report in macports. Shall we include /bin/ps visible in cmake environment in darwin systems?\nIs there no way to hard-code the path of into CMake?\nisn't actually in .\nThe relevant code in cmake is I believe. The package on contains the Apple ps. I'll likely get it building as part of my revamp of the darwin stdenv, and if I'm feeling particularly charitable I might even hook it into cmake. In the mean time, I'd probably just leave this alone, since I don't think we're far off from having a sensible darwin stdenv.\nWe definitely can't use procps on Darwin since it doesn't have /proc. Also, these warnings are not always harmless: MySQL's startup scripts won't work unless it has ps in scope during the build process. Right now the workaround is this\nThis was fixed as part of the new darwin stdenv."}
{"_id":"q-en-nixpkgs-e5c440d10e43d395105e2a681e2406cd2d750e49f5d92098847d62343073ed8f","text":"Project description can produce completions for a range of shells, including Bash. These should be integrated with the pijul package to improve usability. Metadata homepage URL: source URL: license: gpl2 platforms: all"}
{"_id":"q-en-nixpkgs-a1f293d67d4ad432d8fc0e64c287fa39806d74764f4dd2c6cae066ff8897980b","text":"Project description Compute metrics on Python code (e.g. cyclomatic complexity) Metadata homepage URL: source URL: license: mit platforms: unix, linux, darwin"}
{"_id":"q-en-nixpkgs-01f92938c13bbb59300dd859b0678d40863da4389c062fa319c809175f7a9744","text":"We need to propagate the pth loader so packages using us process our pth files, but it makes no sense to have the pth loader itself in a pth file."}
{"_id":"q-en-nixpkgs-b5d401b2f7df1860db058b75af0fdb1aabbf77e1d3e343f5cd374b35d681afba","text":"honk-initdb.service attempts tu run even though already exists, causing it to fail and prevent honk.service from starting. I think honk-initdb.service is supposed to be skipped since already exists? Please run and paste the result.\nHi, Thanks for the bug report. That's weird, because that service shouldn't run since the file already exists. I will try to investigate during the week.\nDo you know if there is any way I can make honk.service believe its dependencies are met, as a workaround?\nWhat I did precisely in I have no clue why this is not working yet.\nI pushed a PR here: Could you try it and let me know if this is fixing the issue ?\nDear is your issue fixed now ?\nPing ^^\nThe condition seems to work but the issue remains?"}
{"_id":"q-en-nixpkgs-05bd29d60426de0d23d5c82094a8c40e7a6d0819049a8c1047c9a8754a432093","text":"Package name: cryptomator Latest released version: 1.10.1Current version on the unstable channel: 1.9.4 Current version on the stable/release channel: 1.8.0 Notify maintainers Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-38b5b588497a6d1ca29a8b493457cc948efd104efcbf067f70b23d3d9e384680","text":"Package name: rio Latest released version: 0.0.27 <!-- Search your package here: --- Current version on the unstable channel: 0.0.19 Current version on the stable/release channel: N/A <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR.\nSeems one of the dependence of depends on which not entered master yet (see ). I'll handle this once it gets ready.\nNot sure why it requires it. Maybe removing this requirement?\nAnd it's an critical component that could not be remove.\n1.72.1 merged into master\n:+1: Looking forward to this as well. Hopefully fixes\nStill not merged into master, merged into staging-next though.\n1.73 has been merged into master.\nI sent the update"}
{"_id":"q-en-nixpkgs-85a3255eff8f82bcc14419f5837f68dbcc4811301ee08363487252ccbf82ab34","text":"Package name: qdigidoc Latest released version: 4.4.0 <!-- Search your package here: --- Current version on the unstable channel: 4.2.12 Current version on the stable/release channel: 4.2.12 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-bbd0a67a25f0272b3f194b1b0a24f977223aff96f2f7fc62893b92a57fce96ee","text":"Steps to reproduce the behavior: pyside6 Please run and paste the result. :sweat_smile: I guess it says not found because i have no channels (i use flakes)\nThis same failure is causing syncplay to fail to install on my system.\nyeah thats how i knew lol\nTo confirm, I can reproduce this issue.\nI'm trying to package something which relies on pyside6. Package builds fine on nixos-stable but fails under nixos-unstable due to pyside6 not building."}
{"_id":"q-en-nixpkgs-5087bc0963f3fa7c14838522d0da4c5bea4ad2118de99eaf80f881e39adc93ce","text":"The , which . I do not have a short reproducer. However, the issue was observed on in a cgo context. The issue can be resolved by applying the following change to : The error should not have occurred. The code under test used to work, and failed after an update of the nixpkgs revision in use from to . n/a The commit range includes the following two commits that touch on : adds to the strip setup-hook adds to the cc-wrapper for response file handling The latter one seems to be the cause of the issue. Similar issues were described in and <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nSo the wrapper picks up the binary from the system instead of the one from the store because the environment is not pure? This makes the behavior of this and other wrappers unpredictable. Maybe it would be better to make sure that the tools used in the wrapper are actually from its real runtime dependencies in the store?\nYes, that matches my understanding. Yes, that thought crossed my mind, too. However, I saw that there is an , which is then . And this import and seems to be . I don't know the context or reasoning for all this logic and didn't venture there. But, that may very well also be a good solution.\nAh cool! So it basically already works that way. Setting actually enables the import. Your explanation led me to find where this logic was first introduced. The discussion there is very interesting, afaiu we need to restore to its original value so things like linker scripts will still find the intended executables. I think that the commit that introduced the should have it before the path gets restored.\nOh, I see, thanks! I overlooked one negation. That seems like it could solve the issue, too. Perhaps it's easier to just resolve before the is reset and then still invoke it later, since . I had opened that replaces the invocation by a portable one. That approach was easier to test in my use-case without rebuilding the world."}
{"_id":"q-en-nixpkgs-c0fd456f96275ea4d47dc7d3d728bc0fe90340f50567493a3daa68c2867caf1a","text":"Steps to reproduce the behavior: as such, so that it builds with Airplay 2 support: Strangest thing is that I have 4.1.1 built and working for a while (probably at ) but following a it does not rebuild anymore, even though and hasn't changed, while the only two subsequent commits are and which I'm not too sure how they could relate. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result."}
{"_id":"q-en-nixpkgs-827c17d681460e80b42a8ed932a827383e34d26720c57599782e1f069c63686d","text":"I have transmission running inside a Nixos container which has been working without issue until I recently updated the system. It now fails to start inside the container, I have tried deleting the container and all its data but the same issue exists when creating a new one. I have also tried creating the path \"\" on both the host and container but it doesn't seem to care. Steps to reproduce the behavior: Run transmission inside of a container Transmission service starts and runs without issue Service status Here's where I have the container and transmission defined in my config. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nSame here, not sure if it's an upstream or nixpkgs issue\nFor the time being and since it's inside a container anyway I'm overriding the root dir of the systemd unit:\nThanks your fix is working for me!\nThe problem is that is listed in : Which got introduced in to The original problem was that was not in the chroot. I'd say mounting all of as read-only is going too far, so a better fix would be to just mount and only if is true. It also seems weird to have this logic be part of Transmission, but whatever. So a workaround is: (And I'm not sure why this wasn't a problem for 3 years)\nI've stumbled upon this issue as well, and wondering if there is anything we can do about it regarding the Transmission service. I had to disable the following is well, before it played nicely in my setup.\nindeed, adding only was one the alternatives suggested in , so we can try your suggestion. IMHO disabling or should not be the default, we could make them use though, would that be enough for you?\nThis is actually still an issue and it isn't limited to containers. Adding the params that posted, unfortunately, does not work for me ;/"}
{"_id":"q-en-nixpkgs-8ceb893ebcf06fed1761387cd65712bc4b4bd9c8a0d994c5149cd9689fd77994","text":"Package name: Latest released version: <!-- Search your package here: --- Current version on the unstable channel: 0.37.1 Current version on the stable/release channel: 0.24.1 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-899ebb741e642eb761d2992a515a9915ad9d7cdc45673abc3f830027e69e8b3c","text":"if a folder is a substring of another folder, we don't delete it even if we have overrideFolders and the substring id got removed from the config. Steps to reproduce the behavior: a folder with id \"xxx\" a folder with id \"yyyzzzxxx\"; folder with id \"xxx\" from config never gets removed folder \"xxx\" should have been removed after removing it from config <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nThanks for spotting the issue and reporting :) It sounds to me like a issue not being accurate... Maybe only the 1st commit in does the job?\nyes, the first commit solves the issue with substring ids. the second commit solves the issue with ids like \"hello/world\". I tried escaping the id to send a DELETE request but couldn't find out how to correctly escape the / in the id\nI'd never imagine someone would do that LOL. What if we'd impose a check in the module that would prevent such ids? Is it even supported upstream?\nwell it works in the webinterface. so it seems supported? also the default value for the id in our nixos module is the key, which also is the path. I usually do something like so all those ids are broken. as a workaround we could replace all / in the id, but that would break interaction with non nixos setups. PUTing the config seems to not have that issue. Were there any problems with PUTing the config if we want to override folders and devices anyway?\nCould you perhaps give more context? I don't understand what kind of a json object is passed to the commands when you do that. I also observed in the below code that \"/\" are taken care of in a special way? As I in , I don't like how the code diverges in the , and it seems to me it doesn't handle the case where you want to delete a stale folder, and you don't .\nyeah the PR sadly isn't complete, I just used it as a test for my local systems where I just override everything. I guess we need to handle individual overrides better, but not sure how. setting wold result in a nixos configuration of when I initially wrote this module this seemed unique enough and like a sane default. I thought about hashing the key for the id, but this would have ended in unnecessary long keys.\nI see. I think that using such IDs for folders and devices is very uniqe usage... I tend towards suggesting to take the 1st commit of the PR, and a message when evaluating the script, in case the folders' IDs contain \"/\", and is true. However, that would not solve your particular, usage. Perhaps we can talk with upstream about it? They are rather responsive from my experience.\nthat sounds like an idea. I'm busy the next couple of days, but you can cherry pick the first commit from my PR and close it and we will continue with this once we have some better idea how to handle things and we maybe have an answer from upstream?"}
{"_id":"q-en-nixpkgs-05254bfb7e82c7a34c3771640732e4f77819a6d0093e10818e10e519fc624228","text":"Project description Colorful, adaptive gemini browser <!-- Describe the project a little: --Metadata homepage URL: source URL: license: gpl3 platforms: linux"}
{"_id":"q-en-nixpkgs-6d34722e099db80277efebd004f8f8f5f99f66b9372d9d3ed71c8b6d8189479e","text":"If you add enough (minimum number I haven't searched for) settings, then the script of applying these settings via api will not be able to execute completely and curl will generate \"CSRF Error\" error, this happens because syncthing resets api key after applying a certain number of settings Steps to reproduce the behavior: services.syncthing several devices and options logs of syncthing-init.service, there will be errors syncthing-init.service will execute entirely with no errors and will apply the entire configuration without errors ! I made a vm with the minimal configuration at which the error is reproduced I used nixpkgs commit Config for vm - Steps to reproduce nixpkgs - vm config - vm - vm, login as with password \"CSRF Error\" and parse error from jq Also during testing I managed to fix this by parsing the api key before each curl call, this helps but looks like a dirty solution <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result.\nHmm indeed we recently changed a few details in the script in: I tried to write a nixos test to test some of these issues, and I experienced numerous other bugs, and I sort of reproduced the issue as well. I Reverted In: And reopened: cc .\nI have this issue with (vm config, that i provided, uses this values too) So i dont think reverting will help\nCould you try to merge locally and test?\nI tested vanila with as patch on top. In both cases, the issue persists\nOK. Thanks for the report!\nAs I said before, getting the api key from config before each curl call solves the problem (an example of this solution is in the patch), but I don't think it's a good solution, probably we should check the result (or/and stdout) of curl, reparsing api key if needed, call curl again, before trying to parse it with jq Patch:\nI also found that to add several different devices you need to have different id's, so I tested with four devices with different id's. With my patch, all 4 devices were to correctly and syncthing-init.service ran correctly and without errors\nI also found out that during my testings in . You are invited to submit that patch, with better formatting. I am also looking into finding a way to add dummy, but valid devices ID, for testings.\nOK I managed to fix the test, regarding checking that the script indeed deletes devices that need to be deleted. I haven't shared it in , because I use there private (sort-of) syncthing's devices' IDs I took from my personal configuration. I'm still looking for a way to generate valid dummy devices IDs. My help request is at syncthing's forum: I also applied a better version of your patch to the (currently draft) . I plan to add many more devices to that test, and maybe do the same for folders. Then I'll also test issue . In the meantime, I'm waiting for a dummy device ID generator.\nI managed to create dummy syncthing device IDs and I am using them now in my PR . I saw your PR , and I think we can concentrate our efforts together on my PR at , where I'm currently failing to reproduce your issue, your comments will be appreciated. I am testing your issue without the 2nd commit in my PR (that is identical to your commit at ), and I'm testing that is fixed by reverting commit just for testing (this works :))."}
{"_id":"q-en-nixpkgs-04c1aa9d7f2fe9537302209de55cb02aa1d249616a2bbf72f64314bafe04d8e6","text":"Steps to reproduce the behavior: llpp Works on 22.05 05 probably depends on insecure with:\nI don't use this package, but I looked into it because why not I guess It looks like it builds if you bump the AUR patch to the latest one but only if you also include (which fixes a derivation bug but was reverted because other packages in nixpkgs have workarounds that stopped working when the bug was fixed....)\nohh, thanks for looking at this! I build it with that version of mupdf and the aur patch, and it does build fine, but I get a runtime error now if run without arguments . Still, it does seem able to open pdfs now, so it's sort of usable :p\nHmm I don't actually seem to get that error when running it without args I don't actually know how to use this program though\nI'll still merge in a bit to fix build for now, but I may have gotten sniped into taking back up again and will hopefully making a staging PR today (because we really should fix that....)\nSorry! :p\nSnipe is in now"}
{"_id":"q-en-nixpkgs-1b4be95db975bd54ecf674e0d3986ebfb106007e4dade0dde5df71429920b8a3","text":"only returns the hostname, even when both and are set. Steps to reproduce the behavior: and . . should return the FQDN defined by and . The breakage happens when is used, which is now the default in NixOS after Stopping causes to return the correct value. reproduces the issue. It broke at , which is when became enabled by default. It's still broken on which is at time of writing. Possibly related issues:\nThis seems to happen because the resolution flow (when nscd is running) looks like: calls , which uses the nodename (), which is the short hostname then calls , which invokes the nss machinery gets the request and processes the request according to the NSS rules; on my NixOS system this is (from systemd) doesn't match (from systemd) does match, and returns a synthesized record; systemd internally determines the hostname for this using only the nodename, so unless the nodename is already the FQDN then we fail here, and it does this before checking /etc/hosts. I wonder if we can fix this by patching out the support in systemd-resolved for looking up the nodename, which will make it fall back to the result from /etc/hosts instead; on NixOS we can (probably) always assume that /etc/hosts will actually have the correct system hostname in it.\nNo, this is most likely We wanted to polish that PR in the next days, and once merged, bump our nixpkgs pin (and resurrect the upstreaming attempts). If you want, can you check if using nsncd from this PR fixes the problem for you?\nAh, actually, I think I'm just talking out of my ass. systemd-resolved seems to be able to cope with this properly when I query it over DNS, so it probably isn't to blame.\ntests fail at that PR. Setting gets a successful build, which then passes the test in , but it fails on the test.\nThis is a pretty weird behavior. What's (likely) failing here is the IPv6 resolution of . The test does pass on some NixOS setup, like my desktop (and the Ubuntu CI) but does fail on my laptop. I'm still not sure what's happening, is present in . There's something wrong with some NixOS setup, it'd be nice to fix that as well. By any chance, did you disable systemd-resolved? Thanks for the heads up wrt. . I'll have a look today.\nThe behavior I described is from the Nix build environment. I used the nixpkgs derivation for nsncd and updated the source to the PR."}
{"_id":"q-en-nixpkgs-ae97421fe83a20c66bf2fb6e9e6c17b3f7383df7e1587f9f7c6e31f15e2b6b5b","text":"Another security release for Exim: 4.96.2. See CC maintainers:\nPart of the report also concerns libspf2. (For some reason, it has been filed against Exim instead...) The patch(es) should probably be cherry-picked as well onto helsinki-systems/libspf2 (the fork that we're using)."}
{"_id":"q-en-nixpkgs-cfc944e3ff75cd8b88d714d9d00b36dc72ad9d24f7b574bf86ed37fab89d7275","text":"Ditto. I have noticed this accidentially whily working at . Nixpkgs version: (run ) cc\nYou can puts secrets in an additionnal initrd readable by root only, and use See\nThere is more generic option now, . We probably want to merge those."}
{"_id":"q-en-nixpkgs-b3f084b1e1c486f9d7abf384470c9aa7e305ae8cac8b831579841780485584b2","text":"Package name: mautrix-googlechat Latest released version: 0.5.1 <!-- Search your package here: --- Current version on the unstable channel: 0.5.0 Current version on the stable/release channel: 0.5.0 Notify maintainers Note for maintainers: Please tag this issue in your PR.\nThanks"}
{"_id":"q-en-nixpkgs-70de878a69963fd650a7b36ad4e5d078d04724de99af3553252b8a3a6b3217f6","text":"introduced a deprecation warning for the option. Upon following the warning and moving all options out of the option and removing it, the build fails with the following error: Steps to reproduce the behavior: the service without using the option the configuration Build succeeds. As a temporary workaround, I tried setting to an empty set but then borgmatic fails to parse the config file. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result."}
{"_id":"q-en-nixpkgs-e7a2d0e600d949ecb1663a7b4d17ecdc6217a5d7e97262740ea3626b88e1eb3c","text":"Attempting to upgrade to kernel v6.5.x results in the arm64 prl-tools package (18.3.0-) failing to compile. Steps to reproduce the behavior: NixOS in a Parallels VM from kernel version v6.4.x to v6.5.x nixos-rebuild to complete without error Using allows an upgrade to succeed with no tools installed. Attempting an upgrade with the unstable version of (19.0.0-) results in a similar build error. Please run and paste the result.\nJust noticed a new DMG is available with Linux 6.5 mentioned in the : Not sure how to patch locally to test this.\nI've already opened for that."}
{"_id":"q-en-nixpkgs-387125a2f0f38186733d341ea4d5c5b917da79451ad8be4c1b72090ab7e986f4","text":"Steps to reproduce the behavior: , or anything depending on it (my failure specifically was triggered by udiskie). Looks like the exact same issue that is being reported in Hydra, on the same hash: Seems similar to reports that it was generated with cython 0.29.3. Unfortunately also seems to be archived and a cursory check does not seem to show a fork floating around.\nRelated:\nUpdate: I was able to rebuild udiskie successfully by changing its build input (and all respective downstream references) to , so I'm guessing this might have only surfaced now with the recent cut to python 3.11 (maybe ?)\nAnother comment looks like . I'm guessing we could do the same here."}
{"_id":"q-en-nixpkgs-b4614fbfe8566b268c08b5cb8960f3b77ef29375bd873c256e172a9b90c95939","text":"Package name: cardinal Latest released version: 23.10 Current version on the unstable channel: 23.09 Current version on the stable/release channel: 22.12 Notify maintainers Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-033761f628d08c61451368a862b3f183f35fd609ca2ea019493bab0c950ef76c","text":"Vesktop opens up to a white screen. It cannot update from tray to latest version. I tried completely removing config files from Vencord but it doesn't change anything. Running Vesktop with flag --disable-gpu makes the Vesktop stuck at loading Steps to reproduce the behavior: Vesktop Launching Vesktop should open up the app normally ! Kernel: 6.5.7-xanmod1 Gpu: AMD (mesa 23.1.7)\nvencorddesktop command doesn't output any logs. When doing update from the gui, it doesn't show anything. Only time it shows something is when i try updating from tray icon. The output is the following:\nThis is a problem with Vencord. It is currently broken with the curren versions of Discord (including the web app) We have to wait for upstream to fix these issues and release a new version of vencord\nAlright, thank you for the fast response! Cheers\nI encountered the same issue, and can confirm it is fixed when the system version of vencord is 1.6.0."}
{"_id":"q-en-nixpkgs-a4f0fc8675ec2c71f838798e6b203344c6e11638e77c590104777d630ec912ef","text":"So, whenever I open Player, it starts installing Roblox, finishes, and does nothing. Even when launched from the website, starts installing, finishes, and most likely just exits. Logs show nothing important. However, when trying open Studio, the error saying \" not found\". Steps to reproduce the behavior: the grapejuice package (just write grapejuice in the with pkgs section) wine-wayland package (not sure if it is required, however I did) Grapejuice (I used rofi (drun), but I suppose $ would work) the message box that pops out (from wine), press \"Install\" to install the wine-mono thingy press the Open RobloxPlayer button, or go into Player, Initialize and Install Roblox, the behavior is the same Roblox installing and doing nothing... I expected RobloxPlayer to actually open, not install and exit... <!-- --\nThis is already tracked in issue"}
{"_id":"q-en-nixpkgs-27ff87609c9aa401dc3d9b566392a6564da4c240af6b5e764aad2ecff5adc80b","text":"I'm trying to use from nixpkgs on my laptop and use remote builders (buildbarn). This fails due to patches that are applied in nixpkgs. Vanilla Bazel has a hardcoded fallback path for bash () . This has the downside of not producing a working fallback for (bash from nixpkgs will default initialize as ). The workaround for this is to default initialize the path variables somewhere else. Currently, this is done . This works fine for local execution, since this is where this function is called. However, this function is never called for remote execution (it is a hack to override things for the local execution environment only). So when using Bazel from nixpkgs locally and enabling remote execution, the shell is set to a bash from nixpkgs during the analysis phase but the variable is explicitly left unset by many actions, leading to the PATH in bash being and resulting in errors when executing tools in a script. I think there are two possible solutions: when using remote execution, do not replace the hardcoded (this is probably not a great fix since it only works for FHS executors) instead of replacing with a regular bash from nixpkgs, add a little helper script that adds shell utils to . This way, any shell rule created by bazel has a minimal set of tools. use when turning on remote execution. This has the downside of only working when the remote builder has a shell in that path. See this for a detailed reproducer: Using bazel from nixpkgs should work with remote execution. Using this version of bazel with remote execution is becoming more viable due to <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- [^1]: [^2]:\ngot reverted, I'll reopen it.\nCould we also remove the related patch here -, or is it still needed for some reason ?\nTo the best of my understanding, it should be sufficient to use a prepared bash (with tools in the PATH) and remove the patch you mentioned. Probably still a good idea to test this in practice!"}
{"_id":"q-en-nixpkgs-10410d0fcbe800047507a4433c9e82a736f4e7e267da37defe66d64348ffc55a","text":"Building the python 3 protobuf package fails for me on macOS. Digging in (by launching ), the cause seems to be that calls instead of . The call above fails, but the same call works after replacing by . Similarly, works fine. That in turn matters because of in the / wrappers, because the system include path is missing in the case. System: mac OS 10.12.5 Nix version: nix-env (Nix) 1.11.8 Nixpkgs version: \"17.09pre108299.\" Sandboxing enabled: grep: No such file or directory :) When digging in, I used a checkout of nixpkgs with hash , but it seems to make no difference.\nseems related.\nuses Python on Darwin and might help here.\nI've had a look at the python setup tools: The relevant line is (rather: appears to be, I haven't tested this) which causes C++ source files to be compile with . Not sure whether this is the right or wrong thing to do... I see two reasonable ways to fix this bug, in a more principled way than : python distutils to compile C++ files using nixpkg to treat like with respect to and friends If 1. is the right thing to do, it might still be nice to work around the bug in another way for now.\nI've also been bitten by this problem once: The core problem is that setuptools pretty much ignores the difference between CC and CXX. This however is not a problem on a standard mac xcode installation, as both and link the C++ standard library there. However our nix-clang does not link against the C++ standard library (and probably should not) and that's why you will encounter this problem every once in a while when dealing with darwin and c++ python extensions. I'm not an expert on the darwin-stdenv, so I wouldn't know how to fix this in a way that does not require hacking our ways around that problem. To hack around this problem you simply have to grep the pythonPackages for . There are a few packages that use a workaround that involves adding to the C-compile flags: See e.g. here:\nHowever our nix-clang does not link against the C++ standard library (and probably should not) and that's why you will encounter this problem every once in a while when dealing with darwin and c++ python extensions. any clue how to solve this issue?\nThe problem does not seem all-too widespread, but there are a few instances:\nAny reason we can't just change to use ? Seems like it would work on Linux just fine, and fix a bunch of other stuff.\nI knew this was an issue in general (), but assumed the problem was something more complicated.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-02ad1a6f45a5a595572378b117b115dbe6df3659988d6486e0495a96fc4470b6","text":"Project description Gruvbox Plus icon pack for Linux desktops based on Gruvbox color theme. Metadata homepage URL: source URL: license: GPL-3.0 platforms: linux"}
{"_id":"q-en-nixpkgs-6243d49522b4f5c2644cc846aaa30f17f3f5464cb1e4cb3b2f6fc4ae1d60b61f","text":"Currently we build with support for . Sonnet typically looks in for dictionaries That means it won't be able to find any dictionaries and thus spellchecking won't work.\nIs this still a problem? It seems that hunspell paths are not hardcoded in hunspelldict.h anymore: But they are still hardcoded in\nYes, this is still a problem. It seems like we should disable hunspell and use aspell, which is configured correctly in NixOS.\nI marked this as stale due to inactivity. &rarr;\nI marked this with \"good first bug\" because I think we need only to change to depend on , and do some testing.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-5fbc2bd2deffa5f3ea9ed822fb5bb7b55944212d65faad109f640c9728b48916","text":"Lily and I talked about this in Matrix, this is the requested issue. Basically it all narrows down to this package lock: Steps to reproduce the behavior: it it hangs forever lock it works It should either give an error or just work. If it helps narrowing the npm version down, the package- was created in September of 2020 and was last changed in September of 2021.\nTo clarify a bit, this isn't something specific to older lockfiles, it just looks like there is a version spec format that our fetcher does not currently support, but probably should"}
{"_id":"q-en-nixpkgs-79d4a77b3ac24ff9398e401c7f64feb80c2a2da86d715f571bb0e731f11121a2","text":"Package name: gdtoolkit Latest released version: 4.1.0 <!-- Search your package here: --- Current version on the unstable channel: 3.3.1 Current version on the stable/release channel: 3.3.1 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers It's a bit of an odd situation because both Godot 3 and 4 are in use, so maybe this one should be renamed to gdtoolkit3 or the opposite for the time being. Haven't had any success yet trying to override the package on my own. <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR."}
{"_id":"q-en-nixpkgs-28bf3fa3ce983bdd0c6d7184dad89ef64d573f50d79026cb3eb88a6428ecb13c","text":"Package name: gitea Latest released version: 1.21.1 () <!-- Search your package here: --- Current version on the unstable channel: 1.20.5 Current version on the stable/release channel: 1.19.4 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers: <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR.\nthanks. I've created:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-be670864a82e5d25f25c3759cd857175c5d3ef50cb6439906afc88cddfe973d3","text":"ssh2john produce error while trying to process ssh key Steps to reproduce the behavior: john in any way - ssh key with password - KALI: Please run and paste the result.\nThis seems like an upstream issue. See\nThis issue 2 years long! Can we bundle older python with john or apply workaround stated in mentioned issue?\nIn the meantime, I believe you should be able to work around this by overriding the python version locally:"}
{"_id":"q-en-nixpkgs-fdfc07fcc964296356047370a1065f826de0728310e55c24722b3a8ddb27b1e5","text":"Package name: handbrake Latest released version: 1.7.3 Current version on the unstable channel: 1.6.1 Current version on the stable/release channel: 1.6.1 Notify maintainers Note for maintainers: Please tag this issue in your PR.\nRevised to 1.7.2 release\nRevised to 1.7.3"}
{"_id":"q-en-nixpkgs-126012b178288cb3e9c25249e096a87c68e9896c8b7967f3a999886f54b2e953","text":"Hi there I installed but the driver is not available on CUPS; the printer (Brother HL-1223WE) is detected correctly. My system is on . Steps to reproduce the behavior: to to add the printer on CUPS.\nSubmitted a with a fix for this issue."}
{"_id":"q-en-nixpkgs-26644850fff55792f704d07b3719409a75ac6ee71b6b077d158c520acbc6aabb","text":"Package name: vcv-rack Latest released version: 2.4.1 Current version on the unstable channel: 2.4.0 Current version on the stable/release channel: 2.3.0 Notify maintainers Note for maintainers: Please tag this issue in your PR. Priorities Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nmay be trying something with so I'll wait a bit\nHonestly I have no idea what vcv-rack is, I just looked for a package I can update to test out a new feature for `nix-update. If you can, take this from from, feel free to update it.\nNot sure if I'll have time in the next few weeks, so this is free to pick"}
{"_id":"q-en-nixpkgs-6ba82bf4cc1c969c49f306d0a6275e0aa125ed82e0d273d32cd1162c4360e961","text":"Building on macOS fails in the test phase, since the path to the nix-daemon's socket is longer than the maximum 108 characters. This is a combination of macOS (long TMPDIR set to the build directory). I'm not really sure whether this is more of a nixpkgs or a nix issue. The problem can be worked around by setting TMPDIR=/tmp manually, but it seems to me that that should not be required... Compare nix commit , which removed a workaround for this limitation. System: Darwin 16.6.0 Nix version: 1.11.11 (building nix-1.12pre5413_b4b1f452) Nixpkgs version: master as of 2017-07-04 Sandboxing enabled: no"}
{"_id":"q-en-nixpkgs-32924df150481cf821f2b96bf2f0dc1a7458dc0f82eba109978d8f915b930d7b","text":"Project description 'cp' and 'mv' utilities with progress bar patches The utilize it (flag -r) to show cp, mv progress. Advanced Copy is a mod for the GNU cp and GNU mv tools which adds a progress bar and provides some info on what's going on. It was written by Florian Zwicke and released under the GPL. This repository exists because the website appears to be dead. You can still find it via the . More info: Metadata homepage URL: source URL: license: GPL platforms: unix, linux Priorities Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThe after running the in the advcpmv repo would be and need the binary with name and for the flag -r to work. So symlink and to and respectively.\nIt is a patch against coreutils, basically?\nYes\nOn a hand, it can simply be treated as a , or even better, a NUR-like overlay. On the other hand, coreutils is a core (pun not intended) package for Nixpkgs, a dependency for things from fetchers to stdenvs. Modifying it has the potential to be a mass-rebuiding mess. Indeed, patches for coreutils should be \"vendored\". Let's try.\nSorry I am so new to Nix and NixOS (2 weeks usage), I don't even know it would mess things up. I don't know much about , isn't it just permanently modify and in place? I have learned something about writing a derivation this week, and here is the .nix file I am using and work just fine for me:\nNix is source-based - it builds the binaries (and artifacts in general) from the sources. Think like picking Archlinux but from the PKGBUILDS. Despite being source-based, Nix can rely on local and remote binary caches to speedup the building. Certainly, when a package source is updated, its old entry in the cache is invalidated and the new version is put in its place. For an application like, say, Zsnes emulator, it affects only Zsnes. On the other, for a library like, say, SDL, its update invalidates SDL and everything that uses SDL - zsnes, bochs, frets-on-fire, dosbox, mednafen... the list is huge. And it applies to everything, from the core utilities like coreutils to desktops like XFCE. provides a better visual analogy.\nThanks for explaining it to me. I think it would be better not touching those basic packages.\nI have tried this and it worked: Now I am thinking if it's better to put this in my (future) NUR overlay or in the main Nixpkgs.\nI think that this issue should stay open because it is not necessary to edit the coreutils packaged in nix. One can just make another package that only includes advcpmv commands like arch does:\nThe problem is that advcpmv is a patch that modifies only cp and mv, while coreutils has a truckload of extra commands. It can cause conflicts with existing packages.\nI just tried to rebuild my config(was working before), got: What's the suggested resolution here?\nFrom this it seems like, the only way out is to not use this package anymore? It's an dependency to some nnn file browser feature. I'll look for alternatives for it.\nTwo weeks ago . If you are interested in keeping advcpmv - and receiving because \"it will break the world\" in a future update of Coreutils -, feel free to pick it."}
{"_id":"q-en-nixpkgs-e2c19866638128d58ce3232f454eb4a8241e4d71bcaa71f7534ad82b63dfdfb9","text":"Project description OpenJDK 22 was released on 19-MAR-2024. The OpenJDK project has been releasing Early Access (EA) binaries for many months now. It would be nice to be able to install OpenJDK 22 as a Nix package and also to begin development of dependent Nix packages (e.g. ) Metadata homepage URL: source URL: (update: is open now) license: GPL with Classpath Exception platforms: unix, linux, darwin, others I'm fairly new to Nix (but loving it so far) so I'm not 100% sure this whether this should be a \"new package request\" (as Nix keeps multiple versions of OpenJDK available as different packages) or a \"package update request\" (since jdk22 is an updated version of jdk21.) If I need to change this or anything else in the issue please let me know and I will try to fix it."}
{"_id":"q-en-nixpkgs-e5ffa001bc4dafa23ae77bccdfba8bd7166c00dd084cf70050c94db767811c61","text":"As of today, this search yields 3 items for Corretto. The \"Source\"-buttons for some of them point to OpenJDK even though the package is Corretto. Corretto has been and is based on the corresponding OpenJDK version (using ). At a first glance, the the usage of looks like the issue. However, when searching for , the \"Source\"-button for correctly points to eventhough the latter also uses .\nTransferred to nixpkgs as nixos-search just uses , which is determined by in . I am not sure what is going on here exactly, but note that does seem to have the correct position.\n:arrow_up: I left some context in the mentioned above. I realize that the actual problem is , i.e. the OpenJDK derivations define their and these get inherited then."}
{"_id":"q-en-nixpkgs-d29918958de6ec85ff31caf45d23b57ad73d24a124c92b1fcb4459414e41cd0f","text":"Steps to reproduce the behavior: type: I have none. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nContext: caused because last supported version for 3.0.0 is llvm 15: Should be easy enough to fix, thanks for reporting."}
{"_id":"q-en-nixpkgs-c8872488bb44ebe5f8b771dee3a966e36a8a5ea8ef6c45cb18eb5efb46e673f2","text":"Maybe related to build from master branch - Hope you don't mind the ping, sorry if you do Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI can't reproduce this. What does give for you?\nMy bad in naming the PR, the package does build but the man output of it points to broken symlink I am building on master branch, commit system x86_64-linux.\nWith some trial and error I was able to pin point the commit which broke it That was merged in See here my build log"}
{"_id":"q-en-nixpkgs-955ddfc34377b103a3454c1202aca565933e93c662f05be19085662887dabb9b","text":"It looks like , so the program crashes on every run. See . No crash. system: host os: multi-user?: sandbox: version: channels(root): channels(username): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-7be524d468c9cba2e8dd3068a6f6a468ab41590f505d74a45f0994e8c19854b9","text":"Apparmour service doesn't reload correctly after a system switch. Steps to reproduce the behavior: anything in config with apparmor enabled nixos-rebuild switch error of apparmor service failing to restart No error, apparmor starting as it should Here my service log: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI'm encountering the exact same error here.\nSame issue here\nSame for the issues after . However do a after that just fine.\nSomeone once told me that the state of AA in NixOS wasn't great. But I don't know first hand. Could be wrong. I lack any knowledge at all to even figure out the root cause of the issue. :/\nAA seems to work fine in general. Just the service restart after a nixos-rebuild doesn't work.\nTested just now on nixos-23.11. AppArmor service restarts without any issues after nixos-rebuild. I think this issue is resolved.\nReopening this issue as unresolved. Currently on 24.05.. and this issue seems to be persisting after a nixos-rebuild: reloading the following units: apparmor.service, dbus.service Job for apparmor.service failed. See \"systemctl status apparmor.service\" and \"journalctl -xeu apparmor.service\" for details.\nI did some more investigation on this. I have one system with a custom apparmor profile and another system with no profiles. On the system I have with a custom apparmor profile, nixos-rebuild works fine. On the other system without any apparmor profiles, nixos-rebuild fails with the message in this bug report. After nixos-rebuild, journalctl reports: In I tried running \"sudo aa-remove-unknown\" from zsh: The error message matches what is in the system log. I think this may be related to the aa-remove-unknown step.\nAs a workaround, adding this to appears to work: After this, the first nixos-rebuild switch will still cause the issue (because the profile is not installed yet). But on the second and subsequent nixos-rebuilds, there is no error message on my system. Can someone else try it and see if it works on their system too? If so, it may help to figure out the proper solution.\nI tried it and it actually rebuilds without error the first time too. Hopefully that helps you or someone else figure out a permanent fix! Many thanks.\nFrom my side the issue seems to be be fixed with the merged fixes. So I will close the issue.\nIf anyone wants to track the merger of the fix :"}
{"_id":"q-en-nixpkgs-047ac40931e0051863838f9150da52f9fb5b679fba18489c3a48cc897c0631fa","text":"Steps to reproduce the behavior: obs-studio Please run and paste the result.\nCouple things: 23.05 will be EOL soon, just a heads up the rev as specified in your \"host os\" line of the output, it pulls from the cache. I can only assume you're using a different rev or some override/overlay. Even rebuilding from that rev in my case produces no errors. Can you provide a config to reproduce?\nMy config is here: I am not using any override/overlay, afaik...\nSetting is causing some massive rebuilds on your end and likely causing the breakage here. That's also why it couldn't just fetch from the binary cache. Try without this line.\nThough FWIW, I'll see if I can get the build working with , it probably shouldn't just break."}
{"_id":"q-en-nixpkgs-3360c271ccfa0164890ffe16d669d572237e509bf07062029b78efa5640cb769","text":"Mathematica fails to start. .... System: NixOS: 17.03.1517. (Gorilla) Nix version: 1.11.11 Nixpkgs version: 17.03.1517. Sandboxing enabled: no\nSimilar issue in ."}
{"_id":"q-en-nixpkgs-ddab0adc4eb82814bb965bf3acdb628f9931406f93513213a9e4603dcc9152aa","text":"Package name: Latest released version: 5.1.1 <!-- Search your package here: --- Current version on the unstable channel: 5.0.3 Current version on the stable/release channel: 5.0.3 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nVersion 5.1.0 fixed compatibility of this extension with PHP 8.3, so the current version 5.0.3 is unusable with PHP 8.3. It outputs these build warnings/errors\nCould you please provide the PR to update this? I think it's a matter of changing 2 lines. Actually, creating the PR would have been faster than creating this issue :)"}
{"_id":"q-en-nixpkgs-ea447d78ded727b242ec0a72a257d4d8cb6fe08e191d95759f43476ff2c77f55","text":"They apparently fixed the issue where the cursor wouldn’t show up in new tabs. Needs backport to 23.11. Thanks! Package name: vte Latest released version: 0.74.2 Current version on the unstable channel: 0.74.1 Current version on the stable/release channel: 0.74.1 Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-56053fd240e858ed4c16c292deedbbfe4ea05abd1759b830ca3f0d10ac77f9af","text":"Add vulkan-tools-lunarg to a devshell and see that the only executable it provides is vkvia. It should also at least build and provice vkconfig Steps to reproduce the behavior: vulkan-tools-lunarg to a devshell to execute vkconfig to debug your vulkan application that it is not found in your path I expected vkconfig to be built and included in the vulkan-tools-lunarg derivation"}
{"_id":"q-en-nixpkgs-043b2d6711b332d7dc4a022cdb30cf7c35554dc0c27343401062af812b1423dc","text":"The file seems to be randomly edited by committers who feel like doing so, thereby ignoring data protection laws governing some regions where our contributors come from (e.g. , , as of before the force-push) guidelines our community gave to itself (e.g. , ). Having our community guidelines ignored at random lets NixOS look unprofessional and can scare away contributors and supporters. Having data protection laws ignored might even lead to expensive liabilities for the NixOS foundation, meaning that the money would be missing when paying for servers or incentivizing contributors. Therefore, there should be some QA when editing that file. A straightforward suggestion regarding this would be to install one or more maintainers for that file itself. These would have to review any PRs touching it before merging would be allowed. The maintainers should provide evidence that they are aware of laws governing personal data in the different jurisdictions we work with, and aware of our community guidelines. commits happening where laws and/or community guidelines are being ignored Not applicable.\nHi, on GitHub and more specifically the nixpkgs repo we almost exclusively handle the technical side of things, so keeping moderation and foundation related discussions low would be greatly appreciated (yes there is some exceptions). We have for general commmunity disscusion. I generally understand your issue, and I'm fearly certian that the foundation takes gdpr serious, and if you actually want to send a gdpr deletion request send it to one of the listed contact methods on If you have a problem with the way someone in a nix* project bevhaves then please message one of the mods privatley or report it in the matrix room. Please keep in mind that Nix, NixOS, Nixpkgs and almost any other Project under the foundations hat are community first Projects and most people particpating do so in there spare time. And that the way you are currently behaving comes across quite childish. And you might also want to note that your github handle, even if we don't track it in the maintainers file, still ends up in the git history anyways because github puts it in the commit messages of pr merges, so not having it in the list basically doesn't matter. The point aobut Lack of QA for changes to maintainers/maintainer- might be valid and if you want to persue the issue I would suggest you do a discourse post to raise awrness and try to convince one or two well-known community mebers to put them selfs as code owners for the file.\nLet's please keep this scoped towards only discussing a QA process for the maintainers file, and not go into other points of disagreement from the issues that sparked this discussion. Proposal in .\nSeems like the default mechanisms for maintainers who get pinged on file changes won't work well here. Most commits are probably contributors adding themselves to the file. If we assume that the TOFU model works for these, only exceptional commits would have to be checked manually, thereby reducing the workload. But this would require a smarter ...\nWell, the maintainer list is currently maintained in code, so the platform where the repository resides should be an appropriate place to discuss how that's done. Of course, one might argue that this is a mistake in the first place considering the unclear situation regarding how to comply with the GDPR with such a setup. If we go that route, then yes, maintenance of the list should also be discussed somewhere else."}
{"_id":"q-en-nixpkgs-04f19b77ba29c074aad76fbbdc402779b95270ba72dc3bd5bb645a0f6635388c","text":"I'm working on upgrading our work repository to 23.11 and am running into an issue building a docker image with . There derivation is: This works with nix 22.05, but fails with nix 23.11 with the following error: Steps to reproduce the behavior: nix 23.11, try to build the above image The image builds and we can set ownership of files in the image n/a n/a Please run and paste the result. Also, we are running branch: , commit\nI tested this a bit. I can change the permissions on the directory as long as I don't reference the unknown user .\nAre you sure you were actually able to chown before hand or was chown simply ignored? It shouldn't be possible in the build sandbox regardless of what tools you preload.\nAs you can see here the chown becomes a no-op: If you would suggest to delete those chown commands, they do not have any effect. We replaced fakeroot and proot with just proot, since proot also can set the uid/gid to root and the combination was breaking things. You can still prepend fakeroot manually but except if you rely on third-party scripts to run without errors.\nAh okay, the idea was to try to ensure the files were owned by the UID that our CI user (outside of the container) uses, so that any file modifications done to volume mounts wouldn't make the files unreadable by CI. It's possible it wasn't actually needed and/or is vestigial. I guess it's a new version of chown that's causing the issue in 23.11? Thanks for the feedback!\nNo. We dropped fakechroot when proot already is enabled as it was causing issues. The chown program was still the same.\nin is included as an example in :\nIt doesn't seem like the is actually a no-op, either, in previous versions. We wind up with permissions issues in CI when we remove the calls, and we're struggling to figure out a way to ensure the directories within the container have the expected user IDs in order to be compatible with bind-mounting volumes as the CI user.\nIf you are able to change filesystem owner/groups without having root, I would like to know how because that sounds like a kernel security vulnerability to me. If you do need to chown, you have to do this at container start.\nI'm not sure I follow. Previously, we were able to use within the docker build process to set the ownership of directories within the resulting container image. That was useful, because we could set the UID of necessary directories to the UID of the CI user. For reference, this is essentially the derivation we are working with: On Nix 22.05, this works, and the results look as you'd expect after running & then running bash on the container: For 23.11, this is the full build output with the error: <details<summaryError Output</summary<pre nix log /nix/store/in5pqapmv1h5q94n4sj961aw5z101vqy-replication-customisation- | cat /nix/store/ffrva7icr8ws2d51955jy302icgmr80s-util-linux-2.39.2-bin/bin: login: /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/login kill: coreutils chfn: /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/chfn nologin: /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin chsh: /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/chsh +++ mkdir -p tmp src cargo +++ mkdir -p usr/bin +++ ln -s /nix/store/m38gwq0w8w7qyjn9s00balyp7cv3m5p9-coreutils-9.3/bin/env usr/bin/env +++ mkdir -p etc +++ touch etc/group proot warning: can't sanitize binding \"/sys/\": No such file or directory proot warning: can't sanitize binding \"/nix/store/q8494z25m447skzy6l4ccyw84nqfcnja-replication-customisation-\": No such file or directory ++ export PATH=/nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin:/nix/store/rvdryxhd9zw6zz7fwwxi6ma55qnblpi8-fakeroot-1.29/bin:/nix/store/811m6jpxyxzhff2wy27xx5r81bqs1dzk-proot-5.4.0/bin:/nix/store/h3ayxf5dk1gdf3s5716fh1ys815qxxag-patchelf-0.15.0/bin:/nix/store/m38gwq0w8w7qyjn9s00balyp7cv3m5p9-coreutils-9.3/bin:/nix/store/01znf87kiw5xx1dj0f7djrnrbg84ij28-findutils-4.9.0/bin:/nix/store/vq4vmndw555m7ld2bi6pq8kr348qvb6a-diffutils-3.10/bin:/nix/store/rwa7qyds01qzxvq7zq3kgnkrzzpw4s66-gnused-4.9/bin:/nix/store/n062zcsmfl9gfp6vfkcg0asb8jjwmy5i-gnugrep-3.11/bin:/nix/store/4v7m4yx07b3anmbznfhihjc8xiizyna9-gawk-5.2.2/bin:/nix/store/mi3pm67ps7c7k11aqki9182ygzg8j503-gnutar-1.35/bin:/nix/store/5c0ancqnpi0cf1h49mv13w68a950s9z0-gzip-1.13/bin:/nix/store/q5cv7r1sh3s4niyq7vm6h1j74394naxs-bzip2-1.0.8-bin/bin:/nix/store/bh7sbl99ygfvw9w96936nrhw46jmcmqq-gnumake-4.4.1/bin:/nix/store/7dpxg7ki7g8ynkdwcqf493p2x8divb4i-bash-5.2-p15/bin:/nix/store/mbh9gx43gsc5av019vsdkazbyiic3f0f-patch-2.7.6/bin:/nix/store/3q6fnwcm677l1q60vkhcf9m1gxhv83jm-xz-5.4.4-bin/bin:/nix/store/a5s2dmba16pw8l52f496gw0b7nli7mbn-file-5.45/bin ++ PATH=/nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin:/nix/store/rvdryxhd9zw6zz7fwwxi6ma55qnblpi8-fakeroot-1.29/bin:/nix/store/811m6jpxyxzhff2wy27xx5r81bqs1dzk-proot-5.4.0/bin:/nix/store/h3ayxf5dk1gdf3s5716fh1ys815qxxag-patchelf-0.15.0/bin:/nix/store/m38gwq0w8w7qyjn9s00balyp7cv3m5p9-coreutils-9.3/bin:/nix/store/01znf87kiw5xx1dj0f7djrnrbg84ij28-findutils-4.9.0/bin:/nix/store/vq4vmndw555m7ld2bi6pq8kr348qvb6a-diffutils-3.10/bin:/nix/store/rwa7qyds01qzxvq7zq3kgnkrzzpw4s66-gnused-4.9/bin:/nix/store/n062zcsmfl9gfp6vfkcg0asb8jjwmy5i-gnugrep-3.11/bin:/nix/store/4v7m4yx07b3anmbznfhihjc8xiizyna9-gawk-5.2.2/bin:/nix/store/mi3pm67ps7c7k11aqki9182ygzg8j503-gnutar-1.35/bin:/nix/store/5c0ancqnpi0cf1h49mv13w68a950s9z0-gzip-1.13/bin:/nix/store/q5cv7r1sh3s4niyq7vm6h1j74394naxs-bzip2-1.0.8-bin/bin:/nix/store/bh7sbl99ygfvw9w96936nrhw46jmcmqq-gnumake-4.4.1/bin:/nix/store/7dpxg7ki7g8ynkdwcqf493p2x8divb4i-bash-5.2-p15/bin:/nix/store/mbh9gx43gsc5av019vsdkazbyiic3f0f-patch-2.7.6/bin:/nix/store/3q6fnwcm677l1q60vkhcf9m1gxhv83jm-xz-5.4.4-bin/bin:/nix/store/a5s2dmba16pw8l52f496gw0b7nli7mbn-file-5.45/bin ++ mkdir -p /etc/pam.d ++ [[ ! -f /etc/passwd ]] ++ echo root:x:0:0::/root:/nix/store/7dpxg7ki7g8ynkdwcqf493p2x8divb4i-bash-5.2-p15/bin/bash ++ echo 'root:!x:::::::' ++ [[ ! -f /etc/group ]] ++ [[ ! -f /etc/pam.d/other ]] ++ cat ++ [[ ! -f ]] ++ echo nixbld:x:999: ++ echo root:x:0: +++ seq 1 10 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 1' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld1 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 2' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld2 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 3' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld3 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 4' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld4 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 5' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld5 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 6' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld6 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 7' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld7 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 8' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld8 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 9' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld9 ++ for n in $(seq 1 10) +++ command -v nologin ++ useradd -c 'Nix build user 10' -d /var/empty -g nixbld -G nixbld -M -N -r -s /nix/store/8d4v7wcr31xh2zh6kcyxnd94r0n8byy6-shadow-4.14.2/bin/nologin nixbld10 ++ mkdir -p /home ++ useradd -c 'ci user' --home-dir /home/ci --uid 1001 --groups nixbld,root --create-home ci Creating mailbox file: No such file or directory ++ cat /etc/group nixbld:x:999:nixbld1,nixbld2,nixbld3,nixbld4,nixbld5,nixbld6,nixbld7,nixbld8,nixbld9,nixbld10,ci root:x:0:ci ci:!:1001: ++ chown -R 1001:root /src chown: changing ownership of '/src': Invalid argument exitHandler exitCode=1 set +e '[' -n '' ']' (( 1 != 0 )) runHook failureHook local hookName=failureHook shift local 'hooksSlice=failureHooks[ local hook for hook in \"callImplicitHook 0 {!hooksSlice+\"${!hooksSlice}\"} eval 'callImplicitHook 0 failureHook' declare -F 'callImplicitHook 0 failureHook' eval 'callImplicitHook 0 failureHook' ++ callImplicitHook 0 failureHook ++ local def=0 ++ local hookName=failureHook ++ declare -F failureHook ++ type -p failureHook ++ '[' -n '' ']' ++ return 0 return 0 '[' -n '' ']' return 1 </pre</detailsIf I remove the chowns, everything in the container is owned by root, as you'd expect, but this leads to various issues in our CI process with bind-mounted directories. Is there no way to change the ownership of directories during the container build process?\nLooks like I misunderstood how fakeroot works. It records those permissions between different programs and can replace them for the tar program:\nJust adding the state of the directories when running the above without the chowns in 23.11. Note that the CI user doesn't even wind up owning its home directory created by . The user does exist, and I can things interactively in this bash shell as root. However, we don't run the container as root during the CI build for a few reasons, including that it needs to run postgres, which won't run as root.\nThis might be a better fix than: Feel free to open a pull request if it fixes your use case.\nI'll pull down nixpkgs and give this a shot, thanks!\nThat does indeed seem to fix it! I'll open an MR now\nCreated , still cloning stuff to run the review suite locally"}
{"_id":"q-en-nixpkgs-bad9210657d79fdc8308cbac4166169e7fadb70be6893796d6c52ec05e32bd51","text":"Steps to reproduce the behavior: stig I got this when switching from 23.05 to 23.11 system: host os: multi-user?: sandbox: version: channels(root): channels(ao): nixpkgs:\nThanks for the report, this is supposed to be fixed in . I will wait for Darwin CI to be green there and merge it. Ping me if I forget. Follow that PR at:\nGreat, thanks!"}
{"_id":"q-en-nixpkgs-e787eb33384c2434433943421bbe8a511a84113e77ac059a874265a75131e7ef","text":"Steps to reproduce the behavior: cryptowatch-desktop download url redirects to a blog Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAs of September the 30th 2023, Cryptowatch was sunset and users were advised to use their parent companies offering Kraken Pro. I'll create a pull request to remove cryptowatch from nixpkgs as there aren't any archives available that contain it that are reliable"}
{"_id":"q-en-nixpkgs-2cfe31d835656c765cb7c8ae89816814b7e73408b176af18e62649f485a7ef6d","text":"While working on some documentation on Nixpkgs, I noticed that there doesn't seem to be a convention for showing shell commands and their outputs. prepends to the command prompt, and output as a normal, not-indented line. adds a comment above the command prompt showing what the output is. uses two different code blocks, one with the command, the other with the output. Additionally, there are other code blocks to show expressions and their result inside the nix repl: prepends to the (error) output. And prepends to the output. Note that the proposal below was written just to have some starting proposal, but not much thought has been given to other alternatives yet. My current proposal is to adopt the following convention: For shell commands and outputs, prepend to any lines that are part of the command, and put outputs in their own lines as-is. If the command has line continuations, don't prepend , but add some indentation to keep it from looking like an output. For example: For nix repl expressions, prepend to any lines part of an expression, and put outputs in their own lines as-is. <!-- make sure this issue is not redundant or obsolete --- [x] checked [latest Nixpkgs manual] ([source][nixpkgs-source]) and [latest NixOS manual] ([source][nixos-source]) [x] checked [open documentation issues] for possible duplicates [x] checked [open documentation pull requests] for possible solutions [latest Nixpkgs manual]: [latest NixOS manual]: [nixpkgs-source]: [nixos-source]: [open documentation issues]: [open documentation pull requests]: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nTalked about this in the docs team meeting, I think this is a good idea! If you make a PR to describe this convention in the I could review/merge it. I guess very generally, the convention is to use the format as you'd see it in the REPL.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-cdc0e86ddd518128bd2c5cc610c86792baf8564f7cfdaa6bf1c9c444e2413959","text":"Project description A tool to help efficiently focus windows in Sway inspired by i3-easyfocus: basically you press a key and hints (\"q\", \"w\", ...) popup on the titlebars, you can then press \"q\" to focus the window matching the \"q\" key. Metadata homepage URL: source URL: license: mit platforms: linux Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-236ddeed156aafa5625b3edd9ce9275fa3c7c73a2b1c788114e7a0bce57cb9ef","text":"DaVinci Resolve Studio is missing a entry for application launchers Steps to reproduce the behavior: should contain a entry for application launchers If applicable, add screenshots to help explain your problem. Add any other context about the problem here. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nSee"}
{"_id":"q-en-nixpkgs-f131079a74886ffd38906493420b8ce3413d905a9658e81a84fd93ed30518a4e","text":"New version of the invidious module gives an error on old system state versions with the following message: Config option doesn’t seem to exist and doesn’t do anything when set. Manually setting gets config to build, but service still doesn’t start with permission error, even after running the suggested reassign command: database seems to be correctly owned by the user: Steps to reproduce the behavior: invidious with local database on system state before 24.05: an error message, try to perform suggested fix. Find out that the suggested setting doesn’t exist and use instead: as suggested by the error message. to start . permission error. Service should successfully start after applying suggested fixes. Error message shouldn’t refer to non-existent options. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThe correct configuration should look like this if your state version is before 24.05: Looks like there is a mistake in the error message ( and are both specified and only the former is correct). Can you please try again with set to ?\nYes, I’ve tried this (steps 2+ of reproduction). I’ve mistakenly wrote in my initial issue, sorry, I’ve tried setting . It gets config to build, but it doesn’t start with the error described in the first post. I’ll amend the issue text now.\nCan you show us the output of\nSure! I see the problem now. I’ve tried to run the command without selecting the database first. After running it with database selected, service starts successfully. Thanks!\nSuper. Can you please look at the PR and tell me if the improved language would have solved your issue?\nYeah, thanks, that would’ve been clear!"}
{"_id":"q-en-nixpkgs-493e6280e7be31b15760a774e3cafbbf960697ad84dc6d316c426d053a744205","text":"labwc-tweaks failes to start no matter what I do I get this error message if I try to open it in terminal: Error opening file: No such file or directory Run labwc-tweaks in the terminal labwc-tweaks starting Please run and paste the result. these 4 paths will be fetched (0.03 MiB download, 0.17 MiB unpacked): /nix/store/1alqjnr40dsk7cl15l5sn5y2zdxidc1v-binutils-wrapper-2.40 /nix/store/4dr83a4wn25zkb5pvsq2y3q3v586jzs9-expand-response-params /nix/store/90h6k8ylkgn81k10190v5c9ldyjpzgl9-gcc-wrapper-12.3.0 /nix/store/wr08yanv2bjrphhi5aai12hf2qz5kvic-stdenv-linux copying path '/nix/store/4dr83a4wn25zkb5pvsq2y3q3v586jzs9-expand-response-params' from ''... copying path '/nix/store/1alqjnr40dsk7cl15l5sn5y2zdxidc1v-binutils-wrapper-2.40' from ''... copying path '/nix/store/90h6k8ylkgn81k10190v5c9ldyjpzgl9-gcc-wrapper-12.3.0' from ''... copying path '/nix/store/wr08yanv2bjrphhi5aai12hf2qz5kvic-stdenv-linux' from ''... system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThe issue might be related to this hard coded path: it seems like that xkeyboard_config provides this file\nUntil recently the keyboard related files were available in . Fixed in"}
{"_id":"q-en-nixpkgs-8e3dc42ded2dd9d0787507856aaf8498397fe02b5caf7f29e6a193e62802953b","text":"Project description Command line program which provides a suite of tools to create shopping lists and maintain recipes. Metadata homepage URL: source URL: license: mit platforms: linux, darwin Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nA nix-shell script here () which provides a decent starting point. cc"}
{"_id":"q-en-nixpkgs-79e3718ee17391f9e025a0d5d0b95cc559bac049a814e0ae5218e2bc8bc2d6ea","text":"Package name: solana-cli Latest released version: 1.16.27 <!-- Search your package here: --- Current version on the unstable channel: 1.14.23 Current version on the stable/release channel: 1.14.23 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHello I was able make the building script work so solana-cli is updated to 1.16.27 expect solana has bug in a which doesn't let me compile. The issue was already reported back back in 22 november 2023: solana-labs/solana So until is fixed I can't do anything more, as soon as it is fixed I will post my pr\nThanks , did you manage to build it for the current version? () On the master of nixpkgs I couldn't even build the existing version of Solana due a rocksdb issue I also found this commit on homebrew PRs. Seems like they patched this part in order to make the build work:\nHello I didn't test it and I don't think I need since I made it able to compile 1.16.27 I appreciate the citation, with the patch used by them I was able to fix it, appreciate your help"}
{"_id":"q-en-nixpkgs-9f2264431e3c40bb9ef4a6f4e2776817beb743764fa592ed9f4d530cdb730a6a","text":"Optipng crashes with a segmentation fault on current as of . points to the library used to throw/catch exceptions. relevant source: Automatic points to being the first bad revision. However, I don't have enough knowledge of this piece of nix to understand why it causes optipng to fail in the . System: nixos-master Nix version: 1.11.11 Nixpkgs version: Sandboxing enabled: yes CC\nReading into the causing commit's comments made me try which seems to work around the issue. However, I'm hesitant to commit this as a fix without understanding the issue here.\nSo some packages use the C compiler instead of linker directly to link. Nixpkgs's cc-wrapper is aware of this, for example, with separate and environment variables. Classic make builtin rules however only have , , , and , so then either in a configure script or make itself they seem to default to , but manually specifying breaks this. Manually specifying , however, avoids that pitfall, and, unlike , will also work for cross compilation.\nHere's a blurb from optipng's script: If I understand your comment, it looks like optipng's build system is doing precisely what you describe: So when isn't set, it defaults to configuring the such that it uses as the linker, and I suppose the theory is that they depend on behavior available when invoking through e.g. (though I don't know precisely what that may be). It would seem reasonable take the (on a per package basis) approach as a solution. I think this can safely be considered a bug in the upstream packages' build systems.\nProbably the use of to pass actual ld flags, e.g. .\nI implemented the workaround in . Feel free to reopen this if you don't think the fix to be adequate."}
{"_id":"q-en-nixpkgs-ae8a323ae59400f97895b4ab79f3cf18e13fa5fa1e3e546a53240862262ebe4c","text":"I would absolutely love for somebody to package up :)\nwill attempt to reserve some time to give it a try in the next days :-)\nI created a first working version which compiles on my machine ( ): Unfortunately I had to use an older revision since the latest revisions require language features from Rust 1.19 and I'll have to bump it in nixpkgs before using the latest revision, after that I'll open a PR :-)\nbtw do we want to add an option which allows to change the coreutils pkg from to ? Something like and every module should start start using .\nI had contemplated using a package override and just make equal to . Not sure how well that would work out... This may be something where is one of the few qualified to chime in.\nPlease don't introduce this as a NixOS option. This rewrite is not compatible with the GNU coreutils. This will certainly break NixOS. The package override will break lots of packages and will trigger a full rebuild anyway. Packaging these for testing is fine. :+1:\nCould you just wrap this into a PR then? I will certainly test this and create some aliases to replace the GNU tools (in the user environment)\nsorry for the delay at first. I'm aware of this. I just proposed an approach to switch the package just for system-wide testing purposes. This is actually the better approach. Furthermore this is something anyone can do locally, so shouldn't have concers with it as it shouldn't affect code in for now :-) I'll certainly do. However we have to wait for since the latest revisions of require Rust 1.19 (I'm currenty using an older revision from July 2017 locally). I checked out this change tried to compile uutils-coreutils against the new Rust compiler, but it cargo builds seem to fail ()\nAh, of course. Let's wait for then.\nafter has been merged it should be possible to finally implement this. Will open a PR in th next days..."}
{"_id":"q-en-nixpkgs-432d9a6a8fa91901bd87894ec34922a7991dec1aeec18024cb81b75efad3b5aa","text":"Steps to reproduce the behavior: curaengine This was on an update yesterday to . Please run and paste the result.\nI opened a PR yesterday:"}
{"_id":"q-en-nixpkgs-1b080775c65f38495dc67e3d886bc17c1ee1462ca1a407eb36bab84835ca3a6e","text":"Steps to reproduce the behavior: with latest master branch is building with the older version of (2.5), because it can't build with the latest version (see: ). Build of started to fail after recent update to version . Notify Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-a516ac45d94169140caea918d976452b87ae203994b0e91c5c622edf3a196752","text":"Using more-or-less the default configuration for (with ), the list of results is printed twice whenever running . For example: Steps to reproduce the behavior: the module like so: the job with some file name that exists on the system Matching files should each be printed once. The issue is fixed if I unset the environment variable. plocate deliberately reads from both and the hard coded default path, even if they are the same: The comment claims that this is also the behavior of mlocate, but I do not have an mlocate database installed and haven't wanted to change my running system to use mlocate. It looks like recently started being set in NixOS in . Please run and paste the result.\nSounds to me like an upstream issue. Can you write a mail to the upstream maintainer? ;a=blob;f=README;h=;hb=HEAD#l40\nSure, I have done so. I'll follow up when I get a response.\nThe plocate maintainer says they do not consider this to be a bug in plocate, and is intentional behavior. Rather than setting the environment variable, could we just make a wrapper script for that uses the NixOS-configured dbpath, if we don't want to re-build itself when the path is changed?\nI would be personally fine to just drop support for a custom place for the db. If people really want to have it somewhere else they could symlink it. Then we don't need a wrapper or to worry about the env.\nI can confirm the behaviour with .\nCould we just take out the variable setting? It isn't needed for or , and just causes this annoying behaviour. If it's needed for , maybe the setting could be made conditional. It would be nice to get something done about it sooner rather than later. This issue has been open for 4 months now.\nPlease take a look at\nI've reviewed the PR and it LGTM. Thanks for doing this! It will be great to have this fixed."}
{"_id":"q-en-nixpkgs-98ecf8718fde149a4b23293b4b9f909f65e15c5dcc52be3ce8b994f7fc8bff66","text":"A recent update broke building sysdig on release-17.03 branch. Nixpkgs version:\nDoes newer versions of sysdig solve the problem?\nFixed in\nThanks!"}
{"_id":"q-en-nixpkgs-6524d40093333ad8e8ac3299302908f53ef01f8cba356a0fb7adce71f3e3e203","text":"Package name: Tribler Latest released version: <!-- Search your package here: --- Current version on the unstable channel: 7.13.0 Current version on the stable/release channel: 7.11.0 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Note for maintainers: Please tag this issue in your PR. This may Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-29f739ce5f3e490b92773bf1061b3a9ea1f234622217cba331a6268cab1ca3bb","text":"Project description <!-- Describe the project a little: --Tools to manipulate pdf files. Metadata homepage URL: source URL: license: mit, bsd, gpl2+ , ... platforms: unix, linux, darwin, ... Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI'll see what I can do, should be good practice working with Gradle apps\nThank you Tom, much appreciated, best of luck.\nBy the way, I have pretty much succeeded in packaging it, and it seemed to run fine. Just forgot about it a little. The only issue with the package currently is that jar files are created non-deterministically by default (timestamps are not set to a fixed time). After using the fixup hook made for this situation (unzip and rezip the jar file) springboot doesn't recognise its files anymore, which is really weird. So it stayed non-deterministic. So I have two options: A) use the non-deterministic package or B) use the prebuilt jar from GitHub. The latter one is much less complex to work with.\nSee if this helps you.. or this discussion from the top (concerning java apps)...\nThanks for the pointers, though the second thing linked is my own post :) I finally managed to figure out how to fix the jar timestamps, so I'll open a PR adding the package\nThank you for your help . Much appreciated.. sorry for the delay. had a rough week last and it spilled into this one. Looking forward to using stirling-pdf on nixos."}
{"_id":"q-en-nixpkgs-4f3e9c8034d629b5f87418f39994c7eaaba61927383ac9a4c529c4f9cb85774a","text":"When creating a dev shell with + , the resulting environment cannot use CUDA with . Steps to reproduce the behavior: a Note that I am using the commit from nixos-23.11, which is the latest at this moment. However, I have tried the commit from a month ago and the problem was there as well. , run with the above flake. and Note that it reports is found, but does not seem to have CUDA enabled. The above can be found . It is expected that and run with CUDA and GPU appears in the device list. I am running this on my NixOS machine with a Nvidia GPU, where nvidia-driver is properly installed. I am also an intensive user and it works perfectly. Also I vaguely remember such combination works correctly a few months ago. I am not particularly sure about the maintainers so I might be wrong. Based on 's commit history and meta I am goingt to and Please run and paste the result.\nLet me know if I can provide more useful information. Any suggestion on how I can continue debugging this would also be very helpful. Thanks in advance!\nhmm interesting... ooc what gpu are you trying to use? have you confirmed that the cuda capabilities specified in your flake are compatible with your gpu?\nAlso please gist the outputs of demonstrating which implementation of jaxlib ends up being loaded, and whether a search for is attempted\nOn this machine it is a 3060 Ti. I have been using on this machine, with the samilar dev Shell, so I can confirm the cuda capabilities specified should work.\nThanks. Let me put the stderr log here: The log is very huge. I tried to search for and it did not come back with anything. I am trying to see how is loaded from it.\nI think it is trying to load from , but I am not sure whether it is the GPU-enabled one or not. I tend to think the above is the GPU-enabled because if I specify in the dev shell, it loads instead.\nAlso just to confirm: what's the output of ? Sometimes multiple jaxlib versions can get pulled in as transitive dependencies (though they shouldn't if everything is packaged correctly). Then python can end up loading jaxlib-without-cuda instead of jaxlib-with-cuda depending on the ordering in PYTHONPATH. From the looks of your flake this shouldn't be the case, but I'm guessing that this is a reproduction distilled from a more complex flake/project?\nYes this is distilled from a real world flake. It seems that in my dev shell is not set. However, running produces: I looked into the of the : Apparently there is only one being pulled in here.\nOk yeah that looks fine.. hmm\nBtw, I tried two more things the capabilities constraint - still the same this on a 3090 machine, a 3080 machine and a 4090 machine. All of them produce exactly the same behavior.\nAttempting to repro this but currently blocked on\nInteresting. I had an attempt to run my nix-based ML project on an Ubuntu machine but wasn't able to get the nix dev shell to recognize the GPU a while ago. Have been using NixOS for development since then ...\nOh yeah it definitely works... mostly. This is the first real issue I've had with nixGL haha\nWill give it a try! Meanwhile, I can try out any thoughts/ideas you might have to debug. Thanks!\nYeah my hope upon getting a repro is to bisect nixpkgs commits to find a known-good commit and then hopefully isolate the issue in a commit range. If you're feeling up to the task it would be quite useful! This would be a great use case for GPU-enabled tests and some CI infrastructure. tl;dr is that build has been green for a while but at some point it stopped actually using/recognizing GPUs.\nGot it. This looks like something I can try. It would probably take a while as I'd imagine CUDA will be built again and again.\nLet me try to build from a few commits on to start with.\nTried from about 3 months ago. Got different behavior: An older version has similar issue:\nAlso confirmed that can reproduce the probem: Similarly, from 2023-12-01 can reproduce the issue:\nThe above seems to suggest that, it is likely that in the past 3 months would not work with CUDA.\nI found the a successful commit on : Please note that this is not a commit from and I believe it is using CUDA 11.7\nNice! Well I guess we have a (rough) commit window now to bisect on. Btw, I'm curious: does work any better for you on master? We don't have a CUDA alias for that one, so you'll need to set\nLet me try that. My understanding is that is about downloading the pre-built binary (wheel) and patch it, while is about spinning up to build the wheel by ourselves And for end users of , they should behave pretty much the same. Is my understanding correct? Thanks!\nTried . Seeing the following error:\nDigged a little bit into this. On the successful branch 23.05 with CUDA 11.7: while on the failed branch 23.11, This seems to suggest that the we have in this case is not enabled for cuda support.\nYes, that's exactly right! Nice find! Huh, very weird. We're building with all the GPU/CUDA flags turned on AFAIK. Also, it's surprising to me that jaxlib-bin isn't working on your system since upstream builds that and we make only minor modifications\nI checked their and found that there is a bazel flag called . Don't know what it is (maybe related to ?), but 's does not have it. I tried to add it and (more specifically, the inside it) cannot build any more. Sorry if I did not make it clear. can be successfully built on my system. The problem is similar to , as it cannot be used with CUDA. It just prompts a different error.\nMore on . Although it can successfully create the \"client\", will report not being able to find cuda drivers: This is on 23.11. The on has been upgraded to , which has the same problem of not being able to find CUDA:\nTracing the code and I realized that the problem is that this line , which in turn calls in failed for . Based on , I think the <versionhere should be . And it turns out that we already have for , which contains . Scratching my head again now ...\nYup, totally with you there. Now if jaxlib-bin didn't build successfully that'd be an easier fix... Debugging with jaxlib-bin can be convenient since it offers \"known-good\" binaries to work with. Hmm, ok this is where output ought to point the way. But looking through the logs you posted earlier it, I'm not able to find the string anywhere... Would you mind posting the logs for the version? OOC how did you come to conclude that <versionis the failing lookup? It's a perfectly sensible conclusion since just about every CUDA package ever has to load it, but just curious re your methodology.\nAh sorry I should have provided more information here. So my journey starts with the error: and this quickly leads to the code in tensorflow: This means that the call to failed. This particular function can be found at , which is essentially . The function is , which is basically a call to , which in turn is just a call to of the here. Reading the above functions, we can see that the offending filename is in the format of . I could not figure out the because it is actually defined , which is in turn injected from . Because I had the impression that I was using , so I believe the version here is . This turned out to be wrong, see below. Thanks a lot for the suggestion! I tried the following Which produces log: Turns out that it is actually looking for . However, it is looking for ! Below are the selected part of the above log: I then went back to check for the : , and apparently it is downloading the prebuilt binary that is tagged with . Note that points to Cuda 11.8. I might be wrong, but is prebuilt binary tagged with supposed to work with CUDA 11 as well? The logic that finds and loads the files definitely is not compatible with CUDA 11.\nFor as long as tf is 2.13: We should add the comment. It's possible that jaxlib and tf actually come with different version constraints, I haven't checked. I guess one then has to pass it ( specifically) . Looks like jaxlib is trying to take over a part of the dynamic linker's job instead of just using DT_NEEDED and naturally this doesn't play out smoothly. They probably wanted to defer loading or something\nI am trying adding an overlay like this: Seems that is now working. I am not sure whether this will have bad implications though. Yeah they might have some reason to defer it.\nThere is one problem though, if I have the following overlay: is ok, but will fail to build because of test failures. See the log: Note that all the failures seem to be related to: I suspect this is jax tests that should run on CUDA but incorrectly ran on cpu - but I am not sure.\nAt this moment, with the help of and (Thanks!), I think I might have a solution to this. TLTR: We can now use (version 0.4.23) on , if you apply an overlay to point to . NOTE: usually you will also need alongside . It works on , but you will have to disable the unit tests and import test. See below in \"what is still unclear\". Detailed explanation: download a prebuilt binary, which defers loading of at runtime via dynamically. prebuilt binary is built against , meaning that will look for . , there is another check to make sure that the CUDA version is exactly (i.e. 12.1 and 12.0 will not work). , uses , which is designed for . The current needs instead of . However, if your environment does not require , just point to and everything should work out. with fixed above, I cannot build unless all the tests are disabled. It does not seem to affect the actual usage of because the test failure is specifically about not finding GPU-version of during test. This should be fixable but I am not sure how at this moment. we want to be able to use instead of to have more flexibility. The current build of seems to produce a library that does not have CUDA support. This remains a mystery and needs further investigation.\nI propose that we template the \"cudaXX\" strings in based on the version of the inputted / instead of hardcoding. That will prevent this kind of issue in the future. It's worth noting that upstream releases builds for both cuda11 and cuda12, so we're in luck there. Unfortunately I'm of minimal use in terms of submitting PRs until is resolved. This reminds me that we should also migrate jaxlib-bin to use the redist packages at some point. Fixing the source build will be its own adventure. I suspect we're just missing some kind of flag that changed in upstream's build process.\nSounds good. Still, for the upstream releases, cuda11 = cuda 11.8, cuda12 = cuda 12.2. The minor version is still important here and should be checked. Agree. It probably requires setting a development environment and build incrementally with to investigate. Blindly update flag and build the whole package again from scratch is quite inefficient ...\nSorry but I was not able to find the build for . I saw the ones under , but not under . Are they the same?\nSearch for \"cuda11/jaxlib-0.4.23\" and you'll find them. This link seems to be outdated, or at least not actively maintained by upstream. We should prob update the comment to point to instead.\nI saw you're using that with Ubuntu? Did you try nixglhost?\noooh no i hadn't heard of nixglhost before... i'll give that a shot!\nWith that I was able to find the one for CUDA 11 (cudnn 8.6) and it worked out with this overlay: Need some advice on how to fix on . Based on suggestion, I am thinking about the version of cuda, assert on 11.8 / 12.2 will be set to if cuda is 11.8, or if cuda is 12.2 the file based on the above version Does that make sense?\nYup, exactly what I was thinking! Though I have a sneaking suspicion that CUDA versions outside of 11.8 / 12.2 might work. Sometimes they'll attempt to and then try , and so forth. So we should try seeing if other CUDA versions work too.\nThanks! I tried 12.0 and it won't work for prebuilt binary originally built against 12.2.\nAh ok maybe not then... damn EDIT: the JAX folks claim that this ought to be possible:\nI still don't understand (and haven't got around to verify myself) whether is functional, and why do we keep focusing on instead?\nis not functional, which is the original report of this issue: can be built However, the built (more specifically, ), was not able to work with CUDA. It seems that the problem is is actually not \"withCuda\". There might be some flags to be set during bazel build, but I was not able to find the reason. We then turn to as an alternative because we need a working at this moment. Turns out that has a different issue, but is easier to fix.\nI also read about that, but my experiment wasn't successful, for example, it will panic about JAX being built with but the current CUDA version is\nLooking at for https://hercules- I see no mention of () Is it intentional we add to but not to the following? (again, I've no idea what I'm talking about)\nSimilar here, I am not that familiar with jax's bazel code. Would like to provide some information here. with CUDA 11.7 works, and at that time looks pretty much the same:\nIt seems that the version of cudnn is not a requirement. Binary built with 8.6 can work with 8.9 at runtime.\nWell, there's definitely a bug in here somewhere. It is def possible that this is part of it. I'm guessing that upstream changed their build flags and we completed a version upgrade without realizing since we don't have a means to run GPU-enabled tests in CI/nixpkgs-review yet. is included in as a means to reduce the number of possible hashes that we have to record. The fetch is a superset of the CPU-only fetch, so we combine them for simplicity. is really handy in these situations where the source build is either broken or not on a version that the user wants. The source build is really quite unwieldy unfortunately. (Similar but not quite as bad as the tensorflow situation.)\nI am working on the fix for . Not sure what is the best way to test it though. I am thinking about: the fix in a branch a that points to that branch with CUDA 11 and 12 on 3.9 through 3.12\nAnother approach, IMHO a little easier but totally up to you: Then you can run this for , , etc and and . That way you can run/test locally without committing changes.\nThe PR is created at\nI am closing this as at least the version is fixed now. Thanks a lot for the help\nFixing the *-bin version is great, but let's keep this open until the source version is fixed as well\nFWIW, the Guix version (there is only a source version) is here: It works fine with CUDA. The definition of the CUDA-less version that it references is here: Maybe there's an obvious difference somewhere in the arguments.\nOk I think I've found the culprit. Somewhere along the way we removed from and/or . Oops! Working on a fix now... Currently blocked on AFAICT we do not package CUB and it is not provided in any of the cudaPackages that I've consulted thus far. If anyone has any ideas on the best way to proceed, do let me know!\nPretty sure it's part of , and we also used to distribute it as part of nvidia-thrust which we deleted\nAh, thanks for the tip ! I wasn't able to find or : Perhaps it has a different attribute name? Assuming a typo, I checked nccl which is already a dependency of jaxlib, but to no avail: Perhaps I'm missing something? Btw, draft PR started here:\nOop, found , but not finding that header file sadly:\nNot \"\"?\nAdding seemed to solve the problem, despite the difference in filenames. Not sure exactly what kind of wizardry they're doing to adjust header names, but it seems happy for now.\nLooking at your issue upstream, I think the names match?\nOop, yeah you're right. I was confusing that with something else\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nIf you're using NixOS is there a platform wide config to always have the cuda libraries enabled?\nI'm trying to compile jax using the commit , which appears to have closed this issue. When I tried to run jax on CUDA using this commit, I would need to compile jaxlib from scratch. During the compilation, I get a hash mismatch error in a library that is compiling. I even tried to use a later commit to get the following output: This is my And this is my (for the newer commit):\nThis seems like a support question, you're more likely to find help on NixOS Discourse or Matrix. But the answer is and this will trigger a lot of rebuilds Interesting, this might be a reproducibility issue with the way we fetch bazel dependencies, in which case this deserves a separate issue"}
{"_id":"q-en-nixpkgs-ac1bfe1dadb5a4c3ef119b2020ce102d78c65c330ab871f8888faa5b25b556e8","text":"Can't start virtual device: qemu libpixmin not found. Steps to reproduce the behavior: Genymotion any Android device it An android VM is started Here is log: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-4c6533fcd438af2d1064cd410328e77e80932ca1c4cd5cece01e08ee14920a04","text":"Package name: Latest released version: 0.3.4 <!-- Search your package here: --- Current version on the unstable channel: 0.2.0 Current version on the stable/release channel: 0.2.0 Notify maintainers Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThe repo has moved to .\nWorking on a PR now.\nGetting errors with having Git dependencies. I need to find the hash of the new file, but Nix isn't giving it to me when I provide a fake hash.\nI can't get a hash, so I'm stopping for now. If anyone else wants to pick it up, it would be appreciated."}
{"_id":"q-en-nixpkgs-473c5d2fef0a556879a529591218999340542741c576b2e10b3959100fe24669","text":"There is a license violation in the package. The package contains the code for under the old MAME license which is non-commercial and incompatible with the GPL code in the rest of the program. I tried to notify upstream, but that didn't seem to go well...\nHello! It seems that upstream actually an option to disable in version of . , from my testing, gives me an error when I try to run it. You can find more info in this . So, i think merging the pull request linked to this issue is the best option right now."}
{"_id":"q-en-nixpkgs-f7f077816439da17386511771aba12996a58500d8db9a7b2e55e9e5dd4fabcb9","text":"Steps to reproduce the behavior: angband macOS Sonoma 14.3 Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\ncould you include entire build log?\nIt works now. THANKS DEVS!"}
{"_id":"q-en-nixpkgs-c095aa5351b88c8d86b828d89996caa6e5a221f51ae476d60b90cd3284e934a5","text":"Package name: gitqlient Latest released version: 1.6.2 Current version on the unstable channel: 1.5.0 Current version on the stable/release channel: 1.5.0 Notify maintainers Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-200dd8d2b69190eb19e812caca33bec02005d359d3fe288d8e6206e839c5a539","text":"Project description \"28Kb, small memory footprint, single binary that run list of commands in parallel and waits for their termination\" Metadata homepage URL: https://await- source URL: license: mit platforms: linux, darwin Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-9cd008f0b29d987e709ecba3015f44e3fed65aa4f8d96052343710c48e77ef7e","text":"fails to build on aarch64-darwin due to failing tests due to a sandbox issue. See also: Tests should pass, allowing the package build to succeed. none listed for pytest-django specifically, for django in general Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nIn a case like this, would disabling these tests be appropriate? They seem to only need access to -- is that permissible in a sandboxed build?\nFuture self: this was fixed by adding , which seems to currently have zero mentions in the nixpkgs manual (or the nixos manual) or Defined in nix: It adds to the sandbox-exec call."}
{"_id":"q-en-nixpkgs-1cc59313f36fc2aadd768293d18a0e39033f78f7b5db544a4e22bc00c4507377","text":"Project description Metadata homepage URL: source URL: license: bsd3 platforms: unix, linux, darwin Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-c0e6c61d33faf392c8db25c52ec062fc475182227d96fed3c0fb2dc648cb276d","text":"Steps to reproduce the behavior: the following : the flake: cc Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nOddly enough, (the original PR ) actually fails as well. Result of run on x86_64-linux <details<summary1 package failed to build:</summary<ul<licemu</li</ul</details"}
{"_id":"q-en-nixpkgs-6ba5b5b777df947cd2895fffeb49e9c73e334de864d072be458c5c26d73b3fc2","text":"Unable to launch gplates on NixOS. A shared object file appears to be missing. Steps to reproduce the behavior: : gplates is expected to launch and be usable. N/A N/A No listed maintainers on Nix Packages but was the last to touch the source for this application. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nProbably adding to and/or adding to might help, but I'm too tired to debug tonight. I'll look into it in the next few days.\nThank you, I was reading about the first option you suggested being a potential fix. I'll see if I can wrap my head around doing it on my own as well this week, though I don't have a lot of experience contributing to projects likes this."}
{"_id":"q-en-nixpkgs-c30474595fa567ec6dac7921be3930d4442f95aa751807029875ee14c4cda4ce","text":"Hardening gives old Clang flags that it cannot handle and fails. A same problem, I imagine, would arise with old GCC versions. A possible solution is to list somewhere what hardening methods are supported and use the list when deciding what hardening to apply. For example: a file : an error. Nixpkgs version: (run ) Sandboxing enabled: (run ) yes"}
{"_id":"q-en-nixpkgs-f7b7435c99150c647111e71eb9191955222d4e93d4a205ce5107f169d2571783","text":"In the NixOS gitlab module, the fleet of Gitlab component services is strongly coupled to the (local) postgresql and – via sidekiq – to Redis. On a systemd unit dependency level, this happens via a relationship and causes the other services to be stopped and (re)started when just one of them is instructed to do so. Unfortunately, there can be several edge cases where the system ends up with not all of the components being in the desired state (mostly up). Steps to reproduce the behavior: This issue is a subtle race condition, so I choose to describe the general problematic setup instead of trying to engineer a reproducer: I've encountered a situation where both and where restarted (stop and start) at system switch time due to changed configuration. This triggered Gitlab to be stopped as well. Unfortunately the service(s) was in its startup (Activating) phase and stayed inactive. The expectation is that Gitlab comes back after the dependency restart. Instead of continuing to tune the strongly-tied, evermore complex inter-service dependencies I propose to decouple postgresql and redis from the rest of the gitlab services. I'll open a WIP pr on this. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHere's my proposal on reducing the coupling between DBs and the rest of gitlab, currently with failing tests unfortunately:"}
{"_id":"q-en-nixpkgs-7fc9156788f23e75cad0f6640ef1888aab7277cf1f15d6f6b9b71fb2c9d25e09","text":"Steps to reproduce the behavior: /run/current-system/bin/switch-to-configuration test Does not crash system/systemd This can be reproduced using this test in sops-nix: Edit to have this change <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nRelated to\nIt seems like it's reusing the same loop device when mounting the new overlay..."}
{"_id":"q-en-nixpkgs-9dd464a877bf1c641edd07d32b819602f9bb84ab2d3b67b960f51b3615135c30","text":"Package name: Latest released version: 2.27.0 <!-- Search your package here: --- Current version on the unstable channel: 2.21.0 Current version on the stable/release channel: 2.21.0 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-1322723339473795f41572f643a2b6ebc4efc613abe028445010c6e2171da491","text":"Mentioned in , the bcc package can also be used as from python. This currently doesn't seem to work? If applicable, add screenshots to help explain your problem. Add any other context about the problem here. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nbcc doesn't depend on kernel anymore, so the second half of the comment you linked at isn't required and we can just move the package & add a compat alias. Technically you should already be able to use bcc as by magic'ing the correct site path, but I agree it's better to move it - patch welcome!"}
{"_id":"q-en-nixpkgs-fe6ee66fca6f8bd617ee7fe4140577eca34c69c177ddc4feeaaf406187703972","text":"Project description GitButler is a git client that lets you work on multiple branches at the same time. It allows you to quickly organize file changes into separate branches while still having them applied to your working directory. You can then push branches individually to your remote, or directly create pull requests. In a nutshell, it's a more flexible version of git add -p and git rebase -i, allowing you to efficiently multitask across branches. Metadata homepage URL: source URL: license: FSL-1.0-MIT platforms: linux, darwin\nis also a Tauri app so the packaging is likely similar:\nI'll work on this over the weekend, unless anyone has already started?\nI don't think building from source is gonna be easy, precisely as we don't have done yet and this project uses pnpm workspaces. You can however wrap the Appimage though.\nI've made Still draft. It does start the application, but clicking buttons results in problems on my system :sweatsmile: For a quick run: Update: it's now mostly working. The signin/login still gives an error and it's still unclear why. Also, I don't think I'll use gitbutler regularly atm, so it might be best if someone else wants to volunteer as maintainer of this package :sweatsmile:\nI have made a PR that builds from source and has no issues with logging in (that just needed ).\nIs this request open to contributions?\nWould be best to wait for a while, if that PR goes abandoned, you can ask the author if it's okay to open another"}
{"_id":"q-en-nixpkgs-25cc7c6c9aa8b9f7015e6a4af4bc2b2a66d129bc069a8394c2243903481f9faa","text":"libstudxml is a streaming XML pull parser and streaming XML serializer implementation for modern, standard C++. It has an API that we believe should have already been in Boost or even in the C++ standard library."}
{"_id":"q-en-nixpkgs-1bd8261ae889b3312cf3098bdf49919159c7be4f5452ffebafd04113b470f427","text":"Package name: curl-impersonate Latest released version: 0.6.0 <!-- Search your package here: --- Current version on the unstable channel: 0.5.4 Current version on the stable/release channel: 0.5.4 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-0cd715d611fea20ffeaa8d5dad41cf9a7b9b73965b501bf08eab52d4b623659a","text":"Package name: Flutter Latest released version: 3.19.0 <!-- Search your package here: --- Current version on the unstable channel: 3.16.7 Current version on the stable/release channel: 3.13.8 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-e4cb276bec86d6950170118e9cc2ed62f2b66fc2cb17f330f942b3f4e908d99d","text":"When using incus with openvswitch, incus needs to have openvswitch in its path for features involving openvswitch to work. As a result, when attaching an openvswitch as a bridged interface in incus, we encounter issues such as veth interface remaining active when it should be cleaned up, not being able to use openvswitch to make bridges, etc. It is not so urgent because it's easy enough to add it to the path, but it caused me some trouble until I figured out the cause. Enable openvswitch, enable incus, run \"incus network create BRIDGENAME —type bridge \" and notice that incus does not detect openvswitch. If openvswitch is enabled, openvswitch will be to incus's path, and the above command will not fail. Can be resolved by adding = [ pkgs.openvswitch ]; Maybe just need to check if openvswitch enabled, and if so, add it to the path by default? <!-- --Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nDo you have a vswitch configuration you're able to share?\nYes, but if I configure it using nixos, I get other problems (incus is deleting the whole bridge whenever I shut down a container and I don’t know what part of nixos is responsible for it), so I have switched to configuring it with networkmanager, where it all works well, as long as I add openvswitch to path. But here is a very simple example with a nixos configuration: = true; networking.vswitches = { br0 = { interfaces = { eth0 = { }; }; }; }; networking.interfaces.vmbr0.useDHCP = true; Just adjust eth0 to whatever device you want to add to the bridge. Then in incus, we can add a profile. Easy way to do this is set up the default profile. Then we copy it to a new one: incus copy profile default ovstest and edit it incus profile edit ovstest Then replace the networking eth0 part with the below: devices: eth0: name: eth0 nictype: bridged parent: br0 type: nic Then make a container on the profile: incus launch images:debian/bookworm mycontainer —profile ovstest It should work fine, but (this is a bug I think I haven’t reported because it could be my own stupidity) when you stop the container incus stop mycontainer All of your configured bridges will be missing, completely. So the way I do it now is enabling network manager. We can add eth0 to a network manager-defined ovs bridge this way: Setting it up like this, and adding the line below, everything is working as expected. (If I don’t add the path = line, the virtual ethernet devices don’t get cleaned up when incus shuts down a container, which is why I filed the bug report to try to have it included default) = { path = [ pkgs.openvswitch ]; Hopefully this answers what you are looking for.\nThanks! Sounds like you have a workaround until this can get fixed in the module. I'm not quite sure why your bridge is being removed when the container is stopped. You could try creating a bridge instead, but then it looks like you should be setting I tried to test this, but creating an ovs bridge yields:\nI think it is because that ovs bridge doesn't exist yet. If you just attach a pre-existing ovs bridge to the container, it will work. Regular linux bridges defined with the nixos configuration also have the same disappearing problem, but with networkmanager work fine.\nActually the error I encountered is a bug.\nI'm not able to reproduce your issue where the stopping of the container removes a bridge from the host. Are you on 23.11 or unstable?\nThis is on 23.11.\nIf you are able, can you try and reproduce the bridge removal using nixos unstable?\nSure, I will give it a go.\nSo, yes, I can reproduce it on unstable. Both with linux switches and with openvswitch switches. Here is all of the configuration related to networking. You can see the commented lines would be how I set up the linux bridges when testing them. = true; virtualisation.vswitch.resetOnStart = true; networking.useDHCP = false; networking.hostName = \"boss\"; = true; = \"eth0\"; networking.interfaces.vmbr0.ipv4.addresses = [{ address = \"192.168.0.2\"; prefixLength = 24; }]; = [{ address = \"0.0.0.0\"; prefixLength = 0; via = \"192.168.0.1\"; options = { metric = \"100\"; }; }]; = \"eth1\"; networking.interfaces.vmbr1.ipv4.addresses = [{ address = \"192.168.1.2\"; prefixLength = 24; }]; = [{ address = \"0.0.0.0\"; prefixLength = 0; via = \"192.168.1.1\"; options = { metric = \"101\"; }; }]; \"\" = true; = false; networking.nameservers = [\"1.1.1.1\" \"8.8.8.8”];\nI have created a pull request to resolve the path issues."}
{"_id":"q-en-nixpkgs-5428c5fdbb95648899f99a8e798fc2c41086c67fe68e9de441b7aeb4a1dd1353","text":"Decent Sampler is supposed to have a standalone app, a VST2 plugin and a VST3 plugin. The package decent-sampler installs only the standalone app. The downloaded archive contains also which is the VST2 plugin and the directory which is the VST3 plugin. Steps to reproduce the behavior: decent-sampler. for the VST plugins 't find them. The VST3 plugin should be installed in of the package and the VST2 plugin too should be in the package, but I don't know where exactly. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-5d90c6fefd1984ab99526e2c3a00cb092886f658bcccca8b8803e5e5c0c29159","text":", it will fail when it can't find check hydra: works, but also generates deprecation warnings. 0.9.3 (as used by Hydra) isn't an official release at Github. 0.9.4 fixes the deprecation warnings (save for flagging a future dependency and one other future.) Bump to version 0.9.4 coming momentarily. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-12de21f67dba9fae1f640e63bcde19dac1c06965b8fbc9b845509f4dc05ea9de","text":"The searx service is trying to start and it fails because the actual name is of the executable has been changed to Steps to reproduce the behavior: ... use newest unstable branch ... try to deploy searx service with ... searx service fails to start searx service should start without any issues The default searx package on Nixos is in Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThanks for reporting, I just encountered this bug as well after I updated to the latest unstable branch. I believe that a general migration from to naming needs to be done to avoid these kind of issues in the future."}
{"_id":"q-en-nixpkgs-66a0df232b7d70189692e0d8e8638399c4f82b1feb4e9ac3b16eb24acb050652","text":"Some packages, e.g. requires nodejs16 (znck/grammarly), but I cannot override the nodejs version of the package to nodejs16 because it is EOF. Steps to reproduce the behavior: nodejs version of some nodejs package by doing the error message. Allow forcing the use of EOL packages, or maybe officially state that this is not supported and mark the package as broken. If applicable, add screenshots to help explain your problem. Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAllow forcing the use of EOL packages, or maybe officially state that this is not supported If you run: , you'll get all the instructions on how to do this. This has been marked as broken:\nSorry, I was not aware that I can get instructions when doing directly. Will do. For marking the package as broken, I mean instead of . Sorry for the bad wording."}
{"_id":"q-en-nixpkgs-0728b2be4311f3a774fa90abcb9764cf8c2d54daafb1f43027d8c37b830d171b","text":"the PR template points to instead of Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI kindly ask you to just create the PR the next time with this really trivial change instead of creating an issue.\nit's only trivial if you have experience with how github works. i had no idea the template was controlled by a file within the repo and not some option in the settings."}
{"_id":"q-en-nixpkgs-88f0d692a4604bfd114490f5e55f9090347811d322e053954752d9e88b251caf","text":"Project description <!-- Describe the project a little: --The hyprland cursor format, library and utilities. Metadata homepage URL: source URL: license: BSD3 platforms: unix, linux, darwin, ... Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-7c48d597ab0d6d9be969f7ee5d4dba8ef4ebf430bd5495754fd7b91e33000fe2","text":"Sessions are not restored on startup of Firefox 30.0. . firefox. the Preference. tab, select for . and . Ctrl-Q to quit the Firefox. firefox again. Firefox opens and Opens the start page. It works for Firefox 30.0 for Mac, 30.0 for Windows, or 29.0.1 for NixOS.\nI can confirm that. The good news is that your session is not lost. It looks like the UI has some issue, but the session is kept fine. So if you re-install version 29.0.1 you will have your tabs (including groups and pinned tabs) back.\nI have rollbacked to 29.0.1 and the session is restored."}
{"_id":"q-en-nixpkgs-cd5f83c46ee17f89f649a6c0dd644e59b4f67de474fb7ff9a7f4e0bf9042309b","text":"Since was merged into the master, swift cannot be built on aarch64. Steps to reproduce the behavior: reported in I'm not familiar with swift or clang, but I'm using the following overlay as a workaround until it's fixed. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nstates that this only fixed the swift build but all downstream packages are still broken. re-opening.\nActually, it seems only dark-mode-notify is broken, others seem fine. Probably because it uses Cocoa or disables sandbox. I don't know for sure.\nI noticed that my personal package does not build either. Maybe it's because of swift itself or swiftpm?\nThe above open PR seems to run into the same issue…"}
{"_id":"q-en-nixpkgs-05a65162450f28ca4c2514fcdb6a2a80b0788e83304d9f2fd03f371d5f55740f","text":"Project description A simple GTK4 based video player for mobile phones Metadata homepage URL: source URL: license: GPL3 platforms: Linux Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHi. Is this a package you're working on? If not, I packaged it but I'm not sure if it's correct. I can send a PR for it as starter. [It's the first time I built anything on NixOS].\nNot currently working on this, but can help test.\nI a PR for this. It would be great if you could check it out. Though running it gives this warning but can play videos. Also I couldn't play x265 videos with it. Is this normal?"}
{"_id":"q-en-nixpkgs-d974538f9fbdf3bd59aed5839a4a9d947aa8ddf521be608f97a6336d08154654","text":"When changing an SVP option during playback or the played file in mpv, starting from mpv 0.37.0, due to , it spams: and SVP no longer reports any connected player at the bottom of its window after that. top left logo, \"Open file via mpv\", pick a video the video is playing, switch back and forth between one of the buttons under \"Do frame rate conversion\" like \"To screen\" and \"Movie x2\" In mpv 0.36 and earlier, it says: and SVP keeps working normally. SVP has no source code available and new versions haven't been released on Linux since 2021, while mpv 0.37 came out in November 2023, making it unlikely to be fixed upstream any time soon. I've made a dirty workaround by to re-add the option, and overriding the source of the mpv package in my flake + home-manager setup. Handling both the external mpv package and the one within SVP's (the mpv that gets used by SVP's in-GUI \"Open file\") gets complicated - here are the relevant parts: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThanks for your investigation! Indeed it looks like modifying mpv to add the option back is the best way. Based on the diff on your fork, the change seems small enough for us to maintain. On the other hand, patching SVP is impossible due to its tamper protection.\nA couple questions: the mpv package gets updated, it might break building SVP with the patch. Do github checks catch these cases, or will the SVP package just be broken until someone else notices? I give you push access to the forked repo, to rebase and regenerate the patch as needed? I rarely update my system, so the repo will likely be left randomly outdated until I have to take care of it. you make the SVP package take and overridable parameters, so that it isn't so hard to e.g. add scripts to the internal mpv? to make a standard external mpv connect to SVP will fail since it won't be patched, is there a good way to indicate to users that they should install in their environment? The parameters mentioned above should probably also be made known.\npackage in nixpkgs will be broken. However I also have this SVP package on my NUR, and have daily CI runs to rebuild packages, so once it breaks I will notice and make a fix. isn't necessary. I can rebase the patch on mainline mpv as needed. will add them. can't think of any good way for now."}
{"_id":"q-en-nixpkgs-d2fbed9b6c51afe1fbbe6011c2fb63d1ca02e969bf05f841c7ccf1d9ba6e67a1","text":"Steps to reproduce the behavior: stig If I try to install it from my config, I get . If I try to run it by nix-shell: is the nixpkgs commit I'm currently using to get a working version of this program. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThe different experience you get with vs is a problem in your Nix environment - probably an old version of nixpkgs you use for vs how you build \"your config\". However the test failing mark this package is now completely broken. See also discussion with upstream at: . I opened that marks stig as broken, and it will close this issue when merged.\nstig works for me with this overlay: I think since addresses\nIt'd be great if you'd share this in a PR.\nOpened at !"}
{"_id":"q-en-nixpkgs-45a7d216a874bd01513e3d8ca6900c0ebaf85b255db6f219c6f39620ee0f71ed","text":"Cannot use hardware wallet with electrum from nixos-unstable channel Steps to reproduce the behavior: to use hardware wallet with electrum (either via creating wallet or loading a wallet) :x: on \"Choose Hardware Device\" screen, be presented with debug error A list of connected hardware wallets when at least one is connected ! Issue is not present in package from nixos-23.11 channel hardware wallet used: Ledger Nano X Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-03d8ffe9f7cc9df35e012c1b6a36b76cb9ab70b230ef22b6e736a7f77a42b96e","text":"Build of failed on x86_64-linux as of . This is currently breaking . maintainers cc: maintainers cc: Other cc: n/a nixpkgs at commit Contents of : This issue was automatically generated by . that created this issue. Internal tags: - -\nThanks and !"}
{"_id":"q-en-nixpkgs-50ca81da248a26e7e4e5d51b64686b19f0d1de2a9d75edb2803ae76c7869d7a2","text":"Steps to reproduce the behavior: .config.contentAddressedByDefault = true; .firefox.package = pkgs.wrapFirefox pkgs.firefox-devedition-unwrapped {}; to build I also have march/mtune set and can't/won't use firefox-bin Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI think this is $outPath of gcc\nI think my fix worked"}
{"_id":"q-en-nixpkgs-baf82a2394c6e634f8105c6ac9f95456277ff0e8e8a7f1399480763d04ac3e98","text":"Package name: koreader Latest released version: 2024.03.1 <!-- Search your package here: --- Current version on the unstable channel: 2023.04 Current version on the stable/release channel: 2023.04 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nplease test"}
{"_id":"q-en-nixpkgs-b602d5a9ee3226f0480b0929cbc4899eae268d60dcbd69d2a153c29729e001c3","text":"fails to find libcdparanoia and flac at runtime. This is a regression from , and appears to be more or less the same issue, except for Plasma 6. Steps to reproduce the behavior: to rip a CD with K3b It works, and lists Flac as a supported format. If applicable, add screenshots to help explain your problem. The fix from should still apply in some form? Manually adding flac to and libcdparanoia to works around the problem. also (maintainer for Plasma 5 K3b) Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nPlease test"}
{"_id":"q-en-nixpkgs-3ca452b105fe6802371885453794ce609514e013b3b046b612c1af5256985dd7","text":"Project description A waveform viewer with a focus on a snappy usable interface, and extensibility. <!-- Describe the project a little: --Metadata homepage URL: https://surfer- source URL: license: eupl12 platforms: linux Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI tried to package it myself however, the following derivation fails: Error: Similar issue: Looks like the real issue has been solved couple years ago. I should update the derivation accordingly."}
{"_id":"q-en-nixpkgs-d7cfcd1e54d77d7f39e29de284bbbfe54f3177bbaf1a0b6ba92898e6d6d09206","text":"Project description A .NET Core global tool to display and update outdated NuGet packages in a project Metadata homepage URL: source URL: license: mit platforms: unix, linux, darwin, ... Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-c7fe1a17a407206b0fb49c358ebb83a75e7b8c124d2df1f68c4b7018ecc6358c","text":"Package name: halloy Latest released version: 2024.5 <!-- Search your package here: --- Current version on the unstable channel: 2024.3 Current version on the stable/release channel: 2023.5 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-b92eca60f3a5d820ad748b15d767f80da1f05643adfac4c5817b21812f751915","text":"Package name: zombietrackergps Latest released version: 1.15 <!-- Search your package here: --- Current version on the unstable channel: 1.10 Current version on the stable/release channel: 1.10 Notify maintainers Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-a9446b827a22acfa531aa905f744d587bb5612b43c5568256ceb39fedff73575","text":"Steps to reproduce the behavior: Authy, stable or unstable channel No additional context, seems like the package of one of the dependencies was taken off snap for some reason Could not run for some reason... Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nIt seems that Twilio have deprecated the desktop apps. . Looking on it seems that they have pulled the snap from Snapcraft. It also seems that the Flatpak has been removed. See and .\nWell, i suppose there is nothing that can be done. Thanks for explaining the situation to me!\nI'm not sure that issue is really resolved, as I think we need to remove the package from nixpkgs as its deprecated. I believe you are the package maintainer. Should it maybe be removed?\nThanks for the tag and explanation, I had read the announcement when it came out and the apps were still available on Snapcraft. It must have been pulled after a while. It's better to remove the package now. I'll do it sometime this week."}
{"_id":"q-en-nixpkgs-e445ce3c9c03068431cf3ff3aff0c9cd5457a06236c459244e26e2ce8d6f1f56","text":"When installing latest Elixir with Erlang 20 via , several of these warnings are emitted through the build process: I'm installing through with (on Vultr FWIW). Though I believe on a new system may be enough to reproduce. The only way I've gotten the warnings to subside if by specifying . System: Nix version: Nixpkgs version: Sandboxing enabled:\n/cc: (by request from )\nI will look into this.\nI see that you have dropped the override. Is this because of these are just harmless warnings? I've switched to the same nixpkgs and could not reproduce the warnings on Ubuntu: Output is here: I'll try to reproduce this on NixOS machine now.\nI was able to reproduce this on NixOS machine:\nThanks"}
{"_id":"q-en-nixpkgs-e705b308d1650b64e9739c8de298aa94ef31564cac89ad2c4ae07c72d5d67d46","text":"This link is invalid: It is used to download the build Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-1051a81441881e15b553630b9f8c68bdc697aef75a627505caa13724acc1a3fa","text":"Project description <!-- Describe the project a little: --BlastEm is an open source, highly accurate emulator for the Genesis/Megadrive. Nightly release should be packaged as stable is five years out of date. Metadata homepage URL: source URL: license: GPLv3 platforms: unix Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-7e6283e3f7fe795bee22b6a23cc3d74fba3b553079d5b17c1be6d8d504f828ec","text":"After a recent update, xonsh is no longer able to launch. Launch Xonsh should start without issue If applicable, add screenshots to help explain your problem. I've found a change of python env that reached recently (). Perhaps it is the change that breaks xonsh. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --- system: host os: multi-user?: sandbox: version: nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nMay not be the same issue, but this doesn't work either:\nThat's a different issue. You should probably use instead. However, if you really need the full python env, unfortunately there's no way to use it outside the xonsh derivation for now. Maybe we should add it to ."}
{"_id":"q-en-nixpkgs-b12a2a927f385b4f8913af6785073d052a44b5c336ba214450870949bdd0f557","text":"On aarch64-darwin, while neovide does build and install successfully and I can launch it from the terminal, there is no app bundle that can be found in finder to launch neovide from there or pin it to the dock to launch from. Once opened the global menu calls it .neovide-unwrapped and the icon in the dock is a blank white page as a generic icon. I can only assume that once is fixed, that this will also be applicable to x86_64-darwin Steps to reproduce the behavior: neovide as either a system app in nix-darwin or a user app in home-manager on aarch64-darwin An app bundle name that will be linked into or depending on and usage I'd like to point out that the global menu title is .neovide-unwrapped to for sure indicate that this is the nix neovide and not the homebrew neovide ! ! <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nFollowing PR MacOS app information a few days ago: We can install the files with next release of neovide."}
{"_id":"q-en-nixpkgs-6136956a2a618df437e39b053873e4867433858530281508bb5d260eef13765e","text":"Run it. No error. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nWorkaround is to symlink your flake to /etc/nixos or to supply flake flag with path to your flake.\nSupplying the flake with a flag works for me. Did not try the symlink.\nCould you please be more specific? I don't understand what you mean.\nAre you using flakes for your configuration?\nYes. I think I got it, the command to run is:\nJust to be clear, how would you normally use here? I might have jumped the gun in changing the name of the issue, and maybe it's more like \"nixos-rebuild: better error messages when missing and flakes both\"\nThe point is that you don't need to search / know where the config is located if you're just listing generations\nI always have to use the argument. And yes I agree with this. The current title, \"nixos-rebuild: Don't search for config when using list-generations\" is better IMO.\nAs a workaround until the PR lands, you can run to suppress the error.\nThank you, this served well to confirm my assumptions. And yes, exactly what I had in mind as the root cause."}
{"_id":"q-en-nixpkgs-c018da5a283528e8abcb6a11cb29a2f432d314b5c998b2114207023f43b0da9a","text":"It looks like shadowsocks service integration is not working at the moment, even with minimal configuration. systemd service doesn't load. Apply this nixos config It looks like nixos will generate ExecStart similar like so Running this will just get a help screen of ss-server (and same for systemd service, confirmed with journalctl) Contents of /nix/store/i74pqxw9wwicclbamqjw2l4w7l48rdpz- is Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThis is caused by ss-libev's encryption method requiring a password but not set. Refined the assertion in fix pr."}
{"_id":"q-en-nixpkgs-2f037cc5e3c165e8a0bfd4b67abe0bfdab97ad3686a91d68a1bcaadecd18f18b","text":"Project description BrainFlow is a library intended to obtain, parse and analyze EEG, EMG, ECG and other kinds of data from biosensors. Metadata homepage URL: source URL: license: mit platforms: unix, linux, darwin Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-6b0441b85d2e0da84326e57387672261007eba5de36c86c0121f197be78334d9","text":"Steps to reproduce the behavior: Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nthere is now a captcha on the page at \"\" <img width=\"381\" alt=\"image\" src=\"\"\nLooks like it's the original website request captcha, I've submitted a PR to prompt it.\nThe captcha seems to have been removed."}
{"_id":"q-en-nixpkgs-8ff43a7750c478bf7da189c664361813ba772696f7f18297e675ca9fad182a02","text":"Steps to reproduce the behavior: python3.12-django-4.2.11 Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHere is what I see on my side so far : You should wait until nose support the python3.12 interpreter.\nThis doesn't mention explicitly, but the issue only exists in currently (PR ). On Hydra e.g. These tests now regressed on *-linux, both pythons, django_4 only.\nI was looking upstream but it may not be related to this issue then : still I'll mention it, apparently a fork for the nose lib is better maintained upstream and compatible with python3.12. There has been test failures reported as well on Python 3.11 upstream :\nFixed in"}
{"_id":"q-en-nixpkgs-ae7cffe4a9406351b076b6162d01c2655d6c161433fa2602fd4b9042669937b3","text":"Package name: rnnoise Latest released version: 0.2 <!-- Search your package here: --- Current version on the unstable channel: 2021-01-22 Current version on the stable/release channel: 2021-01-22 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers Note for maintainers: Please tag this issue in your PR. Tried building with but it got a missing file. There’s also warning about not using SSE3/AVX/AVX2, but Nix only supports the bottom of the barrel, common x86 instructions, but I usually compile this one out with some flags. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]: [^1] or offering a non-proprietary, non-US-corporate-controlled mirror for this free software project. I wish to delete this Microsoft account in the future, but I need more projects like this to support alternative methods to send patches & contribute.\nThe missing file is downloaded as part of the '' It looks like they a step to the build process.\nPerhaps it can be fetched using fetchurl/fetchpatch?\nYou're correct. The relevant file is retrieved in the script . It fetches a single file from the URL , where the version is extracted from the file , currently set to . Hence, we can emulate the download script and execute something similar to:\nAlternatively you could use instead of a separate model + . While we are at it, we should be using the main repository instead just Microsoft GitHub which is explicitly listed as a mirror, not the primary code forge. Something like Referencing both mirrors would be more resilient for when one of these platforms goes down.\nMade a PR , but still having trouble designing a ."}
{"_id":"q-en-nixpkgs-9a653c3d640dfc0b4da75c4443c391d735aebbcd276411492a0d961d63bd94c2","text":"When I try opening the new version of albert (), it crashes and reports that no frontend plugin is found. Steps to reproduce the behavior: albert to 0.23.0 albert Albert opens normally. This issue exists under Wayland and X11. : : Also notifying the author of the last commit Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI think I have relevant infos in my logs. I included the full log (including application restarts and the full stacktrace) as there are some unique messages on consecutive restarts. Note that I use systemd to run Albert. Beware wall of text :eyes: <details<summaryHardware</summary</details<details<summaryAlbert Details</summary</details<details<summaryLog</summary</details"}
{"_id":"q-en-nixpkgs-1bbed85c8f59190f733c2593c2ef1d96b1afb0a44c6f67da4c9f5d6e4fe86c84","text":"Package name: Latest released version: <!-- Search your package here: --- Current version on the unstable channel: Current version on the stable/release channel: <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\ndone in"}
{"_id":"q-en-nixpkgs-54e0a92749bb9a540bdcb933a3ed75b8a558a482fedb746b8bcdf683abb27df9","text":"Package name: calibre Latest released version: 7.9.0 Current version on the unstable channel: 7.8.0 Current version on the stable/release channel: 7.0.0 Notify maintainers https://calibre- Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-11f04ac68f392a4e654a74737d7272aebe9fd01d8e3ad51add3c50e5fb7e1666","text":"XOrg crashes when certain applications are launched. More details here: Steps to reproduce the behavior: any of the IntelliJ IDEs. You may have to set anti aliasing to greyscale. Xorg will crash. Xorg doesn't crash. This is the bug: It looks like a security fix caused this regression. Ubuntu has already applied patches that fix this for Xorg and Xwayland: see this message from the above Ubuntu thread: NixOS should use the same patches until this is fixed upstream in XOrg. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nSounds like this will be addressed by this update:\nxwayland too: looks like updating xorg and xwayland to their latest versions will do the trick then\nxwayland was bumped to the latest version in PR"}
{"_id":"q-en-nixpkgs-4b05af73330ecfb2f37032eb86d1e4e733a13af001d562e5795d7d251274cef4","text":"fails while compiling The same thing happens with Steps to reproduce the behavior: or: is successfully installed. Here is the output:\nFix in\nThanks On the same system that I encountered the error on, I checkout out the branch with your fix and then ran: The installation worked and is runnable. I'll be doing more testing later today/tomorrow."}
{"_id":"q-en-nixpkgs-e6ad6bed6a0b9fb59745e200fc3fd4c4669ab690b6aaa10215d63d562087b90e","text":"Package name: emscripten Latest released version: 3.1.57 <!-- Search your package here: --- Current version on the unstable channel: 3.1.51 Current version on the stable/release channel: 3.1.47 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\ntl;dr: 3.1.57 is not currently compatible with the version of llvm in tree (18.1.5). The latest compatible version with llvm 18.1.5 appears to be Emscripten 3.1.54. The emscripten nixpkg (3.1.51) currently uses an LLVM17 environment which needs be migrated to LLVM18. The configuration as it exists today doesn't work in some cases. (Godot4.3dev6 web template builds failed missing in , prompting finding this issue) Emscripten 3.1.51 seems to be compatible with the LLVM 18.1.5 currently in nixpkgs unstable. Emscripten 3.1.59 (current as of this comment) and 3.1.57 are not compatible with that version of LLVM (In my usage, ran into errors with and ). These flags were introduced in Emscripten 3.1.55 making the last compatible version against the current in-tree dependencies 3.1.54. <details<summaryThis override was able to successfully build the godot4.3dev6 web templates using Emscripten 3.1.54</summary<p</p</details"}
{"_id":"q-en-nixpkgs-f4e3529aeaf72d64158488db9f9c05260360693d542dc35f4c3cf490cc9d350f","text":"Project description LXGW WenKai TC is the Traditional Chinese Edition of LXGW WenKai, where the latter has already been packaged as . Metadata homepage URL: <https URL: ofl platforms: all Below is a working package build script I tested locally, which is largely based on 's script: (I commented out field.\nIf that snippet is working, could you turn it into a PR? It looks good enough for inclusion to me."}
{"_id":"q-en-nixpkgs-69878f074b9be318d2b918763aaa5c04298cd30d2f4bf992bed20257e2eaad5b","text":"N/A No maintainers are defined, but was last to update it Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-37fb74931d817fcd89534c60217519fe656611a872eeb777de31d035bc081a45","text":"Package name: fluent-gtk-theme Latest released version: 2024-04-28 <!-- Search your package here: --- Current version on the unstable channel: 2023-06-20 Current version on the stable/release channel: 2023-06-20 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-bbef23c677f7768515c43b90c31393c9cff23ff1c8821142124ce4beb3809630","text":"uses and to convert PCI device ID to human-readable product name. Currently their paths are hard coded as and . It won't work with NixOS of course. I found some packages that have similar issues. They are patched directly before building: Fastfetch needs to be patched too. Steps to reproduce the behavior: N/A Fastfetch can properly find and N/A Reference: Fastfetch provided cmake config option  for but not for . I can add one if needed. N/A Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHi thanks for the detailed issue report! This should be pretty straightforward to fix. If you could add a config option for the path that would be great! Many thanks.\n.\nwas released with support\nMany thanks! Much appreciated."}
{"_id":"q-en-nixpkgs-3511c469d1c48263b7cf56a1926d402f91445eff6d61688442978d4e922e4692","text":"On nixpkgs commit , running Tailscale with Tailscale SSH connected, then trying to connect to it, results in an error: . Steps to reproduce the behavior: the tailscale node service with (in my case, using a systemd service). to use Tailscale SSH to access the node (e.g. ). that no shell is opened and the connection is closed immediately. A shell opens (if the default SSH behavior is unchanged). This appears to the result of The underlying patch sets the systemd hardening option , along with an allowlist of devices. This prevents from accessing , which is needed to open a PTY for the shell. Please run and paste the result."}
{"_id":"q-en-nixpkgs-31875d8174eb94e6d7b0955e73712b77a3ce6a2b54e12e436f98eb366a931271","text":"Affects our 3.12.17, 3.13.10 and 3.14.1.\nOnly 3.13 is vulnerable now, but I don't know what to do about it. Personally I'd drop it, but it's also possible to follow the Ubuntu-maintained branch if people care for it, or just leave it as it is...\nRemoving 3.13 is fine - we were going to initially keep it because grsecurity was going to use it, but it looks like they've chosen the 3.14 kernel instead.\nI have no idea if anyone else can depend on it, but reverting will be easy.\nAfter finishing build tests locally, I'll push to release and close this.\nWell, doesn't build on release (it does on master), but it's probably not worth debugging. The rest builds fine."}
{"_id":"q-en-nixpkgs-ab37d427f34e16a4ee14b7644b4aa7436c08205c38209c28dad6a967fa7017c9","text":"Updating my falke to latest nixpkgs unstable gives an error that vault is non-free. But my package doesn't use or need vault. My package at has: This gives the error: If I comment out the it works. Steps to reproduce the behavior: to use nixpkgs-unstable instead of nixpkgs 23.11 Should build without errors. As above. ` Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nhelm-secrets depends on vault\nWhat happened to vault? It use to be open source I thought. Did the license change?\nDoes helm-secrets really depend on vault? From the it says: Which seems to indicate that it should be able to use any secret manager, not just vault.\nthe vault license change, which sparked quite the outrage. you ask a question and follow with the answer, that's both rare and impressive ^^'' You can see it here: I suggest you follow the recommendation and accept the new license for vault. If you dont need it, you can remove support for it as well from helm-secrets (but will require some nix skills though you can search the web for plenty of examples)\nHow about switching the dependency from vault to OpenBAO - would that be possible?\nHeh, I didn't realize LF forked Vault too. :popcorn: It's certainly possible, but might have unintended side effects for Nix users. (I realize how annoying that might sound, given your recent experience.) By default, helm-secrets supports and . After a quick glance, it looks like both sops and vals use the Vault Go library, so maybe we don't even need at runtime anyway... I wonder if sops and vals will have to alter their licenses... Edit: It seems the is still .\nI've just tried a modified helm-secrets derivation, removing all references to , against a repo that uses helm-secrets (with AWS KMS backend) and it worked fine. I then switched the config to point use a Vault backend, and that also worked. Note that I used to get a token, but assuming exists or the env var is set somehow, it seems doesn't need a binary at runtime at all. So, I suggest we drop the dependency.\nmight you be able to test It Works on my Machine, and I'm fairly confident, but I'd love some additional verification.\nIt seems to work OK on my system, so far. Don't have a good test case for vault though. Will think about additional ways to test this."}
{"_id":"q-en-nixpkgs-6800beb27e4c6200d9e28022e0e484efacd12ceeee5f52fce950f51e899c8854","text":"After installation notation shell completion is not available. Steps to reproduce the behavior: notation completion Shell completion should be installed. N/A N/A <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-afbf66b13e1a5e87e8c79499860a5f077b609be81b485f45752147b453a871bf","text":"After installation kcl-cli shell completion is not available. Steps to reproduce the behavior: kcl-cli completion Shell completion should be installed. N/A N/A <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-31e7302692a1cc045872a721d580e3326ef92bf2ff01b8395e6d7809fba07c01","text":"Steps to reproduce the behavior: unicode-emoji-15.0 :frowning_face: Here's an example log from my system building with substitutes disabled ( being a dependency of ) Note that the logs from building don't contain anything else. The fetched URL () from the Unicode website seems to be restricted in favor of the 15.1 version, which breaks building. A report has already been sent to Unicode, Inc via on their website. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAs a bit of a context: this error only occurs on forced local rebuilds, eg. because of disabled substition!"}
{"_id":"q-en-nixpkgs-7dda5702282a3f4be32837958c0cf40806962a392317a8fd455d8eab65a65ab1","text":"Package name: feishin Latest released version: 0.7.1 <!-- Search your package here: --- Current version on the unstable channel: 0.6.1 Current version on the stable/release channel: 0.5.1 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nthanks for reporting this, just noticed I hadn't subscribed to release notifications. Shouldn't miss any from now on :) Fixed by (awaiting review)."}
{"_id":"q-en-nixpkgs-bfe694ad2b05a5627eb7b054b8f27b479100efb3ac856ae9a2e0f1edffc869f7","text":"Gives this error: , yet installing doesn't fix the problem. Steps to reproduce the behavior: Meteo That the Maps section works. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI have this exact issue. Hoping for some clarity.\nshould be here:"}
{"_id":"q-en-nixpkgs-bc335a9772de89c8efb16157fe30d394629c74d0b71ded95e3f320266397be74","text":"Package name: tracker-miners Latest released version: 3.7.3 <!-- Search your package here: --- Current version on the unstable channel: 3.7.2 Current version on the stable/release channel: 3.6.2 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Tracker was updated in , however tracker-miners wasn't. It currently suffers from this bug: (and generates a lot of failures on a running gnome system). It was fixed in 3.7.3. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nPRs are welcomed."}
{"_id":"q-en-nixpkgs-3ae8e3a44e0fa69fd1db9a91c1b04be670c35fd8f0d209b7157b13f85088d58a","text":"Package name:material-symbols Latest released version: 2024.05.04 <!-- Search your package here: --- Current version on the unstable channel: 2023.11.17 Current version on the stable/release channel: 2023.11.17 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAre there any relevant changes in the font definition?\nthere should be some new symbols or some symbols' code changed.\nplease do replace OLDVERSION and NEWVERSION in the issue title with the actual version names. In this case this would be like:"}
{"_id":"q-en-nixpkgs-bd231fc431519331292b923bea0025a0561431e03188af6ba5ef0ba1df5835d0","text":"There is an option to save the configuration state that does not work because it tries to run which is not located at this path in NixOS the GUI to File -Save Configuration filename and click save It should save a file to the filesystem with the current configuration state."}
{"_id":"q-en-nixpkgs-5bd14985cdc8cda3df6a9c2252a7fff71e6936169a8445a47f00da607d7f37ad","text":"Project description <!-- Describe the project a little: --Metadata homepage URL: https://mqtt- source URL: license: creative commons platforms: linux, darwin Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAppImage works fine with appimage-run. There is also a deb and snap package\nI use this one\nThat package looks great. Maybe even worth to try to contribute to nixpkgs.\nUnfortunately I don't think the appImage version would be accepted since the source code is available with a build manual so imho a PR to nixpkgs would need to be with a source build. I might look into it\nI maybe have a potentially working version. I'll try to find some macs to test out the Darwin versions and if that's gonna be okay I'll then submit a PR\nGreat news, I can test on darwin."}
{"_id":"q-en-nixpkgs-435dbf1cc2e88b02486427050a40780ba1409ed2e3a1ca869ad58bc7dbbf424a","text":"System: 18.03pre118631. (Impala) Nix version: nix-env (Nix) 1.11.15 Nixpkgs version: \"18.03pre118631.\" Sandboxing enabled: build-use-sandbox = true\nSee, I thought I had fixed it somewhere but cannot find it ATM."}
{"_id":"q-en-nixpkgs-a8109068cdbc6e2d91a56fe44927e8969a1962ff24539780fcef279792680bee","text":"Project description \"A general purpose HTTP reverse proxy and forwarding tool.\" Metadata homepage URL: source URL: license: agpl-3.0 platforms: linux Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-6b75f64b66896f2c541e7c079ce9ca89dfc5e98a85fad88e1fbc3d91cc8d7b0b","text":"hydra went from to as per on a . Steps to reproduce the behavior: to nvimpager <details<summarytest failing on this update:</summary</details\nTo my knowledge the actual functionality of nvimpager is not affected but the default colorscheme in neovim changed and my test depend on it. So a quickfix might be if that is acceptable.\nthis will be fixed by .\nThe neovim update has been reverted in . Closing."}
{"_id":"q-en-nixpkgs-25545a79235c9870f9c6a2d7fa34cbd52e95ead643827d337d993367951f9c4d","text":"When enabling freshrss and rebuilding, no freshrss systemd service is started. Steps to reproduce the behavior: (as well other requirements for minimal config) (with , for me) The only freshrss units are and . to in browser. Unable to connect. A service with (or something?) is created and runs automatically. The http server is accessible from a browser. I initially tried running in a container, but running on the host OS had the same result Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nFreshrss doesn't run as a systemd service. It runs under php-fpm. The associated systemd service is . This is then exposed via nginx. I believe the default nginx configuration will expose freshrss on port 80, so http://localhost may work for your current configuration. The used by nginx for freshrss is . My guess is you'll need to configure it. If you want to access freshrss via \"http://localhost:8080\" you could use a config like:\nSorry I never responded to this! This worked, thank you.\ni guess we can close this then? please report back if i am wrong.\nI do agree with qyriad that this could be explicitly documented better, but I don't have any further issue to specifically track\nany specific location where you want this documented? where would you look something like this up?\nI would probably put in the description of the option."}
{"_id":"q-en-nixpkgs-16d7906b8b84a8908a22c632e45c3f34473e3f9942fcf681ddde50060c61c80f","text":"Slack reproducibly crashes: Steps to reproduce the behavior: to Sway Slack It should not crash If applicable, add screenshots to help explain your problem. Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-bf7b6e1cf4543ca5b1f91c070cd4e0ed8a0c28db763fd98c28d05a680d3daef5","text":"Steps to reproduce the behavior: <details<summaryBuild log</summary</detailsOut of the few packages I've tried installing using , all of them failed with \"detected dubious ownership in repository\". It builds successfully if I use the commit that introduced : Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nGit bisect says is the first bad commit.\nnot sure if the error is intentional, but the changes in that version are explained in\nOkay, it seems intentional: So we have to change the julia builder accordingly, since anything in is presumably owned by , which is not the user used in the build. Not sure if that'd require an upstream fix, I'm not familiar with julia.\nPR here:"}
{"_id":"q-en-nixpkgs-da4c66428c603b7bc3cbc44516083b445f7d920d693264847dac0fa9ae620c5b","text":"When Freshrss is first taken into use the service presents configuration wizard where it asks for language and complains of invalid file permissions. The service should configure these automatically, but it does not. Steps to reproduce the behavior: freshrss with nixos rebuild \"http://<MYDOMAINHERE\" to web browser 's web page opens to configuration wizard Freshrss's web page should open to the \"Main stream\" page. This can be fixed by manually starting which does the configuration. It does not get started automatically when you first enable freshrss though, with no indication that you should start it yourself. Which led in my case to much wasted hours and frustration. Example output from rebuild: ^-- No at the list of new started units Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nCan you see if adding the following configuration fixes this for you?\nHi, seems that it does indeed fix it. :+1: Tested this by starting from scratch: freshrss's nix configuration nixos rebuild accessing freshrss's web page and got 404 error to another location freshrss's nix configuration back in and applied your fix nixos rebuild again accessing freshrss's web page and got the expected main stream page Also another side benefit is that it is now possible to change admin's password via . Previously it required also running of manually afterwards."}
{"_id":"q-en-nixpkgs-68358c4a25685b612cdbac4d08e9f4d83bf6d0712a5dc663570c716921e0183f","text":"Package name: Latest released version: 46.0 <!-- Search your package here: --- Current version on the unstable channel: 46.0 Current version on the stable/release channel: 46.0 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]: Fixes very annoying bug for users with nvidia dGPU notebooks on gnome using external displays. Reports of the Bug:"}
{"_id":"q-en-nixpkgs-3e0b003d6e40d7138deb055a88e8f54d9dad3c7a72161fd9dc32e6ea87a87e10","text":"Package name: python312 Latest released version: 3.12.4 <!-- Search your package here: --- Current version on the unstable channel: 3.12.3 Current version on the stable/release channel: 3.12.3 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-6814f5d88fac5e8a71c5a0d99625063f3d836acbbe8d71a69079e86e6b8330b8","text":"Package name: monado Latest released version: 2024-06-07 (24.0.0) <!-- Search your package here: --- Current version on the unstable channel: 2024-01-02 Current version on the stable/release channel: 2024-01-02 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nPR at\nThank you!"}
{"_id":"q-en-nixpkgs-b3da356c5802b42250086d69f62c5bafe2f588d4af14756ea0e55aa7d50eb13c","text":"Hostname being unset (resulting in ) causes issues for logging. Specifically, with journald, the same boot (identified by (see systemd.journal-fields(7)) has both and with the switch happening at the rootfs pivot of the boot. This makes filtering by hostname difficult. Even more so, when doing log aggregation from multiple machines, filtering by hostname will yield incomplete results. Instead one must build a pipeline that makes looks up all bootids of a hostname and querying all those instead. Steps to reproduce the behavior: behavior described above journal contains the the same as long as it not changed by the user. Hostname should be generally available, even during early boot(?). N/A I'm filing this as a bug instead of a pull request because I'm not sure if there are any unintended consequences of doing this. Aside from the usual Hoping for your input here. There's a reference in the docs giving an example of setting the hostname by hand: (from ). Why this was chosen as an example and not just standard behavior, I don't know. CC/FYI , who the example/was tagged in the commit. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nMe neither. I don't see a reason not to have that file. may be more knowledgeable, but I'd say go for it. Also systemd-based initrd is somewhat young, so I feel like the probability for spacebar heating is somewhat low.\nI think the main reason was to have the initrds be more substitutable between different hosts as this would be a difference between all of them (when having different systems with only different hostnames). I'm not sure how well adding the hostname file would work. My memory is a bit fuzzy on the details (I did a lot of other projects at that time) but I think an issue was that when journald starts up, it imports the kernel logs BEFORE the hostname was set (causing the pre-systemd lines to be mapped to localhost). An alternative implementation that would probably solve that as well (if it's still/even an issue) would be adding the kernel parameter rather than adding the file to the initrd.\nIf I read the code correctly, journald reads the hostname from /proc: -: So the hostname needs to be set system-wide before journald starts up. Systemd does this in --: So on bare metal we can either set on the kernel cmdline or it seems.\nIt seems this is the best choice then, but maybe also as that's more standard and picked up even earlier in the boot process. Unless you're running an old kernel though. () We could just set both.\nI'll run an experiment to see if even solves my problem or one must use . On that note, I also realized there's an equivalent (or slightly harder?) problem with as well :-.\n(So, setting /etc/hostname in the initrd works. Every message from that boot has the correct hostname.) I was also thinking about your folks' usecase some more, and was wondering about one thing ... how are you approaching the UKI/secureboot? None of the mentioned approaches work for that, right?\nWhy not? Most people use for that which signs whatever UKI you have locally with a local key\nBecause changing files or the kernel cmdline will cause different images to be created? Your design shares a single one among multiple hosts, no?\nYes, that's the downside. Although, the cmdline is only baked into the UKI which is not in the nix store anyway. This way we could substitute the initrd and have differing hostnames. I must mention, however, that I personally don't care about the substitutability of the initrd. My main focus was to keep the state as it was before, but if people decide it's time to change that, I'm not opposed to that ;)"}
{"_id":"q-en-nixpkgs-bface8ea401f5718c17838e9f3fc1f44f85f980976d1b1b8889880000909a709","text":"The service will ask docker to mount the binary which is not provided by the Nvidia driver v470. Using NixOS 24.05 with nixpkgs . Steps to reproduce the behavior: should not be mounted when the binary is missing. <img width=\"1234\" alt=\"image\" src=\"\"<img width=\"854\" alt=\"image\" src=\"\" Edited: a workaround Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThank you for the report You can also set in the meantime. I'll open a PR as soon as possible."}
{"_id":"q-en-nixpkgs-591810796cc370bb0c662334fcfcb36b61ff0f4b72ce7b04fb78a8841d9dcb86","text":"Package name: fluent-gtk-theme Latest released version: 2024-06-12 <!-- Search your package here: --- Current version on the unstable channel: 2024-04-28 Current version on the stable/release channel: 2024-04-28 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-7375cfb37ea0c38aa3f622ccc241d4b040c3beeac98f874d719fabc896088c89","text":"Commit seems to have broken rexml. How to reproduce the behavior: It should work as it did in commit . It looks like this is related to , but currently fails on master. system: host os: multi-user?: sandbox: version: channels(root): channels(username): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nShoot! Looks like it's already fixed upstream: Looking to submit another upgrade PR. I just want to check if there would have been a way to catch this, so that it doesn't happen again.\nA test running would've been enough in this case. Do you want me to look into creating one, or to help you with that?\nI confirmed that the upgrade to 3.3.0 fixes github-pages, but I wanted to make sure it doesn't break any other gems. It looks like kramdown-rfc2629 is the only one causing issues. kramdown-rfc2629 was fixed by pinning the old version of rexml as a temporary solution for the issue. I'll pin the old version of kramdown-rfc2629 (the only difference is the rexml pin) and submit a PR upstream to remove the pin (now that rexml has been fixed).\nShould fix it.\nit has not fixed it for me, using selenium-webdriver gem. Details at Should we be opening a new issue for it?\nNevermind, I think this is the entire issue,"}
{"_id":"q-en-nixpkgs-0eabcdf0b1499375dc543dc874cc1ac048fbe1cc7560beb6092ac4ce84eec35f","text":"Steps to reproduce the behavior: This seems to have been caused by although I'm not familiar with the package, so I could be wrong. It is causing to fail as a result. I don't quite understand the organization of the structure, so I'm not sure who the appropriate maintainer is. As a fallback, and were involved in the above linked PR, so they may be of help here. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nsorry for the breakage, i opened to fix this\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nHad the same issue. But I think it's fixed in PR as mentioned You can track the progress of the PR , apparently it hasn't been merged into nixpkgs-unstable yet."}
{"_id":"q-en-nixpkgs-9431e5e830ab0a1a48c56c93a28fe38f674f5a29f6f6eab0cf53cf73a4d95731","text":"Steps to reproduce the behavior: floorp Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHi, Please elaborate - there does not seem to be any actual build failure here. What's the actual error? Are you sure your build machine has enough resources? A browser build is pretty demanding and can take even multiple hours on low-to-mid-end machines. Can you please also post the complete derivation log, as obtained using ?\nI have the same problem (with floorp and floorp-unwrapped). Here’s the error: ! That feature was removed in February, so either update the qcms crate to 0.3 or downgrade the rust compiler version (1.76 could work).\nis the log in the link not enough ?\nOh, I see, Github just truncated the file in the UI. That's what you get from hosting projects on crappy, proprietary software .. Pointing out the exact error too from the get-go would have been helpful in any case. Anyway - seems it broke with a recent rustc upgrade in nixpkgs. This is already fixed with 11.14.1, as it updates the firefox base source. See .\nOkay, I will host my opengist instance for the next time ;) Thanks for your prompt response, waiting for your pr to be merged.\nAny ETA on the PR merge?\nFloorp is now in version 11.14.1 on unstable."}
{"_id":"q-en-nixpkgs-5fd2f06bb5cde7c688ebea8f1e3572b9b49dee7426c70beefdc9adc78c15ffd6","text":"Project description Sysmenu is a simple and efficient application launcher written in gtkmm4 Metadata homepage URL: source URL: license: wtfpl platforms: linux Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-2c86d3de8edc3dadc3a6c899b47234a316e4bd65c18cad82fb692e38f31df604","text":"Package name: bcachefs-tools Latest released version: 1.9.1 <!-- Search your package here: --- Current version on the unstable channel: 1.7.0-unstable-2024-05-09 Current version on the stable/release channel: 1.7.0-unstable-2024-05-09 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-0c8f5976efa2adf115a9b6c5623cd1c71ad892072a3785fc4110062a5ef0adf9","text":"The gitlab  script depends on , which was removed by in . Steps to reproduce the behavior: The update scripts succeeds. (myself) Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAlso looks like nix-update could be used here.\nOk. looks like switching to nix-update would require some other major refactoring because the writes out some json file out. But here is a simpler approach:"}
{"_id":"q-en-nixpkgs-deec6c6764f2e32d9c5908e7f7ad79baab19f3b22765e50e5cf8af72e5e5ad40","text":"Mealie was updated from 1.7.0 to 1.9.0, and it's now failing to launch, even though I'm using the exact same configuration. It actually seems that just enabling Mealie, with no configuration isn't working. What's weird is that the error is referring to , which is a sops-nix directory. I even rebooted, but it's still getting that particular error Use the following configuration: Mealie successfully starts. Here's my original Nix configuration: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nTested on VM, works fine with the base configuration. The problem comes from the part of your configuration, it seams that you need the setup the user to get access to the directory. I'm not familiar with , is there a group you could add the user to so that it can access the directory ?\nYou can allow a specific user/group access to the decrypted sops secret file using the options. See this example in my config:\nIt still doesn't work after doing that. I wonder if it's some state getting left over that screws something up, because it's weird it would still not work after removing everything but . I deleted though, and it's still happening, and it also still works if I role back to 1.7.0. Although, I am using PostgreSQL so maybe it's something in there? Note that it's happening in . Here's my updated Nix file:\nThe issue seems to have something to do with this mealie commit: Mealie 1.9.0 works when the commit is reverted:\nSomebody opened a PR to get this fixed upstream:\nI've also opened , which prevents mealie from exploding when it encounters a secrets directory that it doesn't have permission to read. I've verified that it fixes this issue in my own NixOS deployment.\nhas been merged and should be out in the next release. In the meantime, we can pull in to the currently available v1.11.0 to fix the bug. I'll create a PR for this.\nI'm still experiencing this issue when Mealie first starts after boot, but it goes away after restarting the service. I'm guessing it's because sops-nix hasn't setup the directory yet? Should I open a new issue?\nThat warning message seems to be intentional. The patch always prints a warning before the directory is ignored."}
{"_id":"q-en-nixpkgs-c5f57c9fe9c67a494942a0aa4422a07ddf1b3be814b9179ca1e165a36fa780a6","text":"Steps to reproduce the behavior: gerbolyze Test failures: and Upstream issue . While upstream is working on a fix, the issue can be worked around using the shell (run ) from the following : Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-32111b885cfc0529c3a59ab6beaf564f35ac4e967520898d689c34848ddbdb18","text":"Application crashes with when scrolling back chat to images/videos. I'm not really sure what exactly triggers a crash, but the project's README says that when viewing media. Steps to reproduce the behavior: . a chat room with images/videos. a crash No crash. Nope. I've also tried overriding the package locally by adding gst plugins to , and when it didn't work, to like so: but that didn't help. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --- system: host os: multi-user?: sandbox: version: nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-d11cf43729ad7acd38b770c4ff5090e429b31e755bcd8553146ffaf20e8da93a","text":"notes that this is a fixme.\ncc\nI use Nix as a package manager in Ubuntu, so I can't really test this. In Ubuntu, drops the file in , and the user is expected to change on their own.\nI think the best approach would be to generate an script that returns the path to . Then it would work in the cases where it's installed in a random profile too. As long as it's on the PATH you can use\nPR ^"}
{"_id":"q-en-nixpkgs-f7da54e340c3eacc9429e7c52c397f5545d56bd6bce4d0fa3900e82519c698ba","text":"Steps to reproduce the behavior: unstable defined, enable the ollama service with pkg overridden to point at unstable: without show-trace: with --show-trace: Add any other context about the problem here. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHi! Can you confirm your nixpkgs revision has this change incorporated and you're not overriding the package manually?"}
{"_id":"q-en-nixpkgs-ad49893424e697b5840b4c7c43a8c00f88c03fff643315e0890d311cc80aa69c","text":"Project description Session persistence tool: \"Think tmux, then aim... lower\" Metadata homepage URL: source URL: (Cargo Rust) license: Apache-2.0 platforms: linux note: primarily useful for nixos servers, not workstations!"}
{"_id":"q-en-nixpkgs-876e874df98bab41a408448c0fbab7ca07a9fc7f2522cc86a5f6eb6791f1697a","text":"shell completion does not work with fastfetch fastfatch fish and shell completion works ! Add any other context about the problem here. system: host os: multi-user?: sandbox: version: nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-23aa1975f73277aa15a487c1696be123c1e21d7f60a993b39a3c9a2ff01d5610","text":"Steps to reproduce the behavior: spotdl/asciinema/binwalk/gpodder/autotiling/... Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI am getting the same issue from binwalk after updating flake inputs: the same error also occurs when trying to install the package xsser or pwndbg. So it seems to affect a lot of packages using python.\nAlso having this, I do not have Spotdl installed however. I think it might be a underlying issue with some dependency\nGot the same error with gpodder\nSame with blender It seems multiple builds are failing here\nI also had issue with Blender-hip but different error I am not sure if it is related or not it was issue with OpenUSD which is NVIDIAs weird model format or something(?)\ngetting the same issue for & asciinema logs: autotiling/i3ipc\nSame issue for and .\nI fixed in . The PR just replaces the package (unsupported in python 3.12) with () .\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nSame error with plex-mpv-shim Created for plex-mpv-shim EDIT: appears to be fixed on master already\nseems to be the latest working commit (one before )\nSame with plasma-hud and gnuradio for library thrift.\nupd: also yep, and is also have this issue hi! same think, this time with , im not a big expert on flake and nix, but maybe there some simple fix like to pin old python version or something?\nBinwalk is also having this issue:\nSame issue with\nsee and fixed by\nthe issue is still there. Here the trace:\nMy issue will be fixed by\nI have issue during nix-darwin update\nIssue persist for nix-darwin update using flakes\nSame issue for pony-0.7.17 (pgadmin)\nPlease open separate issues for all packages that haven't been fixed on master yet. This issue was about spotdl. Btw, nose will work on Python 3.12 once or is merged."}
{"_id":"q-en-nixpkgs-cf950f62f7c7ba05446dc268d0cf83ee679af7282ba8495da08525ddf829cfb4","text":"Steps to reproduce the behavior: build .#hplip Looks like more fallout from Python 3.12. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-cf1433320d50baeb5cd763206a8b958a9fb113fdd0feacd59a2eb112993c0f14","text":"Package name: dbeaver-bin Latest released version: 24.1.2 <!-- Search your package here: --- Current version on the unstable channel: 24.1.1 Current version on the stable/release channel: 24.0.5 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-d8ad2db3b872855dc7563bbacaf08702b012755529f186de659fa99eff37c7d8","text":"Steps to reproduce the behavior: floorp or floorp-unwrapped The important bit at the end of the log: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nYep, got the same.\nIt probably needs something like For now, I've just this overlay:\nYep, seems like the cause/fix. I'll send the update to v11.15.0 soon anyway and include this fix as necessary. The package in the binary cache is fine though, so if you just pull it from there instead of rebuilding this massive thing you'd be just fine too."}
{"_id":"q-en-nixpkgs-51f464375ff051201340058bb300cebaa3d73c41cdbaf5d3ae4855c9d13fdccd","text":"I'm running an issue with overriding the rust toolchain since , I pass the following to crate2nix: to set the toolchain when building something, since I don't want to override the toolchain globally. But after this PR it doesn't work because it uses the toolchain in instead of the one I'm passing in. Now I don't really know how to override the rust toolchain to the version I want. is a nightly version I got from This problem is caused by this optimization: Another issue is that the attribute is missing in rust-overlay:\nI'm having this issue as well. I've overridden my input to my flake to as a workaround for now.\nHi, sorry for the delay, I'm back home now from \"holiday\". Can any of you give me a minimal flake that reproduces the issue so I can experiment and find the best way to fix it?\nI'm also hitting this. I don't have a minimal flake but I can at least offer the error message which seems to be missing from the OP.\nLet me know if this fixes it.\nunfortunately it does not, because it still defaults to for the rustc compiler. I've created a reproduction here: As you can see in I'm overriding rustc and cargo to the specific nightly version I want, but when I run I get the following output: This is the rustc version that's current in nixpkgs, not the nightly I specified. Thanks a lot for looking into this\nOne problem with your reproducer is that it uses the for which takes an arg called not . I think that still wouldn't have solved it. I made some changes in the PR that would allow passing a custom toolchain via the but also using . I did some refactoring to get rid of the and variables."}
{"_id":"q-en-nixpkgs-c079d7f167ad4ae391427a8631733f50b84146e15bed0720b7d5ec3d3cb51c81","text":"gamescope 3.14.23 seems to be packaged incorrectly, looks like this automatic upgrade missed some new runtime dependency, as the command now fails when run. Steps to reproduce the behavior: A window is displayed ! I am running on KDE Plasma 6.1, but this issue seems to happen on a standalone session as well. I'm also running Nvidia hardware, RTX 4090, on driver 555.58.02 Written logs: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nDid a quick test to tryout the previous gamescope version, 3.14.22: Seems like the issue was definitely introduced in 3.14.23. !\nAre you running it from nix-shell or something? should be in the package and thus on your PATH.\nNo, it's part of my system configuration flake. Sharing the relevant host file for context, is the only part of my configuration that is related to gamescope (): My configuration lives here: Some changes I haven't pushed yet. You can see by my that 'unstable' referes to nixos-unstable channel, and stable to the latest stable release, 24.05.\nAt the very least you can't mix unstable gamescope and stable WSI.\nThank you for pointing that out, I changed it to both be on \"unstable\" but the issue persists.\nIt looks to me like \"gamescoperreaper\" is a new binary that Valve recently with 3.14.23 that gamescope calls at some point. I was looking into their releases . Not a CPP developer so I am making a some assumptions about the nature of gamescopereaper and the layout of their project. Perhaps gamescopereaper needs to be packaged as part of the gamescope derivation? It doesn't look like it does that currently.\nThe binary is there in our gamescope package, though?\nAgh, seems like adding though the was not enough to add into the path. I solved the issue by adding the package explicitly inside of .\nI think the issue is with the module -- looks like it doesn't actually add the gamescope package into , but instead a wrapper.\nI can confirm that adding the package through the gamescope module makes gamescope not work. Adding it through other means such as does work.\nOk, so possible solutions: 1 - simple but less optimal: add a symlink below this line: 2 - complex but more Nix-friendly: find the exec portion in C code and patch it to find the executable in $out/bin/ 3 - alternative similar to (1): add $out/bin to wrapper's PATH (using makeWrapper argument) Good opportunity to open a PR with any of these solutions, tag the module and package's maintainers, and make the requested changes (which might be even a change of which strategy). Don't worry, the worst thing that could happen is your PR being closed to favorite a committer's PR, it's still seen as a good contribution (and good opportunity to learn).\nOpened to attempt to resolve."}
{"_id":"q-en-nixpkgs-c8b07626a629ef9d8e58f71be4e9be0647bc1c432d98450538251529bc6943fc","text":"Ansel installs well, and can be launch from console with … but it does not appear in Gnome’s application menu, nor it cannot be find in Gnome launcher search Steps to reproduce the behavior: to from channel 24.05 or from unstable to launch it or to find it in applications menu Like every other application (and Darktable), I expect to find Ansel in the application menu, launcher search, and so on\nI don't know anything about gnome, but there's definitely a desktop file in the correct spot: also, is a bot, FYI Probably try rebooting or something.\nThe desktop file is broken: Sounds like the upstream build system is broken – see\nGood catch, thanks. I've sent a PR"}
{"_id":"q-en-nixpkgs-2424cb13886edde0abf16209c11ca6f9514a132c702f4fdb7e091142cbff79e9","text":"I believe converted the following generated line of Bash: into: , so it's , and is the empty string; but when escaped, it became . The result is that the passthru does not generate a lockfile. After I manually change the two into in the definitions of and , I successfully start getting a NuGet restore. Failing passthru: Add any other context about the problem here. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHm, that is weird. Note that cannot be null.\nThis should be (plural). Introduced in .\nI even checked the implementation of and discovered that it's secretly a function rather than as documented, and didn't make the connection. I clearly need to go to sleep."}
{"_id":"q-en-nixpkgs-b15694b94fcea10728c29e1f12e46a01751b501e74bc7fe2864d0c570372ae0f","text":"Package name: wasm-pack Latest released version: 0.13.0 <!-- Search your package here: --- Current version on the unstable channel: 0.12.1 Current version on the stable/release channel: 0.12.1 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. I just tried to make a PR with the change, but I got when I tried to build (while compiling ), and used all of my RAM while running (it did not finish). I'm using a x86-64 machine with Debian.\nI encountered that error too. After reading I tried this: That resulted in a new error log. I don't know what to do but I thought I'll leave it here in the case that I could be useful to someone else"}
{"_id":"q-en-nixpkgs-9f0c8968c76eb74cbaf1e992eb92626d98fa56f47e6af3d91dd0b34b0249eaba","text":"Package name: Latest released version: 1.4.3 Current version on the unstable channel: 1.4.2-1-unstable-2024-07-10 Current version on the stable/release channel: none Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-6f7e9748807337e742daa10f9be7481487857771847cccf309f5ecd2249d9532","text":"doesn't seem to work as a module Steps to reproduce the behavior: livebook service any env variables or use environment file with any variables first of all, its systemd unit does not appear to start by default. so i did and but it doesn't work, i can't access it either by IP or by domain, which is set up in my nginx configuration as a reverse proxy. some strange things happen while the unit is active [1st screenshot]. also there's nothing in the journal except info messages about the unit itself as i see it, livebook fails to start, but the service is active because something is going on with inside this wrapper script i tried to make it work several different ways: from using option instead of my environment file to writing my very own systemd unit in NixOS configuration file, also using home-manager like so: but that doesn't work too. surely i tried starting livebook manually - and it works as expected [2nd screenshot] module/service works and is accessible systemd livebook unit works and is accessible |\n---------------| NixOS 24.05.3214. (Uakari) x86_64 running on a VPS cc Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThank you for reporting this issue. I am ashamed to admit that I have actually known about this for a while now, without doing anything about it. Now that someone else is having problems with the service, I'll prioritize getting something done soon. I'm not actually sure what's going wrong, but it seems like the service hangs in the command. As far as I remember it ends up using 100% cpu and just loop forever. I'll try to take a look at it later this week.\nPerhaps you have encountered this issue as well?\nI've seen this before and I think it was because livebook was trying to generate a in some weird way. This is my current configuration which works: Could you try manually setting a in your service setup, please?\na screenshot from filtering cpu usage ! not so much is used i've tried setting up a cookie and you can see it works - :pleading_face: thanks!!!! ~<\nThe issue seems to have been described and debugged here: The solution is to set in the systemd service configuration. I'll mint a PR."}
{"_id":"q-en-nixpkgs-44b134a8c0fb0e1171401c422f4cf9c146190a7b972314990667a2fa96364b9d","text":"Switcheroo fails with the error in the title. This does not happen in the Flatpak version (tested). (the formats might not matter, but I've specified them just in case) Steps to reproduce the behavior: a PNG file in Switcheroo to a JPEG file. the output filename. should happen. The image should convert without any errors. ! Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nOops, I had in my path when I originally packaged and tested this. I have it as a dependency in and tested it with this time. Thank you for trying the flatpak and reporting it here first instead of upstream :slightlysmilingface:"}
{"_id":"q-en-nixpkgs-34fdbe9fb2fd138034b598a7206b24922d1261dd948cf4bcf1c4f93fadec9b39","text":"doesn't work with channel configuration. Output: Only with channels. Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThat should be enough to fix it, I think?\nDid you try it and did it work? If so, feel free to shoot a PR with that patch.\nworks for me. I didn't check any other flag combinations though. But I guess working sometimes is better than working never.\nIt'd be great if you could create a PR then :)"}
{"_id":"q-en-nixpkgs-759ee4b974b6e8d4391cc4238027b5cca1d46903f4fc924cc80e540d219a9e06","text":"A clear and concise description of what the bug is. Steps to reproduce the behavior: lack of directory with executables In contrast this command works as expected: Create directory with executables to invoke Guake. Observe that nixpkgs 24.05 has and in the \"Programs provided\" section. <img width=\"989\" alt=\"image\" src=\"\"And unstable does not despite both claiming to have version 3.10 of Guake: <img width=\"1201\" alt=\"image\" src=\"\"And both branches provide the same package definition. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nSeems like the cause is the below. Oddly the command fails but the build doesn't."}
{"_id":"q-en-nixpkgs-f2442375e7195ea046ed69469a80f019b15118163d2e0f11e93acd09a461234b","text":"Alpaca does not use system theme configuration. (probably has something to do with flatpak) Affected: GTK4 theme GTK4 polarity Cursor theme Steps to reproduce the behavior: customized NixOS system, example: up and run nixpkgs#alpaca on nixos-unstable: App follows system theme settings: GTK4 Theme GTK4 Polarity Cursor theme ! <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nIt works for me, with . Maybe there's something to do with your theme? (Is that some simple libadwaita stylesheets?) Also try setting .\ndidn't help, I will provide my gtk settings files. (they are symlinks to /nix/store btw) the theme is generated by stylix but under the hood it uses In any other situation I would assume it's a problem with my theme but this is the first time this happened to me on NixOS and I use a lot of gtk apps.\nYes, I've tried with , and it behaves the same. But there are probably other applications that behave similarly, because I set this environment variable a long time ago for similar reasons. I will try to read its source code later. (this may also be caused by me setting XDG paths)"}
{"_id":"q-en-nixpkgs-75c92fe92c9c6af2898a7034526abc06f85a87006ea4c5526cdd5500504677a7","text":"Package name: matrix-synapse Latest released version: v1.112.0 <!-- Search your package here: --- Current version on the unstable channel: 1.111.0 Current version on the stable/release channel: 1.111.0 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThey released 1.112.0 very quickly after 1.111.1, so I changed the issue."}
{"_id":"q-en-nixpkgs-d2cae625be4fcdb9a58caa184a2bee70928105bb3260442b2e5d6bfff4bf400a","text":"Steps to reproduce the behavior: mysql-workbench-8.0.38 Might be related to or Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-5e0f8d22c92b9b7f408d0624dd70e6379987612051de2318bc010ab196901652","text":"Steps to reproduce the behavior: ncdns.goModules with or Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThe goModules derivation can't be built since go 1.17, see I don't think I can fix it."}
{"_id":"q-en-nixpkgs-0fe0e8ef8b64c0ea3f13c4e7fdc2dfedeb9283f2a869398f4beaeb520df0614d","text":"Running the package fails, apparently because it uses as input instead of as would be required to run:"}
{"_id":"q-en-nixpkgs-5257102a2d301ada9d39dca6e08180dd62abb54c0844af89d6d242c54add4309","text":"Project description Microsoft Azure Command-Line Tools Storage-preview Extension Metadata homepage URL: source URL: license: mit platforms: unix, linux, windows This package is distinct from the package that is already in 24.05 and unstable. It enables azcopy utility in the azure-cli, which is useful for interacting with large storage accounts in Azure.\nCode for extension\nplease give a try and leave some feedback. :)"}
{"_id":"q-en-nixpkgs-ed291f23c4ed665f8666f89143f3eaa5c7db6ac5f3677772ba0e57fa072ab518","text":"Looks like CHADtree was updated recently, and it needs a new . was to in a week ago. Steps to reproduce the behavior: CHADtree to open the tree view Should work. :) (Skipping this, since I'm going to push a patch myself.) Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-da5f9b0a8221fdaeb8c3f1f8e55a42c4e7498dd178ac33fa450bc04306f29955","text":"When installing the package and rebooting the desktop takes longer to load then usual and then only displays a black screen even without selecting the whitesur theme. Opening applications e.g. via shortcuts still works. I am new to nix so please wirte if further information is needed. Steps to reproduce the behavior: whitesur-kde package in the your nix configuration Desktop works as expected ! My current configuration: <details</summary</details<details</summary</details<details</summary</details<details</summary</detailsjournctl log: <details<summaryjournalctl</summary</details<!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\ncan you review , please?"}
{"_id":"q-en-nixpkgs-46890a5ff2d56c50cb90446d656b15e3cbe047addf94be403a8f21a298a2050d","text":"Package name: qrtool Latest released version: 0.11.4 <!-- Search your package here: --- Current version on the unstable channel: 0.11.4 Current version on the stable/release channel: 0.10.10 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. -- qrtool 0.11.4 was , and was updated in , but now (homepage), (repo) and even (owner) become 404 Not Found. I have no idea what have happened… Besides, still works. Update Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nSad! I'm camping for the next week... but unless the package and author reappears, we'll need to delete it, sadly.\nCould we wait for one or two weeks? It looks like that is fine, but GitHub somehow locks the account.\nYes, I received an email and will update the source to later today.\nI've opened to resolve this issue.\nHi, The flag has been removed from my account a few hours ago, so this change can be reverted."}
{"_id":"q-en-nixpkgs-51d79b6906888e8be9beb2cddaaa72eba6055f3218ed204b76de18b636a94075","text":"crackmapexec fails to run on the unstable branch. Full backtrace : Steps to reproduce the behavior: Should be marked as broken for python 3.12 and above due to not being maintained anymore, it should be recommended to use the maintained fork . If applicable, add screenshots to help explain your problem. I may PR a proper fix later if I find time to do so <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI don't know if removal was the best choice for this tool but maybe has all the features that had so it's alright. It's just that I'm not aware if there are any bugs that are not in that are in but I understand why removal makes sense since later we'll move onto later version of Python such as 3.13 and further more\nAs is a fork, I guess that the feature set is comparable or even bigger in the meantime. is still available for 24.05 and could still be used after 24.11 with some magic.\nI ran into this issue where didn't have it, it is just a minor issue so it's alright on my side, but not sure if there will be any more issues I'll ran into later :"}
{"_id":"q-en-nixpkgs-2d9a6f97455d65c0b47e21452f96bccf1c1c2bcd8225da524789a1924cc64049","text":"After that upgraded openssh to 9.8p1, variant is broken on M1 darwin. Run: it results in the following error during build: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAdding to fixes the issue for me, I'm now trying to figure out why it only shows up on darwin-aarch64.\nSeems that 9.7p1-9.8p1 change removed the include directive: -- so it's not clear to me why this succeeds on linux/FreeBSD and fails on darwin.\nI'm now trying to fix this upstream, in freebsd ports, where we pull hpn patches from."}
{"_id":"q-en-nixpkgs-e8746acc119bffb228de31436e80cd255a458a397b017e5daa27d4753b71a357","text":"All expressions in nixpkgs are supposed to be MIT licensed, but is licensed under GPL-3.0. as an externally developed project under different licensing terms that we use for nixpkgs. It was merged into nixpgks in without the license being changed to MIT or a compatible license. Every package using is using improperly licensed expressions, which at the time of writing looks to be a bit more than 25 packages. I have two proposed solutions, neither of which are particularly appealing: -license under MIT This would require assembling relicensing signatures from all existing yarn2nix authors which include both contributors when it was in nix-community & in nixpkgs. It's gonna take some time and keep us in license non-compliance in the mean time. , migrate packages to something else The newly looks like a good contender as pointed out by in chat. CCing authors and other relevant parties Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nJust to repeat what I said on Matrix, if we can’t figure out something better in a few days I will open a relicensing issue and start the work of pinging all the contributors. Reimplementing it sounds nice, but I suspect it would take quite a while given the speed at which Nixpkgs language ecosystems tend to move.\nyarn2nix can be easily misused and trigger IFD. I think too that in the long run we should replace it. Maybe yarnBuildHook/yarnConfigHook from are already enough and the package can easily be migrated?\n(Nixpkgs potentially actually being GPLv3 is very much a bug IMO.) Migration is all well and good, but we probably ought to try to relicense anyway: we can’t change old revisions of Nixpkgs and it’s not good for multiple years of Nixpkgs history to be under unclear licensing conditions.\nI guess re-licensing is the better way for now? We can ping all the contributors, from a glance I read most of the names before.\nYes, I think we should probably at least try. I will open a new GitHub issue for it today so that we can discuss here without it getting bogged down in rote responses.\nOkay, here we go I guess:\nOof, thank you for trying to clean up that mess. I just got the notification from the tracking issue and I feel that a big portion of any blame should go towards my direction because of the introduction in , rather than While I (try to) put a lot more thought into licenses today, it didn't cross my mind back then that vendoring with the included license could have such implications.\nNo worries – thanks for trying to improve the Node packaging situation in Nixpkgs and for your quick reply to the ping! This is an institutional failure IMO, and I think that it would have been caught before merge these days.\nseems still active. It might be worth posting the same as on Discourse if you want people's replies to be public (unlike with email)."}
{"_id":"q-en-nixpkgs-1d51b9323c103eddc70d8dbc6055fe15e22fb29a2ffc6cbba2de445bc631a70a","text":"Package name: kodi Latest released version: 21.1 Current version on the unstable channel: 21.0 Current version on the stable/release channel: 21.0 Notify maintainers Release: Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nthanks for the ping! i am on it in the next day or two"}
{"_id":"q-en-nixpkgs-45c8ffc1c7889b62182023011085d52c74b7db06ec08b028a7bb31a57b358add","text":"Steps to reproduce the behavior: Attempt to build vmware-workstation from the unstable branch using the following command: sudo nixos-rebuild switch --flake ~/nixos#framework --show-trace Observe the build failure with the following error related to the --docdir option in the configure flags. Build log Running phase: updateAutotoolsGnuConfigScriptsPhase Updating Autotools / GNU config script to a newer upstream version: Updating Autotools / GNU config script to a newer upstream version: Running phase: configurePhase fixing libtool script patching script interpreter paths in ./configure ./configure: interpreter directive changed from \"#! /bin/sh\" to \"/nix/store/4bj2kxdm1462fzcc2i2s4dn33g2angcc-bash-5.2p32/bin/sh\" configure flags: --disable-static --prefix=/nix/store/9lclanzcvqmx0jc15i06125phchhsbsp-gdbm-1.8.3 --bindir=/nix/store/9lclanzcvqmx0jc15i06125phchhsbsp-gdbm-1.8.3/bin --sbindir=/nix/store/9lclanzcvqmx0jc15i06125phchhsbsp-gdbm-1.8.3/sbin --includedir=/nix/store/f8yb3rhalwdpaig8lwxd4plgbqn8vagk-gdbm-1.8.3-dev/include --oldincludedir=/nix/store/f8yb3rhalwdpaig8lwxd4plgbqn8vagk-gdbm-1.8.3-dev/include --mandir=/nix/store/64jmdcjmdl94l4hbadwqsz7j8qg16dwb-gdbm-1.8.3-man/share/man --infodir=/nix/store/vs7bglvjl3pxp2ffc98v5w7xlinn0q5j-gdbm-1.8.3-info/share/info --docdir=/nix/store/9lclanzcvqmx0jc15i06125phchhsbsp-gdbm-1.8.3/share/doc/gdbm --libdir=/nix/store/6gzwn3dm5qzbvj5009ind2ihvi84csi2-gdbm-1.8.3-lib/lib --libexecdir=/nix/store/6gzwn3dm5qzbvj5009ind2ihvi84csi2-gdbm-1.8.3-lib/libexec --localedir=/nix/store/6gzwn3dm5qzbvj5009ind2ihvi84csi2-gdbm-1.8.3-lib/share/locale --enable-libgdbm-compat configure: error: unrecognized option: --docdir=/nix/store/9lclanzcvqmx0jc15i06125phchhsbsp-gdbm-1.8.3/share/doc/gdbm Try `./configure --help' for more information. Additional context I tried building the package on NixOS 24.05 with the following command, and the build succeeded without any issues: Notify maintainers cc Metadata [user nix-shell -p nix-info --run \"nix-info -m\" system: host os: multi-user?: sandbox: version: channels(root): nixpkgs:\nPlease update the issue's title with the actual package name ()\ndo we really need specific gdbm version during the build? or as a buildInputs?\nI removed the override from and I can build and run vmware just fine. Could you just update the dependency version to 1.24 and remove the install override?\nHow can I do it myself? Download and edit the nixpkg and edit it?\nSorry, I didn't mean \"you\" as in you specifically, but more generally as a third person pronoun. I don't know how either, never contributed to nixpkgs before . I meant for the question to be at nixpkgs contributors to know if what I proposed makes sense. EDIT: Oh, might have misunderstood you. If you want to do the same thing I did then yes, get a local clone of nixpkgs, edit the file and point your system nixpkgs to the local clone. Here's the patch that worked for me: <details<summaryClick to show patch </summary</detailsNote that it doesn't pin the version of gdbm, it will always use the latest version from nixpkgs which may cause with future versions. I'm assuming there was a reason the original was pinned to 1.8.3.\nThanks it's very informative!"}
{"_id":"q-en-nixpkgs-88238bf3a715fcef1f051db4ef5b2621cd8215212ea74afc2e483cbe4229500b","text":"system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-ab07be9828612af6fd72b534c4b3cc37c6cc7427be57725882e563ba6e5fa1ac","text":"Package name: Latest released version: <!-- Search your package here: --- Current version on the unstable channel: Current version on the stable/release channel: <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-3aac4ceee51838d12ddb5b90b7a0449e198d6172351d0375755e607b8afc5b03","text":"Steps to reproduce the behavior: X Add any other context about the problem here. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nAwesome. Now I have malware with 200M+ size and a libcef within.\nInteresting you actually have a fsverity-enabled build directory, so this is probably an upstream issue that nobody has run into while running the tests. Specifically it looks like this line is detecting something is not behaving as it should: I recommended that you open this issue upstream. You can temporarily fix the build locally by overriding the package and setting to .\nAlso, could you post the rest of the line starting with . It is getting cut off by your pager and seems to contain the actual error from the mount attempt."}
{"_id":"q-en-nixpkgs-64697f2f1cf59071db92e5a692055126657b4467145a8e5eee5d8579ca30d5ed","text":"Keeweb opens an empty window, hangs for a few seconds and then the program simply crashes. Looks like a missing libGL dependency. <details<summaryLog without libGL</summary</detailsSteps to reproduce the behavior: Keeweb from Terminal or .desktop file. Keeweb not crashing. Just an empty window with only an icon and title. Can be started with following Command: <details<summaryUsing this overlay to launch it without having to use the terminal.</summary</details<!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-081062f393c7b992f93a703ace534a6f2f56b69763340b111cdd6e14b79d8416","text":"has an assertion on enableSSL that requires certFile, keyFile, and caFile, but caFile is optional with SSL and only used for client certificate validation. ttyd is able to serve over SSL with valid certificates Steps to reproduce the behavior: builds with the provided options Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI've committed a fixed file to my own flake that fixes the assertion (as well as adding the \"--ssl\" flag, which was missing)\nplease view"}
{"_id":"q-en-nixpkgs-6e118049bea9db2c318d5526372948dadc94db9bb1e3f5c822bebcaada507957","text":"Since Zettlr is an Electron app, it needs to be run with certain command-line arguments on Wayland (specifically ). In NixOS, setting the environment variable to should automatically tell packages to use this command-line argument, but this requires integrating it into the package file (see ) for an example. The Zettlr package does not currently do this, which results in pixelated and glitchy graphics on Wayland unless it is manually run with the correct command-line arguments. This should be a pretty simple fix, as it has been implemented before for many other Electron apps by creating a binary wrapper."}
{"_id":"q-en-nixpkgs-29e4b31badfc940063994b7da06ed856598c55de4a504a89843f751b14fbeadd","text":"I'm trying to use GZDoom, but the Vulkan backend isn't working. No errors, but it will silently fall back to OpenGL. Apparently this is the case if GZDoom can't find the correct Vulkan drivers. I have a Radeon RX 5700 XT, so my GPU definitely supports Vulkan. Steps to reproduce the behavior: basic without any mods <kbdesc</kbdthen <kbd~</kbdAt the end of its loading it should say what renderer it is using This is a screenshot of it running with Vulkan from someone else's system: ! This is a screenshot of what it looks like when it's using OpenGL: ! Note the lines. GZDoom uses Vulkan Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-97f11581608cead52171927bf9db97cf8245a23270f0cc15fd1a6900b8af3c93","text":"I noticed that nvim wasn't loading any of my treesitter plugins. I decided to check the plugin directory. hm. something is very wrong here. it isn't pulling in any grammars. let's investigate: ok, so it's looking for a file at . what does the grammar derivation look like? So the parser is ending up at instead of . What does the grammar derivation's source look like? Checking , this code has been in place for years. Checking for , I find It was merged three weeks ago. I see that is open. I couldn't find any information on this problem with GitHub search... Could that PR be merged quicker please? My editor setup is broken now :( system: host os: multi-user?: sandbox: version: nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nI think this is not actually the direct cause of the issue you're seeing since the grammars go through (see the below, which precisely corrects the observed discrepancy) You're not seeing the grammars when you because the grammars are in . I explain how to properly build it in The PR is broken, don't get me wrong, but I think not this directly :p. For why your is broken, see I think it's because of your dependency on and which pulls in a without any grammars but I'm not sure. You can try giving an approval, I'm in the same boat as you (I don't have merge permissions). Apologies for breaking your setup!"}
{"_id":"q-en-nixpkgs-a681b572d12515be29815322e1d7e4e730fa01f20309250bf4fd87999970bcfe","text":"package lists and as dependencies but upstream uses instead. Steps to reproduce the behavior: dependencies of package it to the dependencies listed by upstream and should not be in the dependencies list, but should n/a <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nThat's a bot.\nThank you for the quick turnaround!"}
{"_id":"q-en-nixpkgs-ca25236486f06f68e2b96d78929b7de9d401579f876ecc35ad26c7ad80571a21","text":"A small difference in the dependencies used by aider and nixpkgs causes aider to crash. aider 0.56.0 something should get a stack overflow See <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nwas closed by mistake so opened this\nThanks for opening this issue. Discussion started here It is clear that v1.44.22 of is what causes the issue but I'm not sure what could be done. maybe you have an idea?\nThere is an open issue in the repo, we have to wait for the dependencies to be updated upstream unfortunately. On the next update of this dependency ill be sure to try to run aider. Sorry for the inconvenience !\nWhy don't you use ?\nthe issue on aider's repo was closed the fix is to not use the wrong version of litellm :sweat:\nit's not so simple for python packages. each one is a forest of carefully balanced dependencies.\nLatest aider update matches with the latest litellm update, so we should be good in a couple of days"}
{"_id":"q-en-nixpkgs-909c6b535ce39b0c3229510c4c2e383aeda3fafef0cec5bf22a55ef6f29c5eb0","text":"Package name:crow-translate Latest released version:3.0.0 <!-- Search your package here: --- Current version on the unstable channel:2.11.1 Current version on the stable/release channel:2.11.1 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-3cced79923da271c7aaf53f223739cc9517aedf0a6bb71ac4a48efc96c5b3506","text":"Package name: rio Latest released version: 0.1.13 <!-- Search your package here: --- Current version on the unstable channel: 0.1.13 Current version on the stable/release channel: 0.0.39 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --Note for maintainers: Please tag this issue in your PR. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-eb17c19de12b3a54b024abbdcee0e28a1298dbfbeb430b0360084ae0bcec8ccd","text":"The final merged version of isn't building. Just build with the initial configuration that would give on a UEFI machine, along with: Build successfully and boot into Dom0 on next reboot. <img width=\"1201\" alt=\"image\" src=\"\"During the progress of , the issue isn't happening as of . Might be a regression happened in the last few commits before it's merged. system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nWhoops. That's my bad. Will make a PR in a few."}
{"_id":"q-en-nixpkgs-bf93bee584c302a717bc81f9efd5650c71aee4cb57a0d01cb56676001a226f14","text":"Package name: paperless-ngx Latest released version: 2.12.1 <!-- Search your package here: --- Current version on the unstable channel: 2.11.6 <!-- Type the name of your package and try to find an open pull request for the package If you find an open pull request, you can review it! There's a high chance that you'll have the new version right away while helping the community! --- Notify maintainers <!-- If the result shows no maintainers, tag the person that last updated the package. --"}
{"_id":"q-en-nixpkgs-36edc31c84d43614c40af7e7a0482d525bd50752364a7c707321fb5b4af7233e","text":"No desktop file is installed when installing httpie-desktop. Steps to reproduce the behavior: httpie-desktop. find it in the application list of GNOME. Find it in the application list of GNOME. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --- system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:"}
{"_id":"q-en-nixpkgs-729f5d08942abed31012c74af2cc0a468a13623eabe01b3d188d8139b6a8340b","text":"python27Packages.faulthandler fails to build on the unstable channel Cc () The output shows test failures due to regular expressions not matching. E.g. It seems that the tests are expecting to see in the message, while the actual error says . If it's just that, I hope that it would be easy to fix the build. I am happy to help if given some guidance. Full output: host os: multi-user?: sandbox: version: channels(root): channels(user): nixpkgs:\nThis is an upstream issue. See similar issue Best write them that they should also accept , , etc. Then you can open a PR here that adds a patch to the derivation.\nThanks, will look into that. It probably makes sense to keep this issue open to track the upstream one though.\nNote that a patch is available ( ), it may make sense to apply it (e.g. done for Fedora ) while waiting for a new release. I am still happy to try and do that.\nI don't think that patch will work. It accepts but not .\nSorry, I missed that :(\nI opened an upstream PR: And I applied the patch in Nixpkgs:\nfyi has now been merged and a release is pending (see )... hopefully we won't need the patch after that. Thanks everyone for your work!\nI removed the patches in ."}
{"_id":"q-en-nixpkgs-8e1deabc04a41b7429a65a8e8a0eaca30ea7954c9f0cfe39ea04c2264c58bdcb","text":"Steps to reproduce the behavior: Please run and paste the result. Add a :+1: [reaction] to [issues you find important]. [reaction]: [issues you find important]:\nHydra can reproduce:\nI tracked down the potential source of this failure, and am working on a fix. These builds take 30 minutes for me though, so it's slow going.\nPlease see for a fix. Testing appreciated.\nThanks for the quick patch. EDIT: My way of installing to profile didn't pull the fix, I'll see if I can figure out a way to do it properly using the pr review package\nBuilds and runs... seg faults as soon as I load a model, but that's an AMD driver thing (whole WM crashes)... Ship it :laughing:\nI did see some segfaults, but was able to restart and get it to load on my 6700xt. “Glad” to know it’s known and common.\nNice! Thanks for this - excited to see it roll into unstable soon. Ollama seems to always have some issue - is there a \"working group\" that has some focus on this program and nix? Anyway, thanks again!\nThe problems with crashing aren't nix problems, they're crappy amd driver issues. It doesn't crash if you don't load models close to the max vram available... obviously it shouldn't crash if you do, but amd... so yeah"}
{"_id":"q-en-nixpkgs-083bbf928c01f3cef04be34ed106c6c3b21b127cb62dda038482ceafde191b09","text":"It would be nice to see packaged such that at least an executable is exposed.\nHow about doing it yourself? Have a look at the and at the many examples in If you have questions, just ask them in this thread."}
{"_id":"q-en-nixpkgs-fee0cc4f857355377bf89557a09ac863145731f8bf9d1917057440fbcf772df8","text":"fails with the following error since #diff-: on master The systemd build system changed to meson but the systemd-cryptsetup-generator was not updated."}
{"_id":"q-en-nixpkgs-9f153ea408db8008ca65a272f43afc230171fda8d73318f6a195a3d8ef895602","text":"Neomutt's bundled tool is unable to find . We should probably replace that string in the perl script. system: host os: multi-user?: sandbox: version: channels(root): nixpkgs:"}
{"_id":"q-en-nixpkgs-1e66a6f3f9aeba75200e2bf0f9967ff8eae362f02d14be441be8d3d67dc8244a","text":"I have a Samsung M2022W (usb) that works on a raspberrypi using raspbian (2017-11-29-raspbian-stretch-). I had to pick the ML-2160 driver in CUPS 2.2.1 as recommended by On NixOS (cups-2.2.2): [nawal nix-info system: \"x8664-linux\", multi-user?: yes, version: nix-env (Nix) 1.11.16, channels(nawal): \"nixos-17.03-17.03.1082.\", channels(root): \"nixos-17.09.3149.\", nixpkgs: /nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs If I use services.printing = { enable = true; drivers = [ pkgs.samsungUnifiedLinuxDriver ]; extraConf = '' LogLevel debug ''; extraFilesConf = '' FileDevice yes ''; }; and pick the same driver, I get garbage printed out. I tried to track down the ppd on raspbian, but only found: pi $ grep 2160 /usr/share/cups/drv/splix- ModelName \"ML-2160\" PCFileName \"\" and (i.e. it generated the ppd when I installed the printer apparently). On nixos: [nawal | grep 2160 /nix/store/6f2zx8dl4w0kzgw3wj8f466mvkcgc3dc-samsung-UnifiedLinuxDriver-4.00.39/opt/share/ppd/ML- /nix/store/nmp023llc8fix567f413a5rqbx4sw45g-cups-progs/share/cups/model/ML- Having diffed them, I notice amongst a huge number of differences: [nawal diff /nix/store/nmp023llc8fix567f413a5rqbx4sw45g-cups-progs/share/cups/model/ML- | grep Version < FileVersion: \"2.0.0\" < PSVersion: \"(3010.000) 0\" < QPDL QPDLVersion: \"3\" < *cupsVersion: 2.2 so raspbian is using a version \"2.0.0\" and \"2.2\" of something or other, whereas nixos is using a version \"1.2\" and \"1.1\" of similar things. I managed to get my printer to work by copying the ppd across, but I don't understand enough of cups (nor now have the inclination to learn) to figure out what's going on or how to make nixos do the right thing (i.e. in a word, print). I thought this might be useful for someone else who wants to pick it up? Happy to hack around the raspbian if you need more info, or try out things if you need me to while the sd card contains raspbian.\nCan you find out what package on Raspbian provides the driver? Have you installed proprietary Samsung driver there? If not, can you try removing ? It seems we have latest version of but not the latest of the proprietary driver but you don't seem to use it.\nI'll have a look tomorrow evening - how do i use the new splix without upgrading everything to the same commit?\nFind out your current commit with (look at the suffix), then clone nixpkgs repository (), checkout current commit (), fetch my changes () and cherry-pick my patch atop your current commit (). Then you can rebuild nixpkgs with .\nI can confirm this works. Thanks for the detailed instructions and for fixing it so quickly! NB: I had to do: (i.e. ) thanks"}
{"_id":"q-en-nixpkgs-82c962aa1b35f0050d175917217c79d744aabf6cf3b2c93ca46e76c009ed4ce5","text":"When the user of the SSH connection used by the SSH substituter does not have a shell configured it defaults to , which causes the following error: Which actually is the decimal representation of the first few letters of the following error reported by (thanks for figuring that out). Since with the release of Nix 2.0 it has become easy to setup a SSH substituter, so I expect that more people might run into this problem. I know of at least another that reported it on the IRC. Instead of the serialized integer error it should just have reported the actual error, or at least it should have given a more specific error about the response being bogus. Forget to set for the user of the SSH substituter connection. system: host os: multi-user?: sandbox: version: (from nixos-unstable around 5 Mar) channels(root): channels(matthijs): nixpkgs: (nixos-17.09 around 22 Feb)\nI some debug echo messages to my profile scripts and I got this error as well (with a different number). To fix it, I had to remove any output from the profiles ...\nJust stumbled into this too. I figured should be enough on the server part, but alas.\nThank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as \"still important to me\". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly: for maintainers and people that previously touched the related code and @ mention them in a comment. on the . 3. Ask on the on .\nI'm 90% sure this issue is still valid.\nIt is.\nIt is.\nI marked this as stale due to inactivity. &rarr;\nProbably still\nMade to show a better error message. But I feel that should print its message to stderr instead of stdout...\nI just run into this error running\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:"}
{"_id":"q-en-nixpkgs-25f5b4279c1d0be89ea81b444183047595aa83ba2bc2caa0208dddccf8d124f3","text":"It is an intermittent failure, see , not much we can do about it other than restarting the build. I opened an upstream issue\nI've now tried doing a third build locally for this: all have failed. That doesn't seem intermittent to me.\nIt is succeeding for me most of the time, hydra also managed to build it in the past.\nMaybe try setting ? Not sure if it will help but it has helped with some derivations in the past.\nmanaged to reproduce it even with parallel building disabled:\nInterestingly lowering parallelism is what seems to trigger the issue for me. If I build with all cores (8) it seems to get past that part of the build all the time, I'm yet to see it succeed when building with a single core.\nCan confirm that low parallelism is the problem. is building successfully, which is the first time it has worked for me. That's consistent with what I've seen, not having it build successfully a single time (and it has always only used a single core).\nI also got the issue when doing a nixos-rebuild on my 2-core laptop, but it succeeded on the same laptop when using nix-build, which means that all cores are available for just webkitgtk. This is the error message I got:\nWould it make sense to add the stopgap fix suggested in So something like this: diff diff --git index .. --- +++ -60,6 +60,12 stdenv.mkDerivation rec { \"-DENABLEGTKDOC=OFF\" ]; postConfigure = '' # A stopgap for a non-deterministic build failure when using only one core # Upstream bug: ninja JavaScriptCoreForwardingHeaders WTFForwardingHeaders ''; NIXCFLAGS_COMPILE = optionalString stdenv.isDarwin \" -lintl\";\nOpen the file Add the following in line 258: add_dependencies(LLIntOffsetsExtractor JavaScriptCoreForwardingHeaders)\nThis will be no longer neccesary with"}
{"_id":"q-en-nixpkgs-ab2ea93272487b292f614a2b6235fb7d5c977bf72d9b5bcb2e1540ceee96fc90","text":"Installing Nixos from the 18.03 minimal installation DVD iso, for a fresh 64-bit Hyper-V vm target system, is failing for me at : Running under strace, there is a possible match at but this was from a quick grep from a lot of content, so might well be a red herring. That path is present in the host (CD-booted) system. Full trace . I tried to follow the vanilla NixOS installation instructions as closely as possible. [This is the installation CD system] system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: In case relevant, the hardware- contents are:\nHaving the same issue right now.\nGetting the same using for a fresh installation attempt. Identical installation steps with succeeds.\nConfirmed that also succeeds in my case. :heavycheckmark: :x: :x:\nI'm having the same issue on an Acer ES1-132\nYep, here too on an XPS 13 9370\nI had the same issue. My workaround was to temporarily set after which the install would succeed but not write the boot-listing efivar. Then I manually created that efivar with efibootmgr. I imagine the command would have worked from the installation media, but I'm not sure since I ran it in Debian. (on my system the EFI boot partition is /dev/sdb9) There were also *.efi files generated in efinixos, but I didn't try any since it booted with this.\nI saw this as well, on an AsRock Z77 Extreme4. LSLeary's workaround was fine for me since I prefer to boot via rEFInd anyway. I was able to manually touch or cp another boot entry to the destination (per strace) without any error, so it doesn't seem that the firmware is preventing the creation of the variable.\nis anyone actively working on this? what does severity blocker signify?\nI know this ticket is for a fresh install, but I installed 17.09 and upgrade 18.03 and it worked fine in case anyone needs a workaround.\nAlso seeing this when installing on vmware, but with custom install media built from the nixos-18.03channel as of today (). No problems with 17.09. EFI is certainly enabled on the host.\nI run into the exact same problem while trying to install NixOS 18.03 into VMware.\nI got the same behaviour when I tried to install NixOS 18.03 on Dell XPS 13 (9360).\nSame issue when tying a fresh install of nixos-minimal-18.03..-x86_64- on VirtualBox. \"EFI\" option in virtual machine system settings is enabled. I followed the install instructions as closely as possible.\nCannot perform a clean install on Lenovo X1 Yoga 2018.\nI got the same issue when installing NixOS-18.03 () on a recent Intel NUC.\nA fix has been pushed to the master and 18.03 branches. It will take some time to reach the channel, and for a new ISO to be generated using the fix, but this specific issue should be fixed soon for new installs. In the meantime, should work. Once installed, the option can be put back to the defaults (removed from your configuration) as the issue only affects the installation environment.\nFor the record, I was unable to boot using the aforelinked efibootmgr workaround on a 2018 Lenovo X1 Yoga. I had a pre-existing Linux boot entry, which I removed, but the new entry that I did not appear as a startup option.\nAlternatively,  can be used with grub, This will put the bootloader in the default fallback location. While it is documented as \"hardcoded to try first\", I'm not sure this is right. It will, though, try to boot that specific location when no other option is possible. As a final alternative, adding at in the ESP (the fallback location) can help; rEFInd will scan for existing EFI programs on the ESP partitions, which would allow selecting systemd-boot. Though, all of this assumes that the installation process finished. EDIT: For anyone reading this in the (hopefully) few hours left until the channel bumps and the iso updates on the website This build is the iso built in the same evaluation that *\nSorry to drag this from the dead but the website has not updated to the new Isos. the build mentioned by samueldr is complete and the iso there solved this issue on hyper-v for me.\nConfirmed (what Wolfereign said) - thanks!\nI think this issue was automatically closed by the github merge bot, but shouldn't have been.\n4 days ago, it shouldn't have been closed; github doesn't know about nixos channels. The current iso image on is made from commit , which is after , and . Simply put, the 18.03 iso do contain the needed fix for this specific issue. The iso image on the website was updated somewhere between ~24h ago (when the channel advanced) and a couple hours ago when I checked whether the iso image changed. I have also, just in case, verified that it does not reproduce the fixed issue. I'm using this post to add this note:\nIf I look correctly, the ISO on web has been for over thirty hours now, and that commit does contain . I wonder... can you be more specific? What do you mean by \"website has not updated\"? EDIT: oh, github didn't show me the last posts when I was writing this :-)\nThe website is updated now. When I checked at the time of my post the isos on the website (official download links) had not been updated.\nEdit: Ignore me. We did something silly with efibootmgr to wipe EFI entries during install, which weirdly enough triggers exactly this error."}
{"_id":"q-en-nixpkgs-84618bea72304b7fd61a77f22a3d3285fcaf78839773528544b5911dd9dd9e09","text":"Enchant 2 (now the default) is long supported by most packages. The following packages need to be updated or removed: [x] gtkspell2 (dead): [x] xiphos: [x] subtitleeditor: [x] xneur: (contains the removal of enchant1) [x] gnome2.gtkhtml (dead): [x] gnome2.gtkhtml4 (dead): [x] gnome.gtkhtml (dead?): [x] php.extensions.enchant:\nThank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as \"still important to me\". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly: for maintainers and people that previously touched the related code and @ mention them in a comment. on the . on the on .\nThis issue is still relevant.\nthe list from is not up-to-date. e.g. is already fixed maybe we can also close it, because defaults to version 2 and to version 1.\nI marked this as stale due to inactivity. &rarr;\nUpdated the list.\nNow with xiphos merged You can thick both xiphos and gnome2.gtkhtml4 off the checklist\nPushed the enchant1 removal commit to"}
{"_id":"q-en-nixpkgs-2316ce9eb4a61d1843fb41246adca078575ccbbd822605b385fc3343f9879113","text":"Qemu 2.12.0 has released today. We need to update package. Please run and paste the results."}
{"_id":"q-en-nixpkgs-ad635f95b2a428e0d8ae763a32c70852ed2c48a8e205423ec97b50d37e683c1e","text":"From a recent checkout, where the version of procps is advanced to 3.3.14, happens also for unaltered master. and under nix-shell even weirder with the thing. Just issue the commands above. Doesn't happen with release-18.03, so probably some issue with a more recent autotools version.\nLooks good to me. I should have only changed the homepage in the meta tags to begin with. I was of the opinion, it should point somewhere, where one also can report bugs to procps, was a bit too eager then to get rid of sf. : Please take over maintainership for the package."}
{"_id":"q-en-nixpkgs-008dd8342f2ce3034c95c67ef9819544aa5d9362d9c9576943a90b9399d6bdd8","text":"I am working on a Haskell project that uses . Compiling the project fails with: From my limited understanding of Nix, header files are stored in and is not there.\nThe default does not include the development headers. Try running from a nix-shell:\nThanks for the response. I tried and then . This resulted in the same error as before. Could you please explain why we'd expect using to work in this case? Is there a simple modification that can be made to the derivation so that it includes the development headers or would creating an entirely new derivation, perhaps named , be a better option?\nOk sorry forgot, that you probably need to do this: Which is working for me at least. Alternatively, you should be able to call stack with the --nix arg. (Sadly it does not work currently with GHCJS):\nI have actually gone down the + route. I created a file: and to my file. This resulted in the same error as above. This is why I've gone back now to just using to see if I could get it to work in this more simple setting and then try later on. Just now I tried your suggestion and this also resulted in the same error.\nIt must be something weird going on with Stack. You definitely shouldn't have to manually install gpgme. Anyway, bindings-gpgme is working for me:\nThe above worked for me. It finished by printing out . It did also say ; now sure what that means. Any ideas on why this works but doesn't when I just install by itself? I'm guessing somewhere in the Nix machinery it knows depends on and so it installs it in a specialized way? At the moment I'd like to avoid having Nix manage all the haskell packages. My goals is to slowly move the organization I work at towards Nix. I think a good first step would be to provide developers with a reproducible development environment that contains all the non-haskell deps.\nI was able to get the integration to work with the following: Turns out what I needed was which is explained in the documentation . I still wish there was a way to get this to work without having to rely on a feature.\nThat's very weird! We should be doing this for you in buildStackProject: I get this with :\nOk- I might have found the problem. Try this , You need libgpgerror for the gpgme header to work properly. Update: It looks like we are not properly propagating inputs to generic-stack- is not being pulled into despite being necessary for gpgme.h to work (and a propagated input). /cc\nI've never looked at . it might be screwed up in arbitrary ways.\nIs there any kind of tool to pull in the deps that Nixpkgs would with buildInputs? I think does this but it would be nice to avoid having a script. Something like this: where findAllInputs includes propagated dependencies.\nYeah mutates all these globals cause bash so it's very inconvenient to call more than once. Maybe try dredging up and updating it work as does today :D. This won't be quite the same as there's a few cases of packages hand-modiying the propagated input files .\nbut a far easier solution is just to loop over in bash (or add an env hook to let the looping be done for you).\nHmm... I guess we will need to put it in shellHook too for it to work with . The problem with the env hook is that gpgme shouldn't have to know Stack exists at all. It might make more sense to just replace NIXLDFLAGS & NIXCFLAGS_COMPILE directly to emulate cc-wrapper behavior (which stack should be calling anyway right?) Has anyone looked into rewriting all headers to give absolute path names for calls? That would solve most of these build cases I think & Nix would treat them like it does dynamic libraries. Probably a huge change but could be worth doing & avoids some of the propagated cases where it really just needs the header file (not all though).\nAhh but can add the env hook, and everything will work out.\nThat would be really cool! But CPP sadly makes these things quite hard. Because (CPP) identifiers are ambiguous, one would have to bail out on the first unresolved as further name resolution is ambiguous.\nI to the and removed the line. This worked. Thank you. However, if I use in place of or if I work outside of a , installing and with , my haskell project does not compile and results in the same error that I initially posted. Is this to be expected? If so, can you please explain why?\nAwesome! So when you run it will symlink things into . By default, Nix will set some variables like PATH & MANPATH to pick up things in $, but we don't intend to set up too much besides these \"standard\" variables. Instead, we prefer to do things inside (which I think is also what uses) that are closer to how Nixpkgs are built. Here is maybe what you wanted to do originally (which still works but gets ugly with lots of dependencies): Hopefully we can fix the issue with propagation though.\nThanks for the explanation. So it seems my idea of using to build a general purpose development environment doesn't really align with what the tool was intended to be used for? And just to confirm, the \"issue with propagation\" relates only to ?\nIs this what you were thinking? I am not able to get the setupHook route to be run correctly in nix-shell so the addEnvHooks doesn't work."}
{"_id":"q-en-nixpkgs-4241b5b65500214d0b1007527d9de2f43c43e13d062c61347f8c9e33c3a0a6c7","text":"states that it prints out the stripped name to the , but the stores the value in instead. The functionality in the documentation seems more convenient is there a reason not to switch to it?\nI've noticed this as well. +1 for changing the implementation to print to stdout.\nThis one is interesting. There are maybe two dozen uses of that function, but it would be nice to change it. Maybe I’ll look into it tomorrow. Always happy to trigger a mass rebuild. :P\nWas fixed in the meantime."}
{"_id":"q-en-nixpkgs-59810d0a40e96f50f4a86d30ef1b6ed8eff812f4126ff717c70551f46056c70c","text":"As it stands, the NixOS MPD module runs MPD as a systemd system service. As such it expects that the MPD user account (mpd by default) has control over the audio device. This conflicts with PulseAudio when it's used as a systemd user service (the default configuration). The intent of this change is to extend the MPD module so that it can be configured to run as a systemd user service. The changes remain backwards-compatible so users can continue with a system-wide configuration or go with a per-user config. I have a proof-of-concept of the modified module which I'm currently experimenting with. It can be used in as follows: The result is MPD running within a user slice: [ ] Ability to use a different per user; Probably by providing a option which overrides the other options and simply points to a file in the users home directory (ex. ). [ ] See if it's possible to get realtime scheduling to work. [x] Add NixOS tests. Currently there aren't any. The different options make for numerous potential combinations; I'm to lazy to test all of them manually.\ndid you ever add support for running mpd as a user service? It looks like just tests in preparation for running it as a user service.\nWow, I forgot all about this. Perhaps because I soon after stopped using MDP. The implementation is still on Github:\nAn alternative could be to let it talk to pulseaudio over the network. In the past I have used a config in mpd like audio_output { type \"pulse\" name \"pulseaudio\" server \"127.0.0.1\" } and then some config in to allow this access."}
{"_id":"q-en-nixpkgs-bd0edb3203b13f432d5f8507f064f4287a5d8149a3d247dfa9f4299794188426","text":"MusicBrainz-based metadata collection does not work in . The problem is that the Perl modules cannot be found. Full error message: This should be an easy fix for anyone who knows about Perl in nixpkgs. 1) Install globally (not in ). 2) Make sure includes . 3) Insert a CD. 4) Run .\nThat dependency seems to be included already. I don't have a CD drive. Is there an other way to reproduce?\nI am experiencing this problem too. Even without a CD drive, you should be able to reproduce simply by running . This is an auxiliary program invoked by , a wrapped Perl script that will immediately attempt to import the module in question: and fail as described. I have been able to get partway there by manually prefixing with a PERL5LIB environment variable: That invocation dies with and then presumably I can go find that and add it to the manual PERL5LIB, but now I'm chasing all the dependencies manually. There's no though, right? I don't know how to refer to a Perl that has all required dependencies.\nBackported to release-18.09: .\nCan confirm that this is working. :+1:\nIt's broken anyway. I just tried and it install a huge bunch of stuff but ended up with: BFAIST/WebService-MusicBrainz- Has already been unwrapped into directory Configuring B/BF/BFAIST/WebService-MusicBrainz- with Build.PL Created and Creating new 'Build' script for 'WebService-MusicBrainz' version '0.94' BFAIST/WebService-MusicBrainz- /usr/bin/perl Build.PL --installdirs site -- OK Running Build for B/BF/BFAIST/WebService-MusicBrainz- Building WebService-MusicBrainz BFAIST/WebService-MusicBrainz- ./Build -- OK Running Build test t/Artist.t ........ 1/? URL () Request Failed - Code: 410 Error: Gone # Tests were run but no plan was declared and donetesting() was not seen. # Looks like your test exited with 255 just after 3. t/Artist.t ........ Dubious, test returned 255 (wstat , 0xff00) All 3 subtests passed t/Label.t ......... ok t/Release.t ....... 1/? URL () Request Failed - Code: 410 Error: Gone # Tests were run but no plan was declared and donetesting() was not seen. # Looks like your test exited with 255 just after 3. t/Release.t ....... Dubious, test returned 255 (wstat , 0xff00) All 3 subtests passed t/ReleaseGroup.t .. ok t/Track.t ......... ok Test Summary Report t/Artist.t (Wstat: Tests: 3 Failed: 0) Non-zero exit status: 255 Parse errors: No plan found in TAP output t/Release.t (Wstat: Tests: 3 Failed: 0) Non-zero exit status: 255 Parse errors: No plan found in TAP output Files=5, Tests=15, 7 wallclock secs ( 0.03 usr 0.01 sys + 0.52 cusr 0.07 csys = 0.63 CPU) Result: FAIL Failed 2/5 test programs. 0/15 subtests failed. BFAIST/WebService-MusicBrainz- ./Build test -- NOT OK //hint// to see the cpan-testers results for installing this module, try: reports BFAIST/WebService-MusicBrainz- $ I'm not a Perlie, so I have no idea how to report this or where to fix it."}
{"_id":"q-en-nixpkgs-39b045de1668fc2b87788cd68d21a9d0d135e54051536ddfd48e691c51e448d5","text":"This configuration: If you poke around in the configuration I linked, you should see that I disable NetworkManager and use a static configuration for my main NIC. I can't think of anything special in my config or with this machine that would affect my usage of Wireguard, but I'm also new to using Wireguard. I'm currently on nixpkgs with a few of my own commits on top of . Please run and paste the results.\nIt doesn't seem happy when already exists... I have done two things, both of which were necessary, to resolve this: the link manually () sure the path to the private key exists (if not, you get a different failure in the journal, with failing). (It's possible that fixing the private key path and rebooting would have been sufficient). I'm leaving this open for two reasons: user should receive a better error message in both cases. that we're in NixOS, we should assume full control of the interface and just re-configure it, instead of failing in the setup script.\nI ran into this issue after trying to debug my wireguard setup. My setup is essentialy identical and I get the same error message. Running the start script that is causing the error results in this output. The offending 2 lines are .\nI had to go one step further than deleting the interface. I needed to stop the via then redeploy configuration. But this solved it for me. Nixos should handle this much more gracefully.\nI had to remove the routes with etc.\nThe command is . You can build it in your conf as pre-setup command, but it has the nasty side effect that it possible interferes with other connections.\nI am still running into this issue."}
{"_id":"q-en-nixpkgs-346cd31fd1ee2eaf6db00b447eaab18988c2012b0a009c337c77138bce7d9ef7","text":"Darktable can make use of OpenCL, but the nix package apparently isn't set up properly to allow this. Run : system: host os: multi-user?: sandbox: version: channels(root): channels(lukas): nixpkgs: OpenCL is set up properly on my system: cc darktable maintainers:\nI'd been meaning to fix this for myself for awhile. Should be good now, pending review.\nThanks!\nSorry to revive this, but I am still having this problem in dartable 2.4.4 using intel OpenCL Neo driver. Should ti be already solved? Or the fix is still not present in 2.4.4? Thanks\nDisregard (at least partially) my last comment. I was using the snap version in Ubuntu 18.10. Using the deb, OpenCl works. Where should I submit the bug for the snap version?"}
{"_id":"q-en-nixpkgs-56522ab9ca7efd430594b2c9f91b7063161a26bce7234d1fa46014814d6ed4af","text":"! See the . This is to discuss Darwin-specific issues for ZHF. (or more)"}
{"_id":"q-en-nixpkgs-ea24ca5e644880f185b862b0b87415e766acb4a93b3a2260f88940ad8c9e1523","text":"I fail to compile graph-tool on nixos-unstable. The error message is a template one, thus cc\nThis is caused by\nFixed upstream:\nthanks !"}
{"_id":"q-en-nixpkgs-635c3c247502f50bd7e1ebde54e2e682ff5c72948654f1a86fef996ba69c4193","text":"Minor cosmetic issue: when using kpcli, the prompt shown to the user says \".kpcli-wrapped:/\" instead of the usual \"kpcli:/\". system: host os: multi-user?: sandbox: version: channels(delroth): channels(root): nixpkgs:\nUsually this can be fixed with the --argv0 flag:\nkpcli uses wrapProgram which already sets --argv0 correctly AFAICT. Looks like this is a kpcli specific problem:\nOK, I understand the reason why --argv0 did not work now. Since this is an interpreted script, we are indeed setting Perl's argv[0] to the right value, but within Perl 0) - Sent to just wrap the script manually and thus control the basename of the wrapped script.\nFixed in"}
{"_id":"q-en-nixpkgs-f40aa8c1e782b4601b1de30951df108aa0ff8658eb10b2e4e2f739e0abbf8275","text":"Ulauncher is an application launcher for Linux: Tried to package it myself but I am a rookie and this seems like a bit painful for the first package.\nFor those interested I made a really quick. However, if you build and execute it you'll realize there's definitely some problems.\nMore specifically, the first error I run into halts with the following stack trace: Any ideas for workarounds?\nIf you just examine the result of the derivation you'll notice that things in appear to be in the wrong locations. So that'll definitely make it not work. There's probably something wrong with and I don't think is maintained. So perhaps we can just change the build system if it's easy? Maybe do what  is doing.\nLooking at the beginning of the build log, I noticed: Note that it thinks the version is . Maybe this has to do with ? I assume that is supposed to be replaced with something else at But clearly something is not working. That's all I've got. Cheers\nVery late edit, I managed to change this after the build via: But last executable still cries:\nWould you mind sharing your full nix expression? I can't figure out how to get past the issue. (I suspect you replaced it in ?)\nYes, but I did it after the build. the last status (which is a wreck with too many unnecessary stuff probably, I gave up after trying a bit).\nI'm not sure how to implement this in Nix, but we can get past the error by setting the environment variable to the generated site-packages folder (). In other words: We still get an error but I think this is a step forward?\n:) I guess. I am still rookie on nixland and this made me confused since we have defined under everything like crazy.\nJust so we are on the same page, this is what I am currently using to build ulauncher: I replaced the version substitution code to put the version number where I think the ulauncher source code intends for it to be put. Doing so does not break the build.\nThe means that the is missing. I'm working on creating a . I've run into , meaning that I needed to make to load a few libraries with using absolute paths. I'm using to apply the patches: Applying the patches is the only way I was able to manage getting to build. The patch was applied with during build time. During runtime, however, the patch was applied, but without substitutions: How do we get the patch substitutions working for runtime?\nmodule is provided by which we already have packaged. definitely should not be used."}
{"_id":"q-en-nixpkgs-b6d139966058c5b688ae5ec45b0197d126ed7e49d1d87b24514b3c8a58b60200","text":"I currently cannot execute any 32 bit Linux binaries on my 64 bit system and this breaks updates, since I have 32 bit support enabled in the nix options: I remember that this worked about a week ago, so something has changed. I am new to NixOS, so its possible that I have done something wrong. I used ArchLinux for about 9 years before that and now NixOS for just 3 months and its concepts are still foreign for me. As a long term linux user (about 15 years) I do feel a bit out of place in NixOS. So I am not sure how to fix that myself. Not sure. I have wine installed and support for 32 bit games enabled. Then I normally do just 'nixos-rebuild switch --upgrade' to upgrade, and sometimes 'nix-channel --update'. I also regularly do 'nix-collect-garbage -d'. No idea if that is reproducible or not.\nThis might be from Could you see if it works without that commit? Also verify that /nix/store/4qj9vzggfzsc34ci79ckaim5cjrwc6ns-bash-4.4-p23/bin/bash executes. I can run it on my system. You can run this to verify that your store is not corrupted:\nI solved it now for me. Instead of using the 'linuxPackagestestinghardened' 4.20 linux kernel I switched back to 'linuxPackages419' 4.19.2 kernel and now executing 32 binaries work again. So I guess the hardened kernel does not support executing 32 bit binaries. Does it make sense to have some kind of visible warning here? I normally would not have used the hardened kernel, if that wouldn't have been the only 4.20 version is I get. Its not clear to me when and how those changes in master appear in my nixpkgs/nixos channel (I am on unstable). I would like to use the master branch directly, but I haven't figured out how to best do that yet.\nThe same here, hardened kernel 4.18.17 and build failed. building '/nix/store/6h9asn7rc93dbwfcs6xaj3k0qgxmg0vh-libtxcdxtn-'... while setting up the build environment: executing '/nix/store/xgf4vra77qrww0n228hd80jsdccdp54w-bash-4.4-p23/bin/bash': Exec format error builder for '/nix/store/6h9asn7rc93dbwfcs6xaj3k0qgxmg0vh-libtxcdxtn-' failed with exit code 1 building '/nix/store/cfwj5yjby2xvagyj69zd1yx8hdhi27ay-'... cannot build derivation '/nix/store/maynq4fxxjc217figygdrgg5q9xnw6yy-mesa-drivers+txc-': 1 dependencies couldn't be built cannot build derivation '/nix/store/wv9b4ayvgnpkvwz81zkk0lm9w4v5pd1y-opengl-drivers-': 1 dependencies couldn't be built cannot build derivation '/nix/store/q61vc27rdbcsrhxifiw89w1qdjax1amh-nixos-system-brix-': 1 dependencies couldn't be built error: build of '/nix/store/q61vc27rdbcsrhxifiw89w1qdjax1amh-nixos-system-brix-' failed Thanks\nMaybe the hardened kernel shouldn't set when NixOS' is an ?\nIt should be 64 bit for the x86_64-linux stdenv. But some packages use the i686 stdenv.\nI am not too familiar with stdenv. Is this expected behaviour that a 64-bit system uses 32-bit binaries? This issue is reproducible building as well.\nThank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as \"still important to me\". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly: for maintainers and people that previously touched the related code and @ mention them in a comment. on the . on the on .\nThis should be resolved by - please reopen if there's still a problem."}
{"_id":"q-en-nixpkgs-6d58d99560c981a886e182ffd9e34172321fd5e291759e87076fffd91a7c0367","text":"Here is result of — I'll try to do something about it when I have time but it'll most likely be next year, so please go ahead and fix it if you have time for minor issues.\nScreenshot?\npavucontrol has no binary wrapper, thus icons won't show unless relevant icon themes are installed in the system.\nFixed in master."}
{"_id":"q-en-nixpkgs-3ff732e1a55cc098dc036bf4b1ca8c308b79202bfc929d01f128b812494bc7ab","text":"Trying to compile fails with Install (not ) Create with run observe error\nhas gmp but doesn't have Base or Prelude (which is needed for stuff as simple as )\nFor reference, I think this issue is related:\nYou are meant to use to install additional packages with the compiler. I don't understand why the attribute it broken though from a brief look.\nI bisected to the commit that caused this: This was in the idris update from 1.3.0 to 1.3.1 Currently looking into how this can be solved elegantly\nIn the meantime, is there a workaround for this?\nYes, the PR linked above () fixes this, but I'm not sure what to do with it, because it breaks Darwin in weird ways."}
{"_id":"q-en-nixpkgs-2d00877d14207bebe5c507da7bb2d1e5f200200c229ee25fe0ced7e82d539739","text":"A Go program compiled using nix-shell returns an error when running on a separate server and trying to load a timezone file. file with: $nix-shell $<inside shellgo build $exit $ ./app # name of the app you just created 2019/01/25 20:04:58 All went well: America/NewYork Now copy this file to another computer #./app 2019/01/26 01:01:34 could not load new york locationopen /nix/store/0596k27r6pacyx4jp1bcn8zqyjhji52l-go- no such file or directory The error includes the zip name due to internal reasons the Go team has, but the real problem is that nix replaced with I would be more than happy to send a Pull Request if you agree with the proposed solution of prepending instead of replacing. $ nix-shell -p nix-info --run \"nix-info -m\" these paths will be fetched (0.00 MiB download, 0.00 MiB unpacked): /nix/store/dzvh4f6ihkwg44mmf2awxg564k1s8khg-nix-info copying path '/nix/store/dzvh4f6ihkwg44mmf2awxg564k1s8khg-nix-info' from ''... system: host os: multi-user?: sandbox: version: channels(diego): nixpkgs: version is nixos-unstable-2019-01-22 The file replaces the path that the go runtime looks for the timezone information. Original Go code is: But nix changes the first entry, for This works fine if you end up running your go application in the same nix env you compiled it, but when you deploy your go program to a separate server, the path, that can look like isn't present on the separate server and then the lookup fails. is a very common path in Linux distributions. A better way to solve this would be to prepend the nix only path, as the first entry, and then keep and the other entries as they are in the original go source code. The benefits are that those who compile and run in the same nix environment will use the nix path, but those who deploy somewhere else won't have to do any extra work.\nI doubt anyone ever considered support for non-Nix and therefore I will mark it also as . I'll leave it open for now as maybe it can be supported (I doubt it).\nwould you accept a PR with my proposed solution? It won't hurt nix deployments but will help a use case of building a go program in nix but deploying outside of nix.\ncross referencing a thread on the mailing list about this ticket/issue so both conversations are linked.\nYour explanation and proposed solution seem to be very reasonable to me. I suggest that you submit a PR."}
{"_id":"q-en-nixpkgs-7b513cd926ae3bce27fd6249012cca983a54867642645629e79e38c51ca374a7","text":"When I run `sudo crontab -e\", I get no crontab for root - using an empty one /bin/sh: /usr/bin/vi: No such file or directory crontab: \"/usr/bin/vi\" exited with status 127 Now I agree that I probably shouldn't do that (I should keep my crontab in ), but a better error message would probably be better. Please run and paste the results. system: host os: multi-user?: sandbox: version: channels(root): nixpkgs:\nI also have this issue. Current work around is but this needs to be fixed. See\nPlease feel free to close as this issue has been fixed."}
{"_id":"q-en-nixpkgs-153f5679e99c3c1ff90f4715bb35364f11e564f8a9d469c13c03bce6f5684ffa","text":"When used with latest Nix the \"new\" hash is extracted incorrectly, or thereabouts. This seems to be due to Nix changing the message used to be more friendly to humans: Use an update script invoking common-updater-scripts. Sorry, LMK if repro is needed. Please run and paste the results. I don't expect this to be helpful but happy to provide anyway! system: host os: multi-user?: sandbox: version: channels(will): channels(root): nixpkgs:\nIt is increasingly messy, perhaps we can drop the old regexes, most people probably set instead of changing nix in overlays so this would not be affected.\nWhat about adding a flag to nix, which turns on machine readable error messages with a guaranty that those messages will not change in a backward incompatible way? Would need someone to add that. But wouldn't that be better to parse error messages intended for human readers with shell tools?\nAlternatively, if the build fails, I think we could do something like:\nI'm not sure I follow, but FWIW I ran into this and I set and don't set it via overlay. Looks like the new behavior is in 2.2, the current default for : (err formatting might not work for email responses, maybe? sorry :)) I think it's reasonable to drop support for nix older than what's in the tree-- especially since you'd have to overlay to replace it and as you say would likely do via anyway. Oh... that's probably what you meant! :D\nThank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as \"still important to me\". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly: for maintainers and people that previously touched the related code and @ mention them in a comment. on the . on the on .\nAny updates on this?\nI marked this as stale due to inactivity. &rarr;\nShould have been fixed by\nThanks!"}
{"_id":"q-en-nixpkgs-6700785ee95cad356b59d39c3c69cca406fe80be77af5945976459f5c8d284fd","text":"I've just set on my laptop while on holiday. The timezone never updates (tried restarting, waiting a while). Running I see: I see in that the service is run as the user. If I run: I also see the message . Whereas if I run there is no output and the timezone immediately updates (as shown by an immediate change in the time in my status bar, and confirmed by running ). It seems that the user just doesn't have permission to set the timezone. in (and unset ) to see if timezone changes , look for messages with \"Failed to set timezone\" (will only appear if the timezone it would need to set is different than the current timezone) system: host os: multi-user?: sandbox: version: channels(root): nixpkgs:\nIt worked for me after a reboot. Maybe the polkit rules are not getting appliead during a ?"}
{"_id":"q-en-nixpkgs-5acf0c821d7c22283e2924522d96e55f2c6b8580173528925f53937453d6595d","text":"Using with a doesn't work right anymore after (and the bugfix commit ). I have a simple Docker image I'm using as a , which sets up a user and puts some files in their home directory . (The exact setup doesn't matter.) When I use , the files that are present in the home folder in the base image are lost. This used to work before the commits above. I made this repro derivation: Try running and compare that output to the base image: You will see that the repro image is missing files. CC\n(To be clear, it is now broken for a normal , not a layered one.)\nFriendly ping , can you tell what broke here?\nit seems to be related to the layers ordering. If you both images, you will see layers are reversed. Could you check the associated PR fixes this issue?\nYes, that seems to work! Many thanks :)\nMerged so this should be fixed. Thanks !\nOh no, I found another aspect of this which is still broken :(. I think it's still covered under this issue, so I think we should reopen this one rather than file another, is that okay? There's still a problem when you have a command in the command: the layers from the base image aren't present. Here's my repro from above with that addition: If you run and look at the build output, you'll see the output of the command. Because the base image I'm using has an user named \"user\", you should see a line like this: but due to the regression from you don't see that line. This is preventing me from doing some extra setup that depends on stuff in the base image. Any chance you can see a fix ? :)\nDoes fix this issue?\nYes , that seems to work! Thanks again"}
{"_id":"q-en-nixpkgs-ed21dfbb87c2335fe712c6e2c97ab944415da47b3a131748cb8ba3afe6de2529","text":"References to (5) should have hyperlinks to the article/section.\nThese are references to man pages so there can't easily be a hyperlink. I made a PR that makes sure these are referenced to using , though. So that they at least carry some semantic meaning.\nMy bad I didn't think about the man pages. There are man pages online, too. I found this site: but it'd be nice to have them generated into the NixOS documentation. Having the man page sources piped into HTML would be useful I not a huge space waster."}
{"_id":"q-en-nixpkgs-792daead3c7abfd024b3b9d2055871074f96e2a592f1e95604cb7e8e32ce76d9","text":"the python binding for krita are not build correctly. Please run and paste the results.\nmore details wouldn't hurt...\nwell I don't exactly know what is happening here, to build the python binding, it requires and . So when are to the buildInputs, cmake finds all the dependencies correctly; however, at around 98% the compilation fails due to not finding some sip bindings. Alternatively, if is to the buildInputs, everything works fine as intended. But the python interpreter overwrite the python interpreter in the nix-env, so that doesn't seem to be the right fix.\nalso tried setting the variable, but that didn't work either.\ncc who worked on / in"}
{"_id":"q-en-nixpkgs-0aef27920325644f730d5f3ea3457f297641486ccf41d04d5c550553b421b0a1","text":", (nixos-unstable) (nixos-unstable) (nixos-unstable) (nixos-unstable) Scanned versions: nixos-unstable: . May contain false positives.\nFixed in staging (again) via PR"}
{"_id":"q-en-nixpkgs-380521c705e290e59ea10ff28d0bcd978fa5f3d639c4468c0ae6811837e9a4ab","text":"I stumbled upon this package broken while working on , filing an issue to track it. Tested on commit . system: host os: multi-user?: sandbox: version: channels(root): channels(kalbasit): channels(yl): nixpkgs:"}
{"_id":"q-en-nixpkgs-14faabc686787f0bbf6ae0adda721d6d33280e1acca948941da7317c648af17f","text":", (nixos-19.03) Scanned versions: nixos-19.03: . May contain false positives."}
{"_id":"q-en-nixpkgs-6420370ff3cc65dc7fd0617fe99d0d3cfd6c219c622b19b124f3330c59403fb8","text":"is currently set to just , but I was able to build it fine with on macos (10.14.3) and haven't run into any issues running very basic commands so far.\nWould you mind making a pull request?\nI don't, but I'm not in the loop and won't be of much use if it's any more involved than changing it to .\nSo perhaps should be just ."}
{"_id":"q-en-nixpkgs-53444a6c78f01119439ef0716aea02c8909b3b8bc15df41d0d22565bbe1548c1","text":"When trying to build llvm-7.0.1 on my armv7 machine, I get the following error during the test phase: Looking around, I see quite a few changes to llvm recently (also the aarch64 failure in ). Ideas? Are these tests expected to fail?\njust because I know you know about these things. I know armv7 isn't \"really\" supported, but this definitely used to work, so... FWIW, I have tried to disable DebugInfo tests, using the method already used to disable other tests, like this: But this didn't do anything. I do not know the test system for llvm though, I just tried to identify which cpp file to remove by grepping around. So maybe this is not even one of the right files, or maybe I disabled it wrong. I have not been able to find how to disable the llvm-dwarfdump tests, also. Hints welcome, if you have any!\nI am also able to reproduce these failures. I get the same list of failing tests as what's shown above.\nOh - in case it's not obvious to everybody, I should point out that these are x86-specific tests. Obviously they're supposed to work from any host architecture, but since most developers' machines are x86, I think it's entirely possible that somebody upstream introduced an unintended dependency on the host architecture. I've looked around a bit for any upstream bug report, but so far I haven't found one.\nIf you're in the situation I was in, and are unable to rebuild your ARMv7 system because it has inexplicably decided it needs llvm 7.0.1 as part of that build when it didn't used to, you can add something along the lines of the following to your You will have to adapt it for whichever package is pulling in the unbuildable LLVM, which you can discover by looking at the list of unbuildable packages that nixos-rebuild spits out at the end and finding the one immediately after LLVM. You can verify that it's the right one by checking all- in the repo and seeing that its callPackage parameters specify llvmPackages_7.\nThanks for the helpful workaround,\nThanks!"}
{"_id":"q-en-nixpkgs-332f8ce626ebc6d4ddb7a4db952eb2659256740365ab590dacc9568b3b697dac","text":"migrated from python branch issue tracker\nthis is not an issue if propagedBuildInputs are used - whether that is good or bad is another thing.\nthe python rework solved this"}
{"_id":"q-en-nixpkgs-64ab5d265792e8067b2e203a3e433d7fcb56fcd94016af96e922faf73714748d","text":"migrated from python branch issue tracker\nI'm closing this as it seems no longer relevant. From what I could find, python-virtinst is \"CLI tools for building virtual machines using libvirt\". But it is now merged with virtmanager: <-- contains one single file named MERGEDWITHVIRTMANAGER"}
{"_id":"q-en-nixpkgs-eae44e602d929dbf2672e04968de1117331feb122bd24af20e3bda613f9ed4a1","text":"I'm using the latest from and it works fine, mostly. I notice the following errors during start-up, though: Looking at the build it seems like all recommended dependencies are there so it doesn't seem likely that those plugins weren't built, no? But why are they missing at run-time? And why are others (like the equalizer) not missing?\nAfter some bisecting I found to be the cause of this, reverting it fixes it. I'm not entirely sure why that is though, investigating..\nWas fixed in calf with , which is in the latest version, I have opened a PR to update calf"}
{"_id":"q-en-nixpkgs-a731f764b0a644e865434134a777e27eb489e53b1e6c5a6d18914b645103acd5","text":"Installing the gimpPlugins.resynthesizer2 package does not result in a working plugin; the python scripts are installed, but not the two native executables it depends on (called \"resynthesizer\" and \"resynthesizer\" Try to install the package \"gimpPlugins.resynthesizer2\". Expected result: There should be a few new menu options in the Gimp, amongst them \"Filter -Enhance -Heal Selection...\" Actual result: No new menu options; python scripts cannot find the executables. Possible solution: I made it work by creating a derivation that adds a the following postInstall: postInstall = \"installPlugins src/resynthesizer/resynthesizer src/resynthesizer-gui/resynthesizergui\"; system: host os: multi-user?: sandbox: version: channels(cable): nixpkgs:\nYou need to make the plugins available to GIMP, not install it locally. The following should work:\nJust tried it (on a nix shell) - but same result; the resulting installation of gimp-with-plugins links to a resynthesizer2 package that lacks the native executables, only has the python scripts (which don't work without the executables)\nHmm, there was an issue with , saves us."}
{"_id":"q-en-nixpkgs-cce4fbfb14394aea4a8c680569cf4006ae5967306863ac754919041edf7ac05e","text":"The tests for python27Packages.jinja2 fail and thereby the derivation fails to evaluate. (Full log below) I'm crosscompiling an aarch64 sd image with . I'm crosscompiling aarch64 with nixpkgs .\nI had the same issue too, this fixes it for me.\nFixed by"}
{"_id":"q-en-nixpkgs-68726d523a37baff63c5b8c4934a6a4b3f04ca6fed98a5d8aa114618a2a5afb6","text":"The new option has the example: This is a list of attrsets, but this doesn't match the type, which is Also, it would be nice to have an example config for as well. Besides this, the new declarative configuration options for syncthing have been working well for me. Thanks for including them. CC"}
{"_id":"q-en-nixpkgs-4bc62b52cdc9042d6eb89c532096c9e8322e19b5a9911825e7faa5592bc27354","text":"So today I tried to build on unstable as described in the . This fails on the dependency . Thanks for your information in that guide! I tried to build according to the guide on and (with slight modifications for compatibility) on they both fail in at this point. Do you have any insights as to how you managed to get it to build? e.g. a nixpkgs commit?\nFix in"}
{"_id":"q-en-nixpkgs-a40813bf91b8c1f62e4e8eff1d615a0a0864c99625c705cf9ee7fe56b99cb09b","text":"When the service starts, the following message is printed to the logs: sshguard[]: sshg-fw-ipset: Could not initialize firewall Blocking does not actually work. Blocked hosts can still initialize ssh connections. Default sshguard config. system: host os: multi-user?: sandbox: version: channels(root): channels(XXXX): nixpkgs:\nIt seems like that is due to ipsets being created of different types than intended. I'm not sure where it's getting created with that though: # ipset -L Name: sshguard4 Type: hash:ip Revision: 4 Header: family inet hashsize 1024 maxelem Size in memory: 88 References: 1 Number of entries: 0 Members: Name: sshguard6 Type: hash:ip Revision: 4 Header: family inet6 hashsize 1024 maxelem Size in memory: 104 References: 2 Number of entries: 0 Members: Whereas the script intends to create types which is why it returns with an error. I do not see where in the sshguard is type of ipset getting created. I came across though."}
{"_id":"q-en-nixpkgs-80ba3e01595c471105f4ad185af0cca27ed8e01f1782658bc84169b5e3a1f0f9","text":"is unmaintained, superseded by node2nix and should therefore be removed from nixpkgs Please run and paste the results.\nseems LGTM. Wanna open a PR ?"}
{"_id":"q-en-nixpkgs-5a96472d2202e59a6b9a1dec38e278e403d1dbe52f2c1a387e2e8b290046b5c0","text":"no $ Add to installPhase\nFor some reason netpbm is installing the symlink to . Will be fixed by"}
{"_id":"q-en-nixpkgs-7d593f895b49fe9f58e675708c53d749d5d78b9e53e2d717bdf9ec7120718208","text":"Describe the bug fails on Darwin. To Reproduce Steps to reproduce the behavior: Expected behavior Package builds Additional context Error log: <details<summaryError log</summary</detailsMetadata Please run and paste the result. host os: multi-user?: sandbox: version: channels(yl): nixpkgs: Maintainer information: Possibly related to\nOn the master branch, I removed the derivation from the nix store with and I built it without substitutes; failed with the same message. This is not caused by , but help me discover it due to the update of the requests python package.\nmaintainer:\nI don't have a way to test on joblib for darwin so this is why I didn't have input on this PR. Is there an easy way to test nixpkgs on darwin? Hackintosh via kvm?\nThere is that works with some tweaking but I wouldn't describe it as easy. Another option is to ask @ GrahamcOfBorg to do the builds on the PR if there are not too many iterations to do."}
{"_id":"q-en-nixpkgs-af3d85083a8dd49285a0bfe0f0796c00780156a94589666392b33db0c1f7f1ae","text":"Trying to build zathura on unstable, I get: Additional context I can't really see what might have changed recently (I think it started happening about 2 weeks ago....) for it to no longer find synctex. The texlive package is being included as a build input, which should have synctex Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs:\nI can build zathura on master and the hydra has been doing ok. Am I missing something?\nHm, I'm not sure. I can't seem to get it to build.... \"nix-build -A zathura\" on master gives me the same error\nThis also builds for me on master. Only thing i can think of is your branch isn't up to date.\nWhat do you think could be wrong? My sources are up to date:\nI notice that I get a different drv on my system (), I think you may not be building from the master. IIRC when you run it will build from which may not point to your git checkout. If you want to build expressions from the current directory you need to do nix also can download a tarball with expressions, for example like this:\nIf no path is given nix-build will read and If doesn't exist it will fail. So, command is fine.\nI just checked on another x86_64-linux NixOS system that evaluates to on . If the command is correct then there must be an override or an overlay in or that affects the derivation.\nOh my god, you're totally right. I had an overlay in my , which I think I had to add to disable synctex at one point when it wasn't working. Removing that and it builds fine. So sorry for wasting your guys' time but thanks for the help! (Although I wonder why it says it can't find synctex when it should have been disabled with the overlay?) Anyways, I'll close this issue, thanks again, it was driving me crazy!\nWhat was the overlay? There may still be a bug here.\nThe build indeed fails with . I don't have time to open a PR now but the fix is pretty easy: can you reopen the issue?\nSure - it also fails to build with synctexSupport = false, not sure if your fix addresses that as well?\nThe option was introduced in I think we should just drop variant"}
{"_id":"q-en-nixpkgs-4de7f3c0dbea8113a9799b8c4b3c75f57a3eb3652dc31f4ec9313266bee32b4c","text":"Is it possible to implement mount options for swap partition? This feature could be useful for machines with SSD only drive(s). User could change to options for automatic trimming. of this feature.\nI think this is what you want\nI still can't figure out solution with documentation pointed out. If I put raises an error. I am using NixOS 14.12 stable. As far as I know swap partition does not have mounting point.\nyou may do it like this: even though, i didn't test it, but using like this should work regards, iker\nStill, it would be nice if there were a more idiomatic way to accomplish that.\nSorry for long pause. After some research, setting discard on swap is NOT good idea. (see ). Setting it, would trigger trim command every time block clears, which leads to poor performance when swap is under pressure. Issue can be closed.\naccording to the there is a single variant, which is NOT that bad idea:\nOn top of the case pointed out by of doing a oneshot discard of swap space on startup, there are some situations where there really isn't much of any performance impact. As a simple example, I'm running NixOS in a VM, and the hypervisor translates TRIM/UNMAP/DISCARD commands into calls to fallocate() with the FLPUNCHHOLE flag on the host side. This ends up completing almost instantly due to how my storage stack is set up (and will in most cases where people are using file-backed VM disks), and more importantly, it can be queued on the VM side because it's on an emulated SCSI device. There's literally no measurable performance impact in this case, but there's a big impact on the host side in terms of simplifying space management. Even ignoring use of virtualization, 90% of the performance issue for inline discards in general (not just for swap space) is due to most commodity hardware not being able to queue them, which results in them having about the same impact as an FUA request. Some hardware, however (notably most SCSI and NVMe devices) can queue them, and they have very little impact on performance there.\nsince this issue got new visits again, i've uploaded , which is a start. I've also looked into unifying and , but it seems a lot of work if such a change doesn't get merged.\nWhat if both and options are in use? /etc/fstab gets 2 entries: Interestingly, the nixos-generate-config options is placed in a lower part of /etc/fstab. There is hope that the first entry gets precedence. But I don't know any mean of checking . the issue still unresolved. Can you reopen, please?\nI'm using nixos 21.05 and want to add the \"nofail\" option for swap devices that are in raid1. Options: can send a PR for a option that gets passed to . I have this and can put it up quickly. This mirrors . can send a PR for a specific option. can use , which can support arbitrary options. But then won't be reproduced correctly. Any opinions?\nThe approach would get my vote, as it avoids any possibility of arguments between users and developers when a user wants to use an option the developer thinks should not be used (such as the option that started all of this)."}
{"_id":"q-en-nixpkgs-aed0e9da9052cf6a97cc09c69c6f8288cd1ff06f5cf847448ee719f9cd84a97a","text":"Describe the bug When I'm trying to install he arduino or ino package on MacOS I get the error message has wrong length for hash type 'sha256'. , one of the MacOS entries has a hash that is longer than all the other hashes. To Reproduce Steps to reproduce the behavior: Run: Expected behavior Arduino is installed in a nix shell. Screenshots ! Metadata Please run and paste the result. Maintainer information:\nIn the pull request mentioned above I have corrected the obvious error in the hash. I don't have Nix on Darwin to try it on, can you?"}
{"_id":"q-en-nixpkgs-f669f70f11d2f0e1207e7aae42c3e1be72a3b1fcb29ba1dfca22a72f91b50999","text":"Project description sasquatch is a varation of the unsquashfs utilty that can unpack a much larger amount of vendor specific flavours of squashfs that the regular unsquashfs utility cannot unpack. It is distributed as a patch to the regular squashfs-tools package (so might need some extra work to work with the squashfs-tools package already in nixpkgs). Metadata homepage URL: source URL: license: gpl2+ platforms: linux"}
{"_id":"q-en-nixpkgs-67cfc549f68ef597a18c96ab52385136f55f9434d343a6be8b46151cf506eb3d","text":"Project description Metadata homepage URL: source URL: license: lgpl2/lgpl3 platforms: linux\nPreviously: (which I never finished, sorry, but linking in case helps anyone tackle this)."}
{"_id":"q-en-nixpkgs-5b68f9698f55a44c4183b15b6454838fd4d9a17c21fdddb67179afc5f23fe68b","text":"Describe the bug After installing inkscape either in the user or system profile, no icons are found. When opening dialogs such as \"Fill and stroke\" or \"Layers\" the program crashes. I've confirmed this on two of my machines. To Reproduce Steps to reproduce the behavior: inkscape that all icons display incorrectly to Object -Fill and Stroke crash Expected behavior Inkscape should have the full set of icons and dialogs should be accessible. Screenshots ! Additional Context <details<summaryConsole log</summary</detailsMetadata system: host os: multi-user?: sandbox: version: channels(root): channels(riley): nixpkgs:\nI think you can easily fix this issue by installing an icon theme to your system.\nI believe these icons are part of Inkscape. Possibly issue with gdk-pixbuf loaders – though, I cannot reproduce it within plain xterm session. What desktop environment are you running? Does adding to Inkscape’s help?\nSorry, I should have mentioned - I already tried the following overlay for inkscape after reading the gdk-pixbuf issue and related ones. Unfortunately it didn't seem to change anything after rebuilding and compiling. One of my computers is running sway (launched via startx) and the other i3 + sddm.\nMaybe does not work because of :\nThat worked! I to the overlay and all the icons show up, and dialogs work. I tried looking up docs on but I couldn't find much about it, what does it do exactly?\n:confused: I thought the issue with strictDeps only touched gobject-introspection\nIIRC, it might also affect hook’s not being picked up correctly. Does the following work with enabled?\nYes, that worked! Would the best workaround for now be to add those inputs to the derivation?"}
{"_id":"q-en-nixpkgs-1e6648b2bbeb2bcf6b936d5809cb1bc9b883bffb686bcd87ad1c428c8069689b","text":"Describe the bug After upgrading to 19.09pre, I am always asked for a password to unlock from . I don't recall this happening in 19.03 Some more info on how to set this up: To Reproduce Steps to reproduce the behavior: KDE with KWallet enabled out and back in to launch chromium that a password is needed Expected behavior Password used for login is used to unlock kwallet. Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(root): channels(mbauer): nixpkgs: Maintainer information:\nWhat display manager are you using?\nThis is the sddm display manager.\nI've also noticed this.\nI wasn't upgrading from stable, only from ~a few months ago, so it must have been relatively recent.\nThe issue here is with the pam module then if it's not automatically unlocked. ~Do you use Plasma 5 which adds?~ (read it says you use KDE) I believe this was broken with lightdm and I've recently fixed it there Could you try using lightdm and see if there's a difference?\nYeah, I think this one is KDE specific. Noticed this in my log: Looking at the coredump gives me this: Looks like a wrapping issue. This is what kwallet looks like:\nI'd agree that it's most likely a wrapping issue since that's the only other major change we could tie it too. cc\nI agree that the failure in sounds like a wrapping issue. Could you confirm that exists? Unfortunately, I can't tell which platform integration plugin it's trying to load; I would need to see the environment in which PAM loads.\ntwmn now crashes in QGuiApplicationPrivate::createPlatformIntegration again as well (I remember this happening a long time ago) I remember before I could do but even now that doesn't work\nThis looks like it's not happening anymore on my system. Probably fixed by either or\nActually still seeing it...\nNot sure what's happening so I'll run a git bisect on it.\nShould be fixed with .\nIt works correctly on 19.09, but regressed on current unstable:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nHi, I get this again on 20.09 stable. In my case this happens cause of: Core was generated by `/run/current-system/sw/bin/kwalletd5'. Program terminated with signal SIGSEGV, Segmentation fault. 0x00007fa0556062ec in QObject::disconnect(QObject const, char const, QObject const, char const) () from /nix/store/9fmvcacvkj6s65j3l5pf96nbqwjp73wb-qtbase- /run/current-system/sw/bin/kwalletd5 seems to crash only with Plasma sessions"}
{"_id":"q-en-nixpkgs-1228ceede643130a6c5ecce84a9025a69b5326e2a0dfb34b581f4602c016bc1c","text":"Project description Oh My Fish provides core infrastructure to allow to install packages which extend or modify the look of fish shell. It's fast, extensible and easy to use. Very similar to Oh-my-zsh that is already in NixOs package repo Metadata homepage URL: source URL: license: mit platforms: unix, linux, darwin,\nThank you for your contributions. This has been automatically marked as stale because it has had no activity for 180 days. If this is still important to you, we ask that you leave a comment below. Your comment can be as simple as \"still important to me\". This lets people see that at least one person still cares about this. Someone will have to do this at most twice a year if there is no other activity. Here are suggestions that might help resolve this more quickly: for maintainers and people that previously touched the related code and @ mention them in a comment. on the . 3. Ask on the on .\nAny news on this?\nI am still trying to understand how oh-my-zsh works.\nWell, it just looks like a copy-paste. Just a giant install phase. <details<summaryTREE</summary</details\nIsn't this issue about oh-my-fish instead of zsh?\nYes, it is. And I am studying oh-my-zsh in order to use it as a \"template\"."}
{"_id":"q-en-nixpkgs-c480f32e09468122f741c3d70a69b63610d2ba93f5d27e3dc04abb634247ae41","text":"Describe the bug A clear and concise description of what the bug is. Looks like a new failure for mariadb-server: Currently blocking nixpkgs-unstable\nI strongly suspect the update in is the culprit"}
{"_id":"q-en-nixpkgs-f8480fbfc318d60ecc57f0cc14f8f0371865d42fc5208abf347ad9c4e5486c73","text":", (nixos-19.09) (nixos-19.09) (nixos-19.09) (nixos-19.09) Scanned versions: nixos-19.09: . May contain false positives.\nNeeds patching on master, 19.09, 19.03. CVE-2019-: CVE-2019-: CVE-2019-: CVE-2019-:"}
{"_id":"q-en-nixpkgs-f49bb5d7831f3c8e0e0ead79f8fec8b4812c11c0f29876776a5943ae4030e431","text":"Describe the bug A clear and concise description of what the bug is. To Reproduce Steps to reproduce the behavior: -build -A chefdk ./result/bin/chef ./result/bin/knife Expected behavior The chef executable is there but is broken, the knife executable isn't even present. Additional context broke chef, it worked before that commit. I've bisected it down to that changeset. Problem for me is that I don't know ruby or the ruby ecosystem... This is a problem in 19.09 as well. Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:\ncc\nI also ran into this; I tried to debug and build it myself locally but it never seemed to rebuild when I ran in my local copy of nixpkgs. Even so, showed just the binary and nothing else. Eventually I tried the following diff to force a new build: Which caused to actually rebuild, and for some reason the rest of the binaries showed up in even with no other changes: I'm relatively new to nix, so not sure why bumping the version changes anything.\nInteresting find! Just changing to makes it work as well, even without adding a version to the end of it. I tried keeping and adding a separate attribute but no such luck. There's something odd going on with I'd say.\nSo what seems to happen is that bundlerEnv is two different things depending if you have a or a . Maybe there's a bug in there somewhere: But I made PR's to revert to using instead for now.\nSeems like this is still/again a problem ( missing) on NixOS 21.11 since it has been again mass-replaced in\nAfter initial testing this seems to work. Are there any downsides to using here? (except for the fact that we need to list out executables to be exposed)\nI don't know the ruby tooling in nix enough to know if there's any downsides or upsides of making that change. however, I'm not sure why version is removed? If you would make a PR of that we could ask someone with more ruby insights.\nSure, I can do that next week. IIRC / changes behaviour depending on existence of param. If it is missing, it'll be pulled in from relevant gemfiles in anyways.\nSounds good to me, I'm currently running my install from nixos 19.09, because it was the latest release where I found it working :wink: I guess that's the power of Nix, but it's not a good thing to have broken packages like this. So ping me when you have that PR open and I'll test it it and possibly merge it :)\nHeyho, any updates on this? :) I finally also migrated my work laptop to NixOS and just ran into this because the binary is missing. I am not at all experienced with neither Nix or NixOS but just renaming to in the appears to fix the problem, the presence of the version variable apparently doesn't have any impact on the result."}
{"_id":"q-en-nixpkgs-06f9f0da082c647184518e75d742cf1a566603d8c752b86f8b5a09de12ab2e3e","text":"Describe the bug See https://nix- The particular cases that fail are I believe rust was updated recently (was that in the release?). It doesn't appear upstream has made a release recently, so perhaps if this report is forwarded to them they can direct us to which patches could fix the test failures etc. cc\nAccording to and bug reports, this seems like a ulimit issue. An interesting workaround for hydra could be using , as suggested in but it's certainly isn't necessary on my machine because I've tested the build and it passed it.\nsounds like it would make an already long checkPhase longer :rofl: How about we try ?\nWe can try this indeed. The question is, what are the default limits of a build user on hydra? Are we sure this might make a difference?\nI noticed you've marked sequoia as broken in . Even if a package fails to build on hydra butt doesn't fail locally, does it mean we mark it as broken in its ? Doesn't sound right to me..\nThat's was the motivation for the issue being opened. If we bump in preCheck and unmark it as broken I think we should be good. ofborg can serve as testing perhaps similar to hydra.\nThe point is, that we haven't even tried playing with these settings. I would have tried doing it my self but since I know nothing nor I have any effect on Hydra's deployment etc., I wouldn't even know how to test it. What bothers me is the fact it was marked as broken without any attempts by someone with more influence then me to debug it's failure to build only on hydra. Are we even sure this wasn't a failure that happened only once?\nShould we close this issue with the PR?\nNo, it's still failing , at least for hydra, it runs out of file handles\nA bit off-topic maybe but I generated a Nix expression for sequoia using , as build times for sequoia where around hours for me. It works now like a charm, as every crate got its own derivation. It can be found here, as a flake: Maybe repackaging with could be an option?\nThis may be an option, but unpleasant as the generated deps file usually is 1-10k+ of generated noise. (for local building, this works great)\nYeah, that noise is a bummer. Sometimes I dream of having all of packaged the way we do it with , but I guess that would blow up Nixpkgs indefinitely. But maybe someday when Nix flakes are ready? …\nI'm not sure I have enough time to have a good view on this. What's most important about specifically, is the fact we use upstream's to install many other files that should be part of the package's output - files such as libraries, ZSH/Bash completion files and headers. So I'm not sure how will this go together with the approach.\nIf anyone is interested in a sequoia flake which is built incrementally (vs. everything built in a single derivation), I try to maintain one here:\nsequoia , but it hasn't tried 0.16.0 yet - .\nYep, we need to fix it for python 38 though"}
{"_id":"q-en-nixpkgs-fb38109143b938a286bc3256dada1fceb8d8a9c16ffedf982d23a36570eb6f18","text":"Describe the bug By default and according to bazel documentation (), bazel is looking for a system-wide configuration is . In I decided to override this path and make it point in . It was the easiest solution to pass arguments to bazel in order to fix a few runtime issues. However it breaks the assumption that user can configure bazel using . To Reproduce Use bazel from nixpkgs with . Set some bazel settings in your file Observe that bazel ignores this file Expected behavior I'm opening this bug report to discuss what should be the expected behavior here? Should we respect the ? If yes, I'll open a pull request to restore the previous behavior.\nI'll be in favour of keeping support for all paths specified in the documentation - having a system-wide is something people might want to make use of, so we shouldn't take away that possibility. We currently set the following options in bazelrc: We should either patch these as defaults into nixos, or introduce reading from a distro-provided location (should be coordinated with bazel upstream). I think adding a wrapper is a bit too much for all these options.\nI do agree with your point of view. I'll try to propose a PR which restore the old behavior. In the meantime, I'm open to other comments on this issue. Regarding the options you quote, they are used in a local during build only. The global bazelrc that I provide is way simpler: Both were ad hoc solution for different problems, respectively issue and commit ."}
{"_id":"q-en-nixpkgs-226d74b92aca3d86b0bd1a41f4241ff37d1afb170b915021283e329aad3a9618","text":"Capability wrappers are somewhat limited. They require the capability be part of the ambient set for it to survive execv's, ie. when the wrapper executes the source binary (). This means any program spawned from a wrapped binary will also inherit the capabilities. We really don't want programs launched from gnome-shell to inherit capsysnice obviously. In particular bubblewrap fails to work if it possess any extra caps, breaking at least flatpaks and epiphany. So as far as I can see we have two options: Actually copy to wrapperDir and set capsysnic=+ep. This should be pretty easy, but also quite ugly (see for context). Patch mutter to drop inheritable, we could probably adopt . This should probably work as the ambient set only propagates if inheritable is set. cc\nPatching mutter to drop inheritable works :tada:, will make a PR :)\nDoes the rationale for not merging it not apply to us?\nNo it shouldn't. We only drop inheritable to prevent the ambient set being passed on, while keeping effective and permitted for gnome-shell. (it might be possible to drop the ambient set with , but I went with just clearing inheritable):"}
{"_id":"q-en-nixpkgs-ac275771c026a79347a997b8fe86310934c9e693bfdfaae8c8658da2526df274","text":"on Hydra: cc\nSince it's been building for me, checked if I was sitting on any changes and while the commit message doesn't indicate that it fixes anything (eep), it does seem to fix this issue: Don't remember enough details to justify it fully offhand (again d'oh for not writing it down at the time hehe) but hopefully this helps work towards a suitable resolution :).\nNaturally this breaks in as well.\nYep, broken for me too. Haven't had time to tinker yet.\nI have the same problem, but the commit mentioned fixes it for me.\nSo are we just waiting for this commit to be in the unstable channels?"}
{"_id":"q-en-nixpkgs-f2e5667d874e27524660044cf85dabf68a9226e356d3debc4549b43d3ebc10cb","text":"Describe the bug Trying to rebuild from git master and it complains about fonttools and python version mismatch. To Reproduce Metadata Please run and paste the result. root nix run -c nix-info -m system: host os: multi-user?: sandbox: version: nixpkgs: channels(root): Maintainer information:\nLooks like an issue with trying to build it with Python 2.7.x. The code isn't entirely compatible with Python 3.x nevertheless.\nThe issue seems to be with nototools which we currently have disabled for python3. The last nototools release is apparently fully compatible with python3. I'm working on a fix for this, however, the build now has a strange relationship with noto-fonts that I haven't quite figured out yet.\nIn the latest release nototools seems to still fail, or at least does on mine. The only relationship I found is that nototools is a dependency on noto-fonts.\nyou mean when running , or something similar? That's expected since nototools has been disabled for python3 (python3 support has only recently been upstream). I believe the relationship is actually that nototools is a dep of noto-fonts-emoji. The current nototools seems to require noto-fonts as a dep. I've gotten most of the way to a fix I think (changed nototools version and enabled py3 and what we fetch with the fetchZip in noto-fonts). However, the way that nototools expects noto-fonts is pretty strange (as an example look at the top level makefile where it looks for the noto-fonts repo in nototools' parent directory). I'm having a bit of trouble getting noto-fonts in the form nototools expects. I'm trying not to just copy the entire noto-fonts source into nototools source which seems like it should be unnecessary. Once nototools is fixed for python3 the rest should be easy bc we can just change noto-fonts etc to use python3. If you have any thoughts on how to fix nototools I'd love to hear them, or if you think I'm missing something here pls do let me know.\nshould be a dependency of the derivation. I'm not sure where you're seeing the fact that needs as a dependency. The farthest I've gotten is removing all the errors from the configuration but the build still fails with .\nYou mean that it should be a dependency of or it is currently? In regard to \"is currently\" I don't see it as currently being a dependency. The only place appears in noto- is in the following lines as part of derivation. If you mean it should be, then how you get to build on its own? If you change the nototools version to 0.2.0 which adds python3 support I get errors about missing font directories (these directories come from noto-fonts). What am I missing here?\nProposed fix Still building though...\n, and have the same error."}
{"_id":"q-en-nixpkgs-043737207f88022ae6be87885e40e80a3ae6e911a597b1946a45fc3126a51235","text":"Describe the bug The stable channel was last updated 5 days before. The blocking issue seems to be the build of nlohmannjson timing out. To Reproduce Steps to reproduce the behavior: to local nixpkgs repo: latest changes: stable channel: : That works, so there is no actual issue with the build. The problem is that a test times out on the CI: Maintainer information:"}
{"_id":"q-en-nixpkgs-8616104ddd5623b7fe9bb954d7e495ff0490bf1bd871172fbf858c723a2af719","text":"p4merge is unable to start due to missing library. results in:\ncc\nFixed by backporting"}
{"_id":"q-en-nixpkgs-b814dfce750d92ff34e3ec664a55a651bfff56120c3d1c462c06046241d75e41","text":"Describe the bug The latest macvim fails to install with a build error like below I think this change is the cause: The previous version builds fine. To Reproduce the channel to the latest one. or update macvim: or will see the error back to the previous generation: works fine Expected behavior Should install/update fine. Screenshots N/A Additional context macOS 10.14.6 (18G87) Metadata Please run and paste the result. Maintainer information:\ncc:\nHuh, it builds on my 10.14.6 machine, I wonder what the difference is. Also looking through the build output right now I see on the link command for MacVim. Did you customize macvim, or just build it as-is?\nI guess the link command for the binary doesn't include , just the one for .\nI don't have any customizations.\nOk I’ll look into adding the right flag to the Vim link step when I can.\nCan you test and let me know if that fixes it?\nWill do but I'm quite new to nix and don't know how to apply that change to my local nix env. Any pointers?\nNormally I'd suggest using to build it, but that builds with the sandbox and macvim can't build with the full sandbox (it requires the sandbox to be off or in ). Given that, what you can do is just check out the PR branch. If you have a git checkout of nixpkgs this can be done with . Then you can run to build macvim from that checkout. If this succeeds it'll create a symlink in the current directory that contains the build results. You could in theory install this into your environment too if you like, with .\nCan do or\nworked. Thanks, & !\nThank you for testing. I will merge the fix now."}
{"_id":"q-en-nixpkgs-d495e929aa7788b2fe1b0d736390b33caeeb3953ddf4a4ec4ee4ebd8c382de6e","text":", (nixos-19.09, nixos-unstable) Scanned versions: nixos-19.09: ; nixos-unstable: . May contain false positives.\nCVE-2019-: Fixed at version 0.60.8 (nixos-19.09, nixos-20.03, nixos-unstable)."}
{"_id":"q-en-nixpkgs-c3c6ca9fbf8d2785da6b5eef5accb46d12d929a7af3bc5fbe57f9f0b3e4b1bd0","text":", (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) (nixos-19.09, nixos-unstable) Scanned versions: nixos-19.09: ; nixos-unstable: . May contain false positives.\nSee also: ,\n09 updated in (staging ATM, but should get into channel within days). I trust the bump solves all these CVEs."}
{"_id":"q-en-nixpkgs-189a3d0fd7b808eece8cb5f159cfe066f7a8577a3cb138949b9fd8ef5cb6e655","text":"Describe the bug git fails to build on darwin Metadata nixpkgs at Maintainer information:\ncc\nexample hydra log: culprit failing tests:"}
{"_id":"q-en-nixpkgs-6078fd3301283e69634b082681e4bdbe220ee19d38c9975169f5fed980826205","text":"Nixpkgs has 0.11.6, the latest release is 0.11.8"}
{"_id":"q-en-nixpkgs-db00d565483b8b128ee433859fc7395ee3ee4009611c13ac2ddff4d89084fab8","text":"Describe the bug I'm installing send2trash Python lib from the unstable channel. It should be the 1.5.0 version, but it 1.4.2 in fact. To Reproduce Steps to reproduce the behavior: write the following: send2trash buld results: Expected behavior The expected result should be: Metadata system: host os: multi-user?: sandbox: version: nixpkgs: Link to the nix package"}
{"_id":"q-en-nixpkgs-ae54afa00f43791efd3bde49aa2de03f191294c3afd885d5e766a37a70f0e738","text":"Project description Cryptomator is an application providing client-side encryption for cloud storage. Metadata homepage URL: source URL: license: GPLv3 platforms: linux, darwin\nSource URL:\nHello, I'm a bot and I thank you in the name of the community for opening this issue. To help our human contributors focus on the most-relevant reports, I check up on old issues to see if they're still relevant. This issue has had no activity for 180 days, and so I marked it as stale, but you can rest assured it will never be closed by a non-human. The community would appreciate your effort in checking if the issue is still valid. If it isn't, please close it. If the issue persists, and you'd like to remove the stale label, you simply need to leave a comment. Your comment can be as simple as \"still important to me\". If you'd like it to get more attention, you can ask for help by searching for maintainers and people that previously touched related code and @ mention them in a comment. You can use or on the relevant files to find them. Lastly, you can always ask for help at or at .\nThis is still relevant to me. There is currently no way to get transparent encryption on any cloud-drive.\nguys, whoever is interested in cryptomator testing/reviewing are welcome: we absolutely need it for darwin platform as well?\nLet's first get it merged for linux, we can then base on that and check if we can get darwin working too."}
{"_id":"q-en-nixpkgs-80df0ae06be6f0fc06215148297400b299feba0831ba49cf39ce20c53a3cae2f","text":", CVSSv3=9.8 (nixos-19.03, nixos-19.09, nixos-unstable) Scanned versions: nixos-19.03: ; nixos-19.09: ; nixos-unstable: . May contain false positives.\n03 is out of support - we don't backport security fixes there anymore.\nBut this was also about nixos-unstable.\nThis was a grave mistake from your side. See"}
{"_id":"q-en-nixpkgs-4df5cf3f227dbea115a3d7bd071dd390cc0dd8cedfdc475a9378d20764d82338","text":"Describe the bug I have a NixOS installation with a very simple xserver configuration: When trying to rebuild with master I get an error: To Reproduce Steps to reproduce the behavior: my lines from the above Expected behavior The old configuration should continue to work unless I opt into Metadata Please run and paste the result. Maintainer information: This has to do with cc\nRight, looks like there's a bug when constructing the legacy the session name from the legacy defaults. Specifically when only supplying a default for the window manager. Will fix.\nIMHO, is only meant for mutable state like database schemata, that does not offer downgrade/rollback path. Abusing it for versioning option tree only adds complexity for little benefit, especially when we already employ other transition aids. Nevertheless, this was an oversight on my part. Something like this might fix it:\nWhat is the proper way to update my configuration to the new system? Where should I look for that information? Are there release notes? I couldn't figure it out from\nYes, release notes are included in the pull request: Also the module system should give you upgrade instruction when we fix the oversight. Could you test the patch above?\nYour patch fixes the issue for me.\nClosing since this is fixed.\nWe still need to apply the patch to the nixpkgs.\nOops, thought this was\nYeah, I just looked up that one, but it actually predates me filing the issue and it's currently the latest commit for the file."}
{"_id":"q-en-nixpkgs-8c444a4551105aed4f5b0252d10317e3b0dba7c8af71999d15dd2fcbadaa40d5","text":"Describe the bug I just tried to update my system today. It updated Flexget from 3.0.31 to 3.1.11. Installation went smooth but flexget ultimatively crashes when running. To Reproduce Steps to reproduce the behavior: flexget it it Expected behavior No crash Actual behavior Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(hyper): channels(root): nixpkgs: Maintainer information:"}
{"_id":"q-en-nixpkgs-6ab9eb15ba323717810a87dded45ef5624a2da57b8fa6801b78cc234a83000af","text":"Describe the bug See This is because vmTools.diskImageFuns.debian9i386 doesn't build\nHello, I'm a bot and I thank you in the name of the community for opening this issue. To help our human contributors focus on the most-relevant reports, I check up on old issues to see if they're still relevant. This issue has had no activity for 180 days, and so I marked it as stale, but you can rest assured it will never be closed by a non-human. The community would appreciate your effort in checking if the issue is still valid. If it isn't, please close it. If the issue persists, and you'd like to remove the stale label, you simply need to leave a comment. Your comment can be as simple as \"still important to me\". If you'd like it to get more attention, you can ask for help by searching for maintainers and people that previously touched related code and @ mention them in a comment. You can use or on the relevant files to find them. Lastly, you can always ask for help at or at .\nClosing since indicates that it fixed this."}
{"_id":"q-en-nixpkgs-becccc30792825d48a327fd8c49fcc1f29870dff59bb800d45a0285a0b4da11e","text":"The soundfont that is packaged in is no longer available. The URL in that file is , but that page says: \"Sorry, this is a wrong download link.\". No packages depend on this package."}
{"_id":"q-en-nixpkgs-c7983805ff706241ff1cd37140f0c0581ddbaa0fee0f8a2046f8ab3322c3dfa7","text":"On current master (). Fails with See also the . Additional context The hscolour derivation mentioned here does not contain in its . It seems sufficient to make be a propagated build input for , after doing so I can build ghc itself. There's a lot of Metadata Maintainer information:\nCan you send a PR fixing this?\nDone, see\n[ ] port to\nSince this issue was opened the situation has changed. There's still no ghc on aarch64, but now it fails due to [the lack of aarch64 support][] in (aka ). [the lack of aarch64 support]:\nAfter PR the default ghc builds for aarch64 (, ATM only in branch)."}
{"_id":"q-en-nixpkgs-7eef30cebaaa96e00d3e80bb3f2ff26ac4dc8d0fe124f10ab9d4b0a2ab593b9b","text":"Describe the bug The recent Haskell generic-builder changes for Darwin (in particular according to bisect) have broken . To Reproduce from a nixpkgs at or later on Darwin. This fails with . Additional context is a pure Haskell package, that depends on , which is what actually uses the C dynamic library. So the problem seems to be that fails to include the C library in its closure. Interestingly, this C library has filename not , even though it looks exactly like a dylib as far as I can tell (I'm not expert in macOS linking): So I can see why would fail to be included into by . However, I don't see why broke it. Nothing to do with the C or Haskell packages involved has changed for months. Updating the generic-builder to include as well as files does fix the problem for me, and seems benign, so I'll submit a PR for that. But I don't feel I understand everything going on here, so I'm opening the issue in case more qualified folks can identify a better root cause fix. Metadata system: host os: multi-user?: sandbox: version: channels(simon): nixpkgs: Maintainer information: Ping from the original changes - thanks for all your efforts!\nJust noting here that there is some useful discussion ."}
{"_id":"q-en-nixpkgs-441d4ee5858516ae49b7a7a0ce146be9f019302f55aac221e206c1f21c2bde70","text":"Describe the bug I have installed the du-dust package by attribute name (because if I install by name I get the wrong package), but when I run nix-env -u it gets swapped out for an entirely different dust package. To Reproduce Steps to reproduce the behavior: Expected behavior I would expect the du-dust package to be upgraded if a new version is available, no-op otherwise. Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(frijters): nixpkgs: Maintainer information: Maybe ?\nThanks!"}
{"_id":"q-en-nixpkgs-20a54c2afe207910183b94299fcdb8316f8c7550383fdf1e095ceb32a757fe87","text":"Describe the bug Attempting to start xpra fails with the following X server log: To Reproduce Steps to reproduce the behavior: run start :200"}
{"_id":"q-en-nixpkgs-29bcd12634a6f4b63574cfd389e304a1525b29403d582f581a092ceb9a155ddf","text":"Describe the bug Tribler does not have any desktop icons, and does not show up in any menus. There is a desktop file in But it's not being included by the desktop environment. Maintainer information:"}
{"_id":"q-en-nixpkgs-332f5fc699c6dd23b1c905b36e6962d4050a5f2a7d6ba01a30ee13dd44af1fb8","text":"Describe the bug On a ZFS-enabled 20.03 system: IMHO zed does not need the ability to send email by default. Also mariadb-connector should have a separate -dev output so that its doesn't end up in the system closure.\ncc\nHello, I'm a bot and I thank you in the name of the community for opening this issue. To help our human contributors focus on the most-relevant reports, I check up on old issues to see if they're still relevant. This issue has had no activity for 180 days, and so I marked it as stale, but you can rest assured it will never be closed by a non-human. The community would appreciate your effort in checking if the issue is still valid. If it isn't, please close it. If the issue persists, and you'd like to remove the stale label, you simply need to leave a comment. Your comment can be as simple as \"still important to me\". If you'd like it to get more attention, you can ask for help by searching for maintainers and people that previously touched related code and @ mention them in a comment. You can use or on the relevant files to find them. Lastly, you can always ask for help at or at .\nSeems still important: And: And:\nAdjusted to milestone 21.03. BTW, If someone wants to backport the fix it shouldn't cause any issues.\nTook a stab at this in"}
{"_id":"q-en-nixpkgs-8be087a16f78220fc4d75935f65326600b3354d0d7d4c539277c770b6a8aca5b","text":"Describe the bug After upgrading to NixOS 20.03, sudo often (but not always) segfaults: From the logs: Disabling appears to have made this problem to away, so the problem may be in . Metadata NixOS 20.03..\nMaybe the patch to use a not insecure openssl is the source of the issue for you\nsame issue here - the sourceforge repo is way old, there's a github mirror but not maintained either, and there's a fork of the mirror which includes fixes for openssl 1.1.1:\nI marked this as stale due to inactivity. &rarr;\nHmm, kinda forgot about this\nI missed the fact that the \"supported\" branch didn't actually include the patch to fix this."}
{"_id":"q-en-nixpkgs-d34576a5046dd0fea4af3058c7bd6ca36ab5366ff7249630cfb033e9928a518f","text":"PR broke -- first discovered in Before that PR, the systemd service ran as , so the nginx master process ran as root, and then the worker processes ran as an unprivileged user ( , or by default if thta line is not given). Now the systemd service runs as , including the master process. When you change the default back to root (e.g. because your setup needs the master process to be root, or for replicate the setup as it was in 19.09) using , it does not work out of the box. That is because the module now assumes that nginx will run only as 1 user, and it sets file system permissions that way. But nginx running as root makes nginx drop privileges to or the configured , and the file system permissions are not set up correctly for that as shown in\nFYI this is how I'm handling the situation with : The sysadmin should override and to set the master process user and group, while the and options set the effective user and group.\nThe way the tmpfile rules are currently written is that the group doesn't have write, only read access to . So it doesn't really work well with running the nginx master process as root, and then drop permissions to a regular user. I assume you also don't want to run worker processes as root either :laughing:… I'm not sure how to properly approach this problem. Can you elaborate a bit on what's your usecase as running the master process as root?\nare you able to test this configuration out and report back: Please take special consideration to look over your directory (under ) and the permissions there. I'm not familiar on what user/group permissions the process assigns to log files when it runs with as the master process... possibly a minor tweak to the will be required, but let me know what happens with your test. If this works we need to simply add documentation on how to run with as the master process, as I have explained above.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nJust from the go/no-go meeting, I approved of changes being backported to 20.03\nI've gone ahead and tested the above configuration snippet. I had to make a few tweaks and ended up with this as the required configuration to get to run with as the master process but as the child processes: can you please describe your requirement to run the master process as ? I'm not familiar with so I have a hard time judging whether your case is an edge case or not. depending on the answer from we need to make a decision about whether we need to simply improve documentation, or improve documentation and modify the module code to make this use case better supported. I'm currently leaning toward this being a documentation only thing, but let's wait to hear what has to say.\nI think your suggestion is good, I will try it today for our setup and report back. If it works, then I agree that it is sufficient to document your approach in the release notes (and perhaps as a description in one of the config options (probably the one for is best). On my use case, it is: being able to configure nginx the same way it was running on 19.09 for smoother migration, also so that one can upgrade to 20.03 already before fixing all the other things one's nginx uses that may rely on the master running as root\nNixOS does not complain when options on non-existent systemd services are set (should it?) -- that took a while to figure out :grin:\nI have checked your approach now on my staging deployment; it works (with the typo removed). What we should add to the changelog:\nglad to hear. Thanks! :tada:\nPR: We should cherry-pick it on after merge.\nThanks for finding the best solution!"}
{"_id":"q-en-nixpkgs-a6ddc4591101cf05326d400c8b9a23c334c6ca801c7fbfa36cd396c0c8aa16dc","text":"Project description An advanced menu editor that provides modern features in a clean, easy-to-use interface. MenuLibre is developed for users who want to manage their menus without ever needing to read the spec or use a text editor. Budgie, Cinnamon, GNOME, KDE (Plasma), LXDE, LXQt, MATE, Pantheon, Unity, and Xfce are all supported. Metadata homepage URL: source URL: license: gpl3 platforms: linux\nThis project now managed on GitHub: Code: Bugs: Translations: Wiki: Homepage:\nThere's a module helper function that can generate files. Not as easy to use as menulibre of course, but reproducible.\nThank you. I edited the description for it to be accurate\nI marked this as stale due to inactivity. &rarr;\nWell, it would be good to have a formula for this.\nI marked this as stale due to inactivity. &rarr;\nThis would be really useful. I would write a derivation but I'm not confident in my Nix skills to do so."}
{"_id":"q-en-nixpkgs-c7d8544dd41ac9e4ce5e6b5611e89e61994202a2458e0598f414d8f48ed6d57d","text":"Project description Code Intelligence for Vala Metadata homepage URL: source URL: license: LGPL2.1 platforms: unix, linux, darwin, ...\nany relation to\nThere are some efforts to merge the two servers but otherwise they are independent.\nHello, I'm a bot and I thank you in the name of the community for opening this issue. To help our human contributors focus on the most-relevant reports, I check up on old issues to see if they're still relevant. This issue has had no activity for 180 days, and so I marked it as stale, but you can rest assured it will never be closed by a non-human. The community would appreciate your effort in checking if the issue is still valid. If it isn't, please close it. If the issue persists, and you'd like to remove the stale label, you simply need to leave a comment. Your comment can be as simple as \"still important to me\". If you'd like it to get more attention, you can ask for help by searching for maintainers and people that previously touched related code and @ mention them in a comment. You can use or on the relevant files to find them. Lastly, you can always ask for help at or at ."}
{"_id":"q-en-nixpkgs-1b523e982678c75fdedb379837c4f3bde71cf31db80f73623b3e71d6c72a61b9","text":"Describe the bug addon in chromium does not work. After install addon show errors: \"Failed to connect to the main process.\" \"Verify the installation of the plasma-browser-integration package. The running environment of the Plasma desktop environment must be at least version 5.13.\" \"Specified native messaging host not found.\" Manually run from console not help. To Reproduce Steps to reproduce the behavior: Add to : Expected behavior No errors. Additional context I found issue Everything looks like the same problem but in chromium Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Metadata Please run nix-shell -p nix-info --run \"nix-info -m\" and paste the result. system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:\nTry adding the following line to your configuration:\nnow And also does not work. Maybe I need to reboot?\nThat issue is actually unrelated, it has to do with a patch only in our source build of firefox.\nI make test with: in . The same error.\nI can't reproduce this issue. I just tested in a vm using with this configuration: Are you using plasma5?\nYes plasma5. Looks like it worked for me too. Now all errors are gone. Thanks !\nWhy are we adding to systemPackages when using the chromium program module? It does this for you.\nJust an oversight on my part. You're right, adding chromium to systemPackages is not necessary.\nHello, I'm a bot and I thank you in the name of the community for opening this issue. To help our human contributors focus on the most-relevant reports, I check up on old issues to see if they're still relevant. This issue has had no activity for 180 days, and so I marked it as stale, but you can rest assured it will never be closed by a non-human. The community would appreciate your effort in checking if the issue is still valid. If it isn't, please close it. If the issue persists, and you'd like to remove the stale label, you simply need to leave a comment. Your comment can be as simple as \"still important to me\". If you'd like it to get more attention, you can ask for help by searching for maintainers and people that previously touched related code and @ mention them in a comment. You can use or on the relevant files to find them. Lastly, you can always ask for help at or at .\nThanks - the fix worked for me! Can you please comment on what's the role of package that you install?\nI marked this as stale due to inactivity. &rarr;\nProposed fix in"}
{"_id":"q-en-nixpkgs-3ee044724fce8a7ab32114cbbcfd28c6473ebf3d1af97be169cef606ab7e1be3","text":"To Reproduce Additional context I am building the package on an x8664 machine via distributed build, if that makes any difference. I have also tried to build with llvmPackages8 through 10 and I got the same result. Notify maintainers Metadata Maintainer information:\nThis looks just like Unsure why it's reoccurring.\nCaused by\nThis is native linux i868? Interesting, it is deciding to build nothing but headers, and then only the dev output is present. I am curious what it decided to build before that filled .\nAh! I bet it was auto-detecting and that gave it something to build, but now the correct arch is manually specified.\nI can confirm that reverting fixed this issue.\nhopefully fixed in\nI'm seeing something similar for 32-bit arm on recent master () , ([build log][]). Do we need to make a version of for ? [build log]:\nI suspect that since (and previously ) on i686 is really only building the bare minimum: whereas on : Missing are asan, ubsan related libs and much more. I should probably open a new issue for this."}
{"_id":"q-en-nixpkgs-8e67be1ff4332f8cc75dc459cbc61a505ab948879275a6d473081b24b7124678","text":"Project description is already packaged in nixpkgs, but the upstream project has been taken over by new maintainers who have ported it to Python 3 and are continuing development. As far as I know this new version is also used by Debian/Ubuntu (it has python3 dependencies in those distros). I have a local attempt at a new derivation and it Works For Me, but it has a patch in it which I don't think it should need. Also, maybe the original maintainer may want to do the update? If desired I can open a PR with my current state of affairs. Metadata homepage URL: license: GPL2 platforms: Linux\nI have limited spare time these days (months), so please be my guest"}
{"_id":"q-en-nixpkgs-b686b40256030217326ca88302270af6d108fbd905e55743e16a91004b4593c5","text":"Describe the bug Edit: After looking at it for a while, it seems the real problem is that this function is too strict: I think it should be something like this instead: To Reproduce Because comes before alphabetically, it's mounted first (which fails). You can read the file using: Expected behavior The list of file systems to be topologically sorted. Additional context Originally found this while trying to try out root-on-tmpfs. Notify maintainers Couldn't find anyone. <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Metadata system: host os: multi-user?: sandbox: version: channels(root): channels(jakob): nixpkgs: Maintainer information:\nThis is also a problem when using since the dependencies are specified in not or and for this problem is not enough to fix the issue. The option sugested by in would solve it. Actually the only reason the current ISO and netboot images aren't broken because of this (they use ) is solely because and comes alphabetically before .\nI marked this as stale due to inactivity. &rarr;\nThis is still an issue, and for me a fairly big one, since I can't generate secrets with agenix with my current setup:"}
{"_id":"q-en-nixpkgs-74354b926635d24da635297d2857f55e9fd1a3c279e90c1fc2644ad56fbbd4d0","text":"Describe the bug When receiving email notifications from , for example after a scrub, the hostname is missing from the subject and from the body of the email: To Reproduce Steps to reproduce the behavior: up ZFS and zed ( and ) a scrub for the email Expected behavior Hostname should be present in the subject and body of the email. Additional context Try adding the following in your configuration and notice how the returns nothing: Notify maintainers Metadata Please run and paste the result. Maintainer information:\nI just noticed in my logs: So the environment is probably missing something for the command to be available.\nRequires net-tools\nPlease make a PR to add it to the ZED path as we did for the other tools required in ZED.\nI will but I have never committed to this repo so it may take time before I learn what is required.\nAlso, it would have been nice if such a problem would not have been silent, maybe by using in bash scripts.\nyou need to add it here: You can rebuild your own NixOS using the instructions from here:\nI submitted the PR"}
{"_id":"q-en-nixpkgs-75154e4f54d432a81a0f4864735f0315310b15d02ebcf620ae5bd0f368b42393","text":"Describe the bug does not start up. To Reproduce Install and run in the terminal. Expected behavior A TypeScript REPL should start up. Additional context It's possible that the peer dependency on TypeScript is not handled properly. Notify maintainers Metadata Please run and paste the result.\nI am not sure how to fix this as it is not a dependency of the tsun package, it is defined as a peer dependency and this expression is generated from a simple JSON file. Of course it should be fixable with overriding and specifying typescript as a dependency, but I don't see any other packages that have these kind of overrides, thus I am not sure how to do it for node packages. I know some other generated packages have some overrides that adds missing dependencies/applies patches and things like that. For now you should be able to include both packages (typescript and tsun) to get it to work.\nI've tried , but this yields the same error message as above, which I produced with my main , where I have Is there a more idiomatic way of installing , either locally or globally?\nAh, it is a bit confusing, but seems you need nodejs as well, this works for me: It seems without nodejs it won't setup the which is used for the modules to be found.\nIn this works for me. But when I have all this in my and install a new generation, fails with the same error as before. I've tried having in but then I get after installing from .\nIt is the that needs to be set, so this would work: I see now that there are other overrides in the so I should update this package with this code there.\nThe correct fix might also be this: As the reason why it is not included by the autogenerated expression is due to it being part of peerDependencies which is not included by default in the node2nix command."}
{"_id":"q-en-nixpkgs-3251bd772458f22bc638db7055204754746b3f9e7688f3d5e149b28c97dc47d5","text":"To reproduce: create new (ethernet) connection, with ip4 set to link-local. When ing the connection, the following errors are thrown: An reveals the problem: I thought the following would fix it (): but this creates a new problem:\nadd that new user when NM is enabled in the nixos module.\nDid suggestion work? I was just bitten by this issue.\nDid not try it futher yet. If you want you can try it."}
{"_id":"q-en-nixpkgs-ea24907e58721fb155a69f516505eb517708afb155b15d3e8d0fda691000179c","text":"The metasploit package no longer builds. I've updating it myself, but I haven't been very successful. If someone could point out my mistake I would gladly correct it and submit a PR.\nIs this issue still relevant? , (pinging you because you are able to close this if necessary)\nLooks relevant to me, as there was no commit to metasploit that updates or fixes it. Why do you think I isn't relevant anymore?\nCurrent version: 3.3.1 Last stable version: 4.12 We definitely need a maintainer. is there a tag for ?"}
{"_id":"q-en-nixpkgs-c9098bd77768f76b0b2c41f4d626f3f0490d0e61bbb2476fc4ab837254c169f0","text":"Project description This flag is available in most other major distros. It is necessary to enable screen sharing on wayland inside of chromium, combined with xlr-desktop-portal-wlr and pipewire. It requires the build flag and a dependency on libpipewire02. See as an example. Metadata homepage URL: source URL: license: mit, bsd, gpl2+ , ... platforms: unix, linux, darwin, ...\nthat would definitely be nice to have :) I actually also noticed this but unfortunately cannot test this easily as I'm not using screen sharing. If you could contribute a PR that would be awesome (but be aware that Chromium does unfortunately take quite some time to compile). Alternatively (but less preferably) we could make the required changes with the next update and test the core functionality (as always) and you could test screen sharing after the new version is available in the binary caches. Additionally we would need to package in Nixpkgs to make this work on wlroots based compositors. On Gnome/KDE this might work out-of-the-box, but I don't know what they currently implement/provide.\nwas just merged w/ . I'll try building chromium w/ the flag - as you say it takes a while, so I can leave it overnight and give it a shot.\nhow does the situation look like with ? Is the flag available there, as in, screensharing with wayland already works there?\nyes, it's available in google-chrome. ! I'm still trying to get it to work with xlr-desktop-portal-wlr - I've done this in the past, but it's not working out of the box yet with\nAnd just to make sure, you also did set as described there?\nyes I did. I'm almost got it working I think. There are a few issues out of the box: -chrome is missing the pipewire 0.2 runtime dependency, so I it as (I'm on 20.03 so pkgs.pipewire is 0.2): .pipewire service doesn't work - it doesn't seem to be setting up the path correctly. I'm trying to fix this now: My current config on 20.03 looks like: Where is the same package that was recently merged.\nyeah, I'm currently already looking at it - pipewire wants to execute , but it's not in . We can patch to contain the absolute path. I'll check if it works and will get back to you.\nBtw, we /really/ need a sway NixOS test, and better documentation on how to use sway. I was trying stuff until it worked, and does autologin me into a sway, but there might be a smarter way to do this, without also starting X.\nThe sway recommendation is to just run sway from the terminal after login - that's how I do it at least. I was able to workaround the environment issue by doing (but this is a kludge I think and shouldn't be necessary). But unfortunately it's now crashing and I'm not sure why:\nYeah, even my sddm config I also had to , otherwise pipewire complained about not being able to reach dbus: Afterwards, it segfaulted, as it wasn't able to find ; I patched to be able to find the path to , and that error message is gone. Now (apart from not being able to find bluez or rtkit) it /might/ only be some version incompatibilities, not sure: I'll see if I can get as far with your chromium PR, will merge that then, and also submit the pipewire patch.\nOkay, with the pipewire patch and , It seems I'm able to share at least my own application window :tada:. I don't know yet why the alacritty window nearby isn't shown, and why \"Entire Screen\" is black, but that's probably due to me running sway inside X via ssdm.\npipewire PRs:\nI don't think you're using pipewire in that case. xdg-desktop-portal-wlr only supports full screen sharing atm. If it is black that means it's not working. By the way I had to add pipewire02 to my system packages to prevent xdg-desktop-portal from crashing. But now there seems to be some incompatibility in the pipewire service logs when chrome tries to connect:\nyou should be able to cherry-pick the to not have to manually add this to $PATH.\nThanks! I'm suspecting google-chrome is compiled with a too old version of pipewire. Enabling pipewire debug logs, it seems to be sending an incompatible format to the pipewire server (similar to ). I have got this working with chromium compiled against pipewire 0.2.7 in the past, so I'm pretty sure it should work assuming the client library is recent enough. Once the chromium package has that flag enabled I'll try to retest.\nI just confirmed my last comment by building chromium locally w/ rtcusepipewire=true. It works perfectly for fullscreen sharing.\nHmm take that back - it worked the first time but not anymore :/. Not sure what lib is missing where.\nSeems like xdg-desktop-portal is crashing:\nThis might be yet another version incompatibility problem. I propose merging (as it's built with 0.2 everywhere else too), and then see if we can get it to work if the system pipewire also is 0.2. I asked upstream about these problems in - maybe 0.3 needs to support 0.2 clients for some time until we should switch.\nHeads-up: the bad client situation should have been fixed with the latest pipewire release, PR at\nThanks for your help! I'm still getting the xdg-desktop-portal crash, however. I'm guessing there is something missing from my dbus environment - or perhaps it's trying to access a file/directory not available in nixos.\ncould you open a new issue with some more details, like logs etc.?\nOkay, I made"}
{"_id":"q-en-nixpkgs-98248a9a5044ea22e10e216d13b821fff727806a81558c9913775e1d8b53c7ab","text":"Project description Cross platform chewing user phrase editor Chewing editor is a widely-used bopomofo input method for traditional Chinese, and chewing-editor is the tool to edit the database of user phrases stored by the input method. Users of chewing will need chewing-editor sooner or later. NixOS have chewing include as a fcitx engine () that can be used in the module, but not chewing-editor. Metadata homepage URL: source URL: license: gpl2 platforms: unix\nThe corresponding pull request is merged. Closing."}
{"_id":"q-en-nixpkgs-143c58b16a9aba94abc45ce9bac5deb14c6c735c892da6214019789e1052a7c5","text":"Describe the bug pitivi crashes with gtk errors on startup. To Reproduce Steps to reproduce the behavior: in terminal and press enter the error message: Additional context This happens at least on wayland/sway. Notify maintainers as he did the last change to the packet and no maintainer is defined. Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:\nI get a different error with nix on Ubuntu on commit : On on Ubuntu it works.\nPR in ."}
{"_id":"q-en-nixpkgs-e3f15f8d663cdd830fef99f01fd280bdec00500729c9d7b343bdc835f244952b","text":"Describe the bug cannot be built from current master due to failing tests. To Reproduce Additional context Relevant logs: Notify maintainers Metadata Please run and paste the result. Maintainer information:"}
{"_id":"q-en-nixpkgs-e8d6940c814d44bb05626d3bdfbf8a3b4252126edba580ffbaa2a2097d265702","text":"Describe the bug All last builds on Hydra have failed because of (top level package). E.g. Nr Status -- -- 1 Cached failure (log, raw, tail) (propagated from build ) First broken build 2020-06-05 19:14:56: Nr Status -- -- 1 Failed (log, raw, tail) (propagated from build ) Log: To Reproduce I can't reproduce the issue locally though, does compile. Expected behavior Hydra job should not fail on mariadb. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Metadata $ nix-shell -p nix-info --run \"nix-info -m\" system: host os: multi-user?: sandbox: version: channels(theo): channels(root): channels(arsleust): nixpkgs: Maintainer information:\nReproduction: are you sure you tried it on darwin? (That's what fails on Hydra.)\nProbably caused by , Darwin doesn't have .so 's right? cc"}
{"_id":"q-en-nixpkgs-9997c54473c62b577ea68a1ffba3eb522378a21818677f95bd8e41518a0bc778","text":"Describe the bug Size of aarch64 sd_image has suddenly blown up a lot. whereas on commit it was To Reproduce Hydra: - - This blocks the channel ATM.\nIt seems this is caused by my PR here: We switched to , but does not erase the original file, while does. I created PR which hopefully fixes this by passing to ."}
{"_id":"q-en-nixpkgs-17447c6fd0b7f6d1fd9812c221934cb1b99adbc1cbc122a5382c0012c918c41e","text":"Project description Caprine is an unofficial and privacy-focused Facebook Messenger app with many useful features. Metadata homepage URL: URL: mit platforms: linux, darwin\nI marked this as stale due to inactivity. &rarr;\nI'd love to have this, too. I tried to take a stab at it a while back (using Atom and other Electron apps as a base) but it kept core dumping. I'm going to have to give it another try soon.\nI have a binary package PR ( ) dangling for a long time. You may want to take a look."}
{"_id":"q-en-nixpkgs-7d334e580dd3d0d06803aeb6de15b33d8d68a9dbea8432f95299e9f2ce5a9d23","text":", CVSSv3=8.8 (nixos-20.03, nixos-unstable) Scanned versions: nixos-20.03: ; nixos-unstable: . May contain false positives. Cc\nFixed in"}
{"_id":"q-en-nixpkgs-d34925cb5595ee331c28d3a40347cea1e65dcec8d883721742dc0fe14a3cc1dc","text":", CVSSv3=7.5 (nixos-20.03) CVSSv3=7.5 (nixos-20.03) CVSSv3=5.3 (nixos-20.03) Scanned versions: nixos-20.03: . May contain false positives. Cc Cc\nCVE-2018-: imap_mail fixed in 7.2.14 and 7.3.0\nCVE-2020-7067: EBCDIC decode fixed in 7.2.30, 7.3.17 and 7.4.5\nCVE-2019-: fixed in 7.2.30, 7.3.17, 7.4.5\nmerged - closeme"}
{"_id":"q-en-nixpkgs-8e3c3a014ce8b0b9c512b0d50d2cc46fc98c636ee833e2c61b803c24aec5ef45","text":", CVSSv3=7.5 (nixos-20.03, nixos-unstable) Scanned versions: nixos-20.03: ; nixos-unstable: . May contain false positives. Cc Cc"}
{"_id":"q-en-nixpkgs-053d854a4c85f23c5e690704695d95855faadabedad49d97f43e11326e31716f","text":"Describe the bug lacks an option to exclude some packages. Expected behavior xfce service should provide an option like in GNOME3 service. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-240230eb82e56d42ec5077975a567d79ba80cc67f230eb827559739898c332cf","text":"borgmatic is simple, configuration-driven backup software for servers and workstations. Protect your files with client-side encryption. Backup your databases too. Monitor it all with integrated third-party services. Homepage: Source: License: GPL v3 () Platforms: UNIX, Linux, Darwin (I guess anywhere Python is supported)\nI marked this as stale due to inactivity. &rarr;\nUp!"}
{"_id":"q-en-nixpkgs-0f8a2ea87ca0cd96f63d6a233054fcc67e39f1a09e1593be0d3d78c935c36b2c","text":"Julia 1.5 is out, see: and we don't have it packaged. More over, see how defines our other Julias: Several issues: is pointing to a TLS version which is old IMO. Certainly there are breaking changes seen in , but since Nixpkgs is a somewhat rolling release distro, are we going to keep these older versions and default to them for ever? I've once tried to reach for the community to hear if there's a strict policy regarding such scenarios, but Anyway, I'd like to propose a change that will keep only the LTS version and the latest - currently 1.4, and perhaps make the attribute , point to the latest. Any thoughts from maintainers or users? cc\nAlso, now I noticed all of our Julias fail to build, e.g:\nI agree. And the 1.5 release candidates are out, so this will become the latest version in the near future. Last time I looked, the package built the same deps for all versions, including 0.7. In versions 1.0 and later, there are actually fewer dependencies (e.g. arpack and fftw have been dropped), so the package should be cleaned up to not include these unnecessary libraries in the build. I also think the default version should be the latest version (i.e., unless somebody explicitly installs julia-1-lts, they should get the latest).\nOK I'm glad I have your support. I'll let this issue stay open for a while and hopefully I'll be able to take care of this a month or so from now.\ncc\nBack when Julia tests were not so messed up, I think there was often 1 previous version. I guess with 1.3 needing overriding to no tests, it is fine to just have julia-1-lts and Julia 1.4 as julia once Julia 1.4 gets built fine with some tests.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nHaven't had a lot of time to work on julia in nixpkgs, so just posting this as a FYI. We used multiple versions of julia at work, but we use a custom derivation, similar to this: We also do not run the tests, but in general it has been working great. Previously we also used the binary builds patched with patchelf, as there were some differences in the structure of the results of the vanilla binary build provided by julia and what we got when building from source. We gave up on using 'system' libraries, and build the releases that julia provides, as julia is very specific in the versions for its dependencies (LLVM patches mostly).\nI tried to use that tarball in order to write a derivation from scratch. Here's my attempt: It fails with: Naturally I didn't further attempt to add curl or wget to the inputs, as that would fail in a sandbox as well. I've opened an issue upstream: complaining that that tarball doesn't really contain all dependencies.\nI ran with sandbox enabled, and the expression I linked with works for me. So I think it is fine to just add curl to the deps, as with the full tarball, it isn't actually used.\nThere is another approach to avoid running tests: using precompiled binary. I tried downloading the 1.4.2 binary from and then afterward ran on the binary, then I got a running Julia interpreter just like that.\nWe did the same for a while using code similar to this: Unfortunately, we actually sometimes need to patch Julia, so we switched to the julia expressions from source, that I posted earlier.\nThanks that expression works well thanks to . I'm testing it now.\nOK so after a long build, I got a working Julia derivation with these references: And this closure size: Which I must say it is satisfying. The question is, is this how we'd want it to be built? Using all of their dependencies and practically non of ours? Some of their dependencies are required to be patched, according to and they also specify that using the OS' libraries is not officially supported? Anyway I'd happy to get help on deciding this.\ndoes this lead to Julia tests passing? I think for every dependency they patch, it is OK to just use their patched version. I guess it could be nice to use a separate build unpatched (but using the exact same version sounds like a likely restriction), but whatever dependency does not work right away can be switched to bundled.\nI think it did, I've set and the build passed but I haven't inspected the log much, but here it is: Currently produces this log but there's no PR yet as I think there are many makeFlags that should probably be set as well. If anyone feels confident to open a PR with this feel free to ping me for a review. I agree with the idea, but what's bothering me is that it's not absolutely clear what versions are patched and which are not - the docs vs what's actually found in is not totally inline. Also, it seems that some of the patches are old, and were suggested upstream (to the libraries) but they weren't merged there for some reason. I'm still investigating, See My current target is to decide what libraries to use our version vs their version, with the preference to use our version, in order to hopefully reduce closure size due to sharing shared libraries. The idea is to decide this by checking whether the libraries have included the patch Julia suggests, and then whether the library's maintainer has released a new version that includes Julia's maintainers' patch, and whether we ship that new version.\ndoes this lead to Julia tests passing? I think this log is cut short before the interesting part…\nI've uploaded it here:\nHmmm. Reading the log does not convince me the tests were run… Pity.\nSince 1.5 was just released, the title should probably be updated to be 1.5.\nDownside with the patching approach, is users lose the ability to compile against MKL, which has HUGE performance increase (last I saw showed that MKL was ~4x faster at matrix multiplication!!). Would need to also package JuliaPro to get MKL. Of course, anything working nothing working, although would be a shame to lose compilation options.\nSays\nThanks garrison, I saw that, but I have a feeling it won't play nicely with Nix...although if we go patching route, perhaps could do that during buildPhase\nAre you concerned about specifically or all packages that use BinaryBuilder? There are a lot of them. More on this at\nAs I understand binary builder, that all happens in userspace, but modifies , which is in /nix/store\nThe installation instructions for do not seem to require root. My guess is that it stores the modified sysimage somewhere in . Am I wrong?\nI just tried it with Julia 1.4 installed with Nix, and as I suspected fails to build due to\nWe need to have a interface - that would support building binary packages, and make the interpreter aware of them. I wish I was a bit more motivated to do this, but upstream disappoints me regarding the humble task of compiling Julia as is, with the right selection of a few system dependencies - no response at all at & . I'm also dis-motivated to explore this aspect, since their build system is of plain makefiles, and they keep patches of dependencies from so long ago, without documenting whether they are needed / used or not...\nNix behind It is a shame to see julia's support in NixOS languish. It seems pretty clear to me that upstream is only interested in supporting the configuration obtained by cloning their repository and typing on a stock Linux distribution. I understand their perspective in many ways, as having julia and all of its packages \"just work\" requires a lot of care, and building with unsupported configurations is very likely to result in brokenness somewhere in the package ecosystem if not in julia itself. Julia is a bit special in this regard because of its emphasis on performance and also its need to support interfaces to other languages/frameworks that often have weird requirements. I have found that typically, there is very good reason for these patches, even though at times they may not be documented except in git and github history. (Admittedly, I am less involved in upstream than I used to be.)\nI think Julia team initially evaluated Nix as their binary distributor () but decided against it—there was a discussion somewhere but can’t find it. IIRC lack of Windows / desire to tie into native pkg management systems were factors. Julia’s does a pretty great job, and uses a solver to properly resolve dependency versions based on constraints. I’ve been incredibly frustrated by the withPackages pattern, at least in python. I constantly run into problems due to nixpkgs ignoring version constraints. There are some other nice advantages to nix of course: binary dependencies are shared across language ecosystems for example, and there are lots of supremely annoying packages (I’m looking at you CUDA...) that Nix makes delightfully easy to install, use, and maintain parallel versions. But until Nixpkgs supports multiple versions across the board, and has a solver baked in, I’m honestly not sure Nix brings enough advantages to bail on Edit: I think the approach done by poetry2nix and Mach-nix are the way to go for Julia packages + nix. yarn2nix is another shining example. I would love to see a solution in this vein—didn’t mean to douse any enthusiasm for people interested in helping improve the Julia situation on NixOS :)\nMaybe it is worth to reaching out to Julia Computing folks at to get their attention on ?\nWe'd probably need to put this information in our attribute sets, just like our Python packages know what deps they need. Hmm that's a bit surprising to hear as the Python ecosystem seems most well organized. Have you ever encountered a real bug with a mismatched version of a dependency with a specific version specified in a package's ? IIRC fails if a version is used and it's not exactly what's available. If you've ever felt there's a certain python package of a certain version missing, have you considered contributing it? Even if not, was it too hard to override it? What do you mean \"across the board\"? That might be much more harder then it sounds, due to the constraints of Nix it self when being evaluated everytime you run and friends, see It's not a question of whether Nix brings an advantage or not, it's a question of whether there are any Julia packages that are built at other distros imperatively with a , and this produces (or even worse, downloads) shared objects or executables into (not in ), linked to Julia or other Nix deps that do reside in . If the answer is yes, than every Nix user will experience breakage after Julia is updated in his - not necessarily if it's a version bump of Julia, it could be a mere rebuild of Julia by hydra. Yea Mach-nix is an awesome idea well implemented ( thanks for the inspiration :heart:). I've also reached a similar conclusion at . However, this is my opinion for Rust and Go, and I tend to think that for a small ecosystem like Julia, perhaps it'll be too much - it shouldn't be that hard to manage all of these pacakges like Python is managed in Nixpkgs.\nI opened\nFrequently. Especially with Scipy, matplotlib, bokeh, sklearn, tensorflow, pytorch, etc... Often, nixpkgs used to have it until it was updated. Except in very rare circumstances, we do not keep more than one version of a python package around. I have no interest wasting time working around bugs that authors tried to spare me by specifying precise versions. It seems a sort of cognitive dissonance in our next community that we all care so deeply about reproducible builds, but for convenience often disregard the officially supported dependencies. As a result, there are often subtle bugs in software on NixOS that do not appear on other platforms. (Eg x11vnc crashes, shared memory problems, etc). Nix supports crazy high specificity in versioning, well we past semantic versioning down the the bit in many cases. And yet in nixpkgs, by convention, we remove the old version anytime we update. It’s understandable that Hydra cannot build everything, but it’s always bothered me that I have to pin nixpkgs if I want a build to stay functional over 1+ years. is mostly declarative. Of course, Julia is constantly re-compiling code, and until there is true support for precompilation, I don’t see how it’s possible to use nix store. This is due to multiple dispatch. Suppose I wrote a library to do an efficient k-nearest-neighbor query using some fancy free structure. Perhaps the core algorithms simply require a function for your types. When you write this new function, Julia has to compile all the code in the k-nearest neighbor library for this new type. Thus, it is not possible right now to compile a package on its own as the core pattern in Julia involves extending the function of other packages via multiple dispatch. You can only compile a package in context of the code that is running. Therefore, the Nix store approach is not appropriate. See\nI see. You are talking about \"authors\" of Nixpkgs' expressions? I'm still a bit perplexed, as I earlier said, and correct me if I'm wrong: should fail if is used in , and the exact version is not available. Sometimes (), we do hardcode such version requirement changes in our expressions, but hopefully that's done only when it's justified. I agree (maybe a bit ). Generally I think, every dep's version related issue should be handled by someone like you, who's familiar with python packages you use. I'd at least (open PRs) to fix the issues I encounter :). I see. I guess the situation is more complex than I thought, but I don't think I'd be able to form a strong opinion until I'll get a Julia interpreter from an expression I'm satisfied with. Thanks for the link :).\nUnfortunately, this pattern is done constantly (including in my own contributions!). The problem is that, except an extremely rare cases like tensorflow, nixpkgs does not allow multiple versions of same python package to co-exist. Every time we update a package, we remove the previous version from nixpkgs. Confounding this issue, we often disable tests because sometimes the effort is just too high when developers bake in assumptions that aren’t compatible with the Nix ethos (for some reason this seems to happen more with tests then source code...). The most frequent cause of failure is the code needs to download something to run the test. (I’ve tried to lobby in the past that the test phase should be permitted to have Internet access as long as it doesn’t modify /nix/store, but I digress). In any case, the end result is that using a python package on Nix typically means you aren’t using the exact dependencies that upstream declared, and you’re more likely to have a package that despite having changes in dependencies has never had the test run. Thus, amazingly, I’ve had more luck with Julia packages working on NixOS than Python when all other binaries are present in environment.\nI don't think there's a strong policy in our Python ecosystem against multiple versions of packages - if this is what upstream recommends. Anyway, the whole topic is out of scope for this issue. Feel free to open a new issue and cc me and who's working hard to ship a 20.09 release.\nYes, sorry getting a little off topic here. I mainly mention it to suggest that our efforts in Julia that start from scratch Can learn from these issues. I’m not too familiar with the new binary builder, but was quite nice in that it would first search for libraries before attempting to install. See this issue for some context: Note that a core Julia dev gave a pretty great endorsement on Nix (albeit back in 2014): I think it might be most productive to do this development outside of nixpkgs, and do it in Julia to hook in with nixpkgs as needed. That way users could seamlessly use Pkg in Julia. I’d love to and have it just work instead of the current hack that I wrote that a lot of people seem to be using ()\nShort summary of the Discourse thread: for LLVM and libuv it is essential to use the patches/forks. For other packages it is probably unnecessary to use the exact version specified by upstream.\nThanks for being there with me :). I think I'm gonna work on this once we are through with , to avoid merge conflicts.\nTIL Guix manages to package Julia 1.4.1: What's interesting, is that they but they build it with Julia's LLVM patches fetched from the internet - they don't rely on Julia's build system to build it's own LLVM. For similar dependencies which the real system's version can't be used, they do the same. I think we should try to copy this approach.\nFWIW I did that in Fedora originally, but I stopped. Since LLVM makes breaking releases very quickly, you'll probably want to move to a new version that Julia doesn't support quite soon.\nMaybe I don't understand, but why should I move to an LLVM version that julia doesn't support? Also, aren't the and versions don't introduce breaking changes between minor releases? We could stick to since that's what Guix does and I think Julia uses this version as a base too.\nI don't really know how NixOS works, hence the \"FWIW\". If you used versioned LLVM packages and keep older major versions then it's probably OK.\nJust offering a user perspective: Julia provides a package manager that provides mostly reproducible (in a practical/scientific reproducibility sense rather than a bit-identical sense) environments easily from manifests. Increasingly, these environments are isolated from the host OS by binary packages for dependencies being served by Julia. I don't want Nix to manage dependencies that can handle for me: Julia already handles those excellently and conveniently. I just want Nix to handle the (usually few) system dependencies. What I would like as a NixOS user is for Nix to provide me with a working Julia 1.5 derivation that provides: julia, a FHS wrapper or something providing compatible versions of the few things that Julia's binary packages need (just libc, libstdc++, libgfortran; I think. ). When I say \"compatible\", I mean that these libraries should work with the binaries and with the julia binary provided by Nix. Not all Julia packages are BinaryBuilder-ised yet, so I'd also want an easy way to specify that I want compatible versions of other packages installed (e.g. Qt, cairo, etc). For packages that need python, I don't know if I'd like Julia or Nix to handle installing and managing that. The simplest approach would be to tell Julia to do it (there is pretty good support in Julia for distributing isolated environments of python + packages + dependencies; better than in , imo). I don't know how doable any of that is in Nix, just describing what a good outcome would look like to me.\nI think this is quite doable. I would propose that we add a new package, , that is a \"batteries included\" build that contains system libraries for all common Julia packages. Would this approach work for BinaryBuilder? I haven't used this yet. If not, we could fall back to a FHSuserenv, but I'm reluctant to do so if at all avoidable as there are outstanding nix bugs () that create seg faults in some heavy data processing workflows.\nI think that is a good proposal and I think it will work with BinaryBuilder packages (e.g. SDL2_jll). In any case, some approach that works with BinaryBuilder packages should be found because those are a critical part of the Julia environment now :) Some important packages (e.g. GR) still want to build and install their own software dynamically, and that doesn't work well in Nix, but most of those can plausibly move to the BinaryBuilder system (PR for GR ), so I think letting those be broken would be tolerable (and a big improvement over the status quo of Julia on nix). Perhaps we can get something working quickly by just patchelf'ing the official Julia builds for now? (I don't understand how all of this stuff works!)\nIs there anything presently blocking just getting julia 1.5 into nixpkgs? I don't care about ecosystem working with nixpkgs, just a good ol' fashioned julia install.\nyes, see\nIf you just want Julia and don't mind about binary packages breaking you can just download Julia from the official site and patchelf it. On Sun, 4 Oct 2020, 09:32 Tyler Benster, <:\nOn the there was also the idea floated to patch to patchelf dynamically loaded binary dependencies.\nGuix has a Julia package and it works quite well. Perhaps some ideas can be borrowed from there: It'd be really useful to get Julia working on Nix. It's the only big outstanding thing that doesn't work and prevents me from migrating some servers in my organization.\nI tried to copy their expressions in and it doesn't work yet, but there's also another seemingly working attempt by at .\nSo, do we want anything else now that 1.5.3 is packaged (:greenheart: ? I think we: 't need clean up the expressions a bit - not use the file if it's not shared. I did (2) along with the julia_10 fix in As for (1) I will need a bit of feedback.\nI guess some people like me still use julia_13 via test-disabling overrides, so a bit of time after 1.5 reaches channels for smooth transition could make sense?\nMaybe backport 1.5 to 20.09?\nOK so I disabled the tests for julia13 in and made the default there as well. All merged pull requests referenced here should be backported."}
{"_id":"q-en-nixpkgs-c21b394ebf31b86e5786cce424b48c9cdfc687cffa7b4dbfff0f24ee7318c03c","text":"I would like to use to encrypt my home directory. is already packaged (as , for some reason) What's missing is the PAM setup (). Given that exists for EcryptFS, I would like to request support for via the option, or something similar. I am aware that EcryptFS is available, to achieve the same goals, however it seems to be considered deprecated by other distributions (, ) and has some which I am not comfortable with.\nThis is just to say hi and to share that I too am interested in this feature\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-da80197d21139924e57021df01721212c8d42f04bae06a3c157a33805f5998bb","text":"Project description homepage URL: https://xd- source URL: license: mit platforms: linux\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-4fdbd1496da177ea600e60b819cd5aed3814d746666ede483231ec2fde6cf4ac","text":"Describe the bug arrow-cpp cannot be installed on Darwin To Reproduce Steps to reproduce the behavior: -env -i arrow-cpp Expected behavior Successful installation Additional context nix-build -A arrow-cpp from nixpkgs master fails with this error: Notify maintainers Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(lclaude): nixpkgs: Maintainer information:"}
{"_id":"q-en-nixpkgs-ff7b26f229ec22629feef5d8d699a8c2aea5ff26ba56eb87047b30f330d230f9","text":"Describe the bug Putting a valid certificate name in will cause nixos-rebuild to fail. To Reproduce Steps to reproduce the behavior: this line in . This command will fail. Expected behavior finishes sucessfully. Additional context This error message is related to : I guess this error is caused by switching to python 3. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Metadata Please run and paste the result. Maintainer information:\nisn't compatible with Python 3, it seems. We'll need to update or carry a patch.\nMRTC:"}
{"_id":"q-en-nixpkgs-f2a96f6056fa57cb2f97675603fde5ea3aa8761850f86c7267ed867e5c4757ee","text":"Describe the bug doesn't run on Darwin, due to some Perl / shebang errors. To Reproduce Steps to reproduce the behavior: Darwin, run the only produced binary , ie. following output is produced: Expected behavior The executable should output its help text when run without arguments: Additional context I managed to get it running using a wrapper script that runs . Notify maintainers Metadata Maintainer information:"}
{"_id":"q-en-nixpkgs-964491dc6f9aa316a0f1d83f8970f6ffa305cdae29c094f043350c7a39d617bc","text":"Describe the bug Bisect indicates as the likely point of breakage. cc To Reproduce\nI can't reproduce the error hydra has, it fails earlier for me on current master:\nWe've had a staging-next merge in the meantime; this symlink problem was fixed by .\nLocal build works with my changes in . Should this be tested via ofborg? If so how do I do this (can I even do this)?"}
{"_id":"q-en-nixpkgs-1cb2ca66a502b860d7a0326b8398a741ca66318f139e65bb64034338ebf31cf6","text":"Describe the bug Recently I noticed that the Electron applications I commonly used ceased to be able to connect to the internet. Spotify, Discord, Slack cannot connect anymore. To Reproduce Steps to reproduce the behavior: latest nixpkgs channel Slack, Spotify or Discord cannot connect to their respective servers Expected behavior They should be able to connect to the internet. Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs:\nGot a similar issue, but Discord and Spotify works. For me it's slack-specific. EDIT: Nevermind looks like I am getting only slack from unstable channel :facepalm:\nI don’t think it is Electron specific. I see the same issue with qutebrowser\nI confirm I can reproduce it with . Might be WebKit related ?\nI have issues with qutebrowser aswell, with every https:// site, http:// works. Getting errors like this: But my qutebrowser uses QtWebEngine, not WebKit afaik. I've rolled back so can't double check.\nI've been able to track back the issue to the commit Before this commit it works perfectly. EDIT: It seems that there is an update related to in this commit (see ) I believe that could be related to this specific change.\nInterestingly, it seems that not all Electron apps are affected. Signal still works, for instance, but Discord and Slack don't.\nThis also affects Bitwarden and potentially Steam. Temporarily using nss344 for the nss package in Bitwarden & Discord works.\nSteam indeed doesn't work.\ncc\nFor reference, so this (hopefully) shows up in search results: ERR_FAILED is the error code qutebrowser shows.\nI'm using the canary release and it's working, unfortunately it demands an update from 0.0.104 to 0.0.105\nSlack also seems to work with , but not\nI see that and had some similar issues with other programs while handling , they fixed it by simply use a previous version of for those programs. Maybe we should do the same ? I'm not completely sure this is the right solution since the list of broken programs seems to be quite long.\nRambox (electron) also seems to be affected\nProbably switch the default to the older version? At least that's what this sounds like at a quick look. (i.e. only use the new one for the new Firefox and Thunderbird branches)\nThanks I'm not sure how to achieve this, I will check it out later but if somebody with the knowledge can help that would be really cool :).\nSo much for \"NSS 3.54 shared libraries are backward compatible with all older NSS 3.x shared libraries. A program linked with older NSS 3.x shared libraries will work with NSS 3.54 shared libraries without recompiling or relinking. Furthermore, applications that restrict their use of NSS APIs to the functions listed in NSS Public Functions will remain compatible with future versions of the NSS shared libraries.\" I guess maybe all of this stuff isn't \"restricting their use of NSS APIs to the functions listed in NSS Public Functions\". We can switch the default to 3.44.x, but I'm not sure if that will be supported after Firefox ESR 68 is EOL, which will be in a month iirc.\nA temp workaround for now is to chuck something like the following in an overlay to build the derivation with the older 3.44 NSS (using discord as an example here):\nAn alternative workaround that I'm currently using is to replace with in .\nArch has discord (and everything else) using and I'm assuming it works\nWe use , but I believe this is unrelated ?\nis 3.55, if you want to give that a go. Probably doesn't help much, though.\nIm aware we're on just wondering if it's an issue affecting other releases on other distros, or a exclusive thing They also appear to have been using for a month too:\nMaybe we broke something in the makefile -gyp based build system transition. Apparently that all came down to the default make target being wrong, so what you could do is revert that whole change, append to the makeFlags, update to a newer release, rebuild everything and see if that fixes anything.\nSteam, Discord and Bitwarden stopped working for me.\nSlack stopped working for me as well. I don't use Steam, Discord, or Bitwarden (the Linux app-- I use it in the browser only). However, using fixed it, as per Thanks!\nSo for now?\nNot sure if related but dropbox doesn't currently work for me.\nThis might be related, but vscode doesn't connect to the network either. I tried overriding it with nss 3.44 but apparently it doesn't depend on it:\nVSCode is pulling nss from atomEnv which is it at On Wed, Jul 29, 2020 at 22:58, Matthew Levan <:\nGreat. So how would I go about succinctly overriding 's within the expression?\nMaybe an overlay for atomEnv. I'm not good with the nix language so I don't have any specifics.\nHow is everyone using ? It does not exist for me and it clearly does not exist in the stable repo.\nNo results found.\n!\nDoesn't show up on stable qutebrowser. ! Works on firefox but can't test unstable qutebrowser because it can't access the internet.\nWorks here on stable qutebrowser !\nAnyways to get back on topic here... How do overlays work and is this the appropriate quick fix for someone like me waiting for a PR to be merged into unstable? theGeekPirate <:\nI can't replicate it with so I have no idea what's going on.\nThe simplest solution is to use the stable version of whichever application(s) is/are causing you trouble until is merged.\nGreat, thank you. theGeekPirate <:\nFor someone that has many unstable packages that are broken, is an overlay the way to fix the issue for every conceivable package that depends on the troublesome package without having anonymous functions filled with overrides everywhere?\nMy Nix-fu isn't great, but I'm assuming there's a way to override nss to use that unstable package globally, instead. Hopefully someone else more knowledgeable can step in and help you.\nIn case anyone is wondering how to use a package from another channel, .\nThank you for the merge. I've renamed the issue so people that are searching with \"my [app] is not connecting anymore\" won't get confused by this quite specific nss issue.\nThank you all for finding the source of the problem fast.\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nThis issue has been mentioned on NixOS Discourse. There might be relevant details there:\nBoth channels have updated already to contain the fix."}
{"_id":"q-en-nixpkgs-017826ce4a4e49dc58adf01e2126d1ea8318ff7715f3f87cc1e81c931492a7dd","text":", CVSSv3=7.5 (nixos-20.03) CVSSv3=7.5 (nixos-20.03) Scanned versions: nixos-20.03: . May contain false positives. Cc Cc Cc\nBackporting update that resolves these ().\nFixed in 20.03 and master."}
{"_id":"q-en-nixpkgs-a56b468a42312ff95c533a7f257b02d97cb988de6d8f84373957da9af58d4a42","text":"Describe the bug xscreensaver-demo is both a way to preview individual screensavers and the frontend for configuring xscreensaver. It should be installed alongside all the other xscreensaver binaries; however, it isn't, although its man page is. To Reproduce Steps to reproduce the behavior: Expected behavior is available in . Notify maintainers (raskin) Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:"}
{"_id":"q-en-nixpkgs-8e6d6727bf7701785970a4e83e2c2edafedfcb5cd6fceeded8653bd4d7ed554a","text":"Dokuwiki login page does not work (timeout) when https is not enabled in the webserver. Steps to reproduce the behavior: Add some user using the usersFile option and disable and options. Click on \"Log In\" in the top right, fill in form, and click the login button. The page now redirects to a non-existing https url. Disabling these options should allow Dokuwiki to work on a non-HTTPS enabled webserver. Currently when trying to login, the page will time out as the wiki redirects the user to a https version of the url. While the default https enabled webserver is good for security, it is not always needed or wanted. For example, when you have the wiki on an internal trusted network with no access from the internet such as a wireguard VPN of a few personal devices. In that situation, the SSL adds no value and makes it harder to use, as you would need to distribute self signed certificates to all your devices."}
{"_id":"q-en-nixpkgs-7e2e6dcadeabe65c2ed7e80aedb212f90148891146a05a70deb236d2736c968c","text":"Describe the bug error: --- Error --- build-remote build of '/nix/store/x4nz3kqn24vlz257hm8mnsvrn6ggfsrs-v4l-utils-' on 'ssh://root failed: builder for '/nix/store/x4nz3kqn24vlz257hm8mnsvrn6ggfsrs-v4l-utils-' failed with exit code 2 error: --- Error\n-------------------------------------------------------- nix builder for '/nix/store/x4nz3kqn24vlz257hm8mnsvrn6ggfsrs-v4l-utils-' failed with exit code 1; last 10 log lines: ../../utils/libcecutil/cec- ../../include/linux/cec.h ../../utils/libcecutil/cec-htng.h ../../include/linux/cec-funcs.h ../../utils/libcecutil/cec-htng-funcs.h /nix/store/j8vysakw78bpgngba32hfwwikqda9yx2-bash-4.4-p23/bin/bash: ../../utils/libcecutil/cec- /usr/bin/perl: bad interpreter: No such file or directory make[3]: [Makefile:782: gen.intermediate] Error 126 make[3]: Leaving directory '/build/v4l-utils-1.20.0/utils/libcecutil' make[2]: [Makefile:482: all-recursive] Error 1 make[2]: Leaving directory '/build/v4l-utils-1.20.0/utils' make[1]: [Makefile:591: all-recursive] Error 1 make[1]: Leaving directory '/build/v4l-utils-1.20.0' make: [Makefile:518: all] Error 2 builder for '/nix/store/x4nz3kqn24vlz257hm8mnsvrn6ggfsrs-v4l-utils-' failed with exit code 2 error: --- Error\n-------------------------------------------------------- nix 1 dependencies of derivation '/nix/store/2544nk822aa38adwvdgk8kc1zyxf1wgd-system-' failed to build we need another . PR incoming... I'm on"}
{"_id":"q-en-nixpkgs-6075387a78b9041ee87f3234570a8f62893d3be7bf1192751090863070096f85","text":"Describe the bug as described in , the build fails with: To Reproduce Steps to reproduce the behavior: Since this is a dependency of openjdk14, reproduce with: Additional context This broke somewhere between and ; I'm bisecting now. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Metadata system: host os: multi-user?: sandbox: version: nixpkgs: Maintainer information:\nI could've sworn I tested this back when I was working on ... rebuilding from those commits now to confirm.\nYeah, somehow the release fixes this, but I don't know how."}
{"_id":"q-en-nixpkgs-ea5cc36448ede7e86895c0254863c211162117b60deb1e16dbb60e012079358a","text":"Describe the bug A clear and concise description of what the bug is. If I try to use the function in from the command line I get an error: To Reproduce Steps to reproduce the behavior: is printed to console Expected behavior A clear and concise description of what you expected to happen. should define so works. Screenshots If applicable, add screenshots to help explain your problem. Additional context Add any other context about the problem here. Currently NixOS modules (, ) handle this for the most part, but it would be nice to avoid duplication and have this defined once - in the package instead of modules. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. -- Metadata Please run and paste the result. system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:"}
{"_id":"q-en-nixpkgs-11862c58ecdf9f68e9ab680012bf52611e87269c02977f921e8327fa65a1c002","text":"Describe the bug peek fails to build on macOS To Reproduce Clone then Notify maintainers Log\nI marked this as stale due to inactivity. &rarr;\nsadly its not supported though getting keybinder3 to compile would be good\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-a2e22587dc92383699629f23eef8de255806e2d948fdb22f8bbcd56c1731e023","text":"Project description It's a tool to create pet-container using Podman. The container is fully mutable and maps your user home to the container. Metadata homepage URL: source URL: license: Apache-2.0 platforms: Linux\nI took a brief look at this a couple of months ago after it was rewritten in go. I'll post what I had if I can find it. It uses meson/ninja as a build system which makes it kind of non-standard and a bit awkward to package compared to other go projects, might be easier to skip it and specify the build/install phases manually. IIRC it also needed completion/profile.d paths to be patched in, had issues trying to mount other paths on NixOS and it wanted flatpak for some reason.\nI also took a shot at getting toolbox working a while back, and wrote a nix expression for building the package, but decided to try it out when podman is rolled out in a stable NixOS channel. Now that 20.09 is released, I'm trying it out but I get cryptic errors like the following: I might look into this further when I get the time, but people are welcome to look into this issue to try and fix it. Meanwhile, here's the nix expression to build it: Note that in order to use toolbox on NixOS, you seem to need the following config: People seem to be against usage of buildGoModule (see ), so if people are interested in seeing this in nixpkgs, someone probably needs to port this to buildGoPackage.\nHi , I modified a little your nix expression I a patch to check if exists (it not exists in my system for instance). I also changed to copy to and it's mapping this file inside the container. I'm having the same error. So I tried to start the container using podman, this way: Maybe the problem now is because o cgroups. I don't now.\nI had to submit an upstream patch to fix an error introduced in newer builds with ownership of the toolboxRuntimeDirectory. I think this would be good to have, so I'm interested in carrying on the work. I was able to build toolbox (thanks for the derivation). The biggest issue seems to be this. Trying to start a container with yields: See this issue. Seems that the reported issue occured when host and container glibc don't match. The toolbox binary gets bind-mounted into the fedora container, which may cause issues without the requisite store paths also being mounted. See: Toolbox is kinda leaky by design, so this might be a problem. <details<summaryContainer info </summaryWARN[0000] Error initializing configured OCI runtime kata: no valid executable found for OCI runtime kata: invalid argument DEBU[0000] using runtime \"/nix/store/jkg2hzkxc2g05gsbpkb8sybh34j78syh-nvidia-podman/bin/nvidia-container-runtime\" INFO[0000] Setting parallel job count to 73 DEBU[0000] Called inspect.PersistentPostRunE(/nix/store/h9was55f3ir6b2sq5vc96cpgns3qbfz0-podman-2.2.1/bin/podman --log-level debug inspect --format json --type container fedora-tool </details\nI have it running! Hello from vim installed via dnf. It's blocked by my upstream patch which you can follow here: <img width=\"1278\" alt=\"Screen Shot 2021-01-22 at 4 20 05 AM\" src=\"https://user-\"<img width=\"634\" alt=\"Screen Shot 2021-01-22 at 4 21 13 AM\" src=\"https://user-\"\nThat's great I think toolbox is a great tool to have in NixOS.\nAgreed! Please test the PR if you have time :) Thanks for your work!\nVery happy to see your interest in Toolbox! I just wanted to point out that Toolbox doesn't actually need Flatpak proper. It's just the daemon which is a D-Bus session service that comes from but is completely separate from the rest of Flatpak. In fact, in Fedora, we ship it is a separate isolated sub-package. The purpose behind this daemon is to keep certain configuration files in the container's synchronized with their counterparts in the host. eg., , , etc.. Anyway, good news is that these days, since Toolbox 0.0.97, the daemon is no longer used for newly created containers, because the functionality has been absorbed into Toolbox itself. One big reason was to enable rootful Toolbox containers (ie., ). Older, preexisting containers will continue to use the daemon, though.\nJust a minor note. The actual issue is See the related for an explanation on how we solved it. Ideas for better solutions welcome. :)\nI marked this as stale due to inactivity. &rarr;\nThis issue is important to me.\nAny update on this? Can this be merged to upstream and closed?\nIs there anything that you need from the Toolbx project?\nThere are two remaining roadblocks. To clarify, I no longer use nixos and will probably not work on this (I still am an active toolbox user, just on Fedora :)) glibc updates. Any time the hash of glibc changes in the nix store, all toolbox containers will break (and need to be recreated). I'm sure there is a clever way around this, either a service that runs after a nix-rebuild that updates the bind mounts, or patching toolbox to update the bind mount on each toolbox enter. These seem very nix (or guix) specific) SDDM issue, there's nothing I think toolbox should do about this. Nixpkgs could consider carrying the sddm patches until merged upstream. Anyone is free to takeover , I won't be continuing it. The PR \"works\" with the caveat of point 1 not being handled, so you'll have to periodically recreate toolboxes.\nafter the container creation, it's possible to change the bind mounts and ports? Is there a file to edit?\nI found this alternative to toolbox\nThere's no support for doing it easily through a Toolbx-specific configuration file at the moment. However, you can script something through one of the shell start-up scripts."}
{"_id":"q-en-nixpkgs-582f97b476ad8947a85ba29e365b88ab6e6ff8187ba137872d62552ae77f4e37","text":"Project description cchat-gtk appears to be a generic chat frontend cchat-discord appears to be a discord-capable backend gtkcord3 was the predecessor and was quite nice, I'm excited to get this packaged. Metadata homepage URL: source URL: license: gpl3 platforms: linux\nupdated OP to reflect that we should package cchat-gtk/cchat-discord now instead of gtkcord3.\nI marked this as stale due to inactivity. &rarr;\nstill important to me\nI have this packaged in my own config:\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-e369dd0dc9303e1339bd792a880bf9a9f3d850e2fe48d9148915a19295aea389","text":"Describe the bug It's easiest to see in this picture: ! To Reproduce Steps to reproduce the behavior: system with nixpkgs at how init complains about /dev/fd/62 Expected behavior No error pertaining to /dev/fd/62. Screenshots See above. Additional context I don't actually remember when this started, because the boot logging scrolls so fast I missed it a few times (and I might have updated nixos-unstable-small at some point). The problematic line is: According to the blame, that hasn't been touched in 5 years, so I don't know why this would start cropping up now. Notify maintainers <!-- Please @ people who are in the list of the offending package or module. If in doubt, check for whoever last touched something. --Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:\nWhy I think we should block on this: there should be no scary pointless outputs in the init script in nixos. Same if the activation script was being loud and scary for no reason.\nSo, I attempted to bisect this, but due to build failures (and the size of some builds) I was unable to pinpoint. However, I've now set up a machine that I can use for those huge builds, and will try to go further sometime soon. For anybody interested, below is my ( and ):\nOK, I re-bisected (by removing all the skipped commits and building them on a machine) and got this result: cc What are the chances this is an issue with the package/one of our patches and not with upstream? New bisect log, for the curious:\ndid this happen in a VM, or on your real machine as well? What's with the `chown: invalid user\" in there? Is this related?\nI can reproduce in a VM and on my machine. Ignore everything below the fd error -- screenshot was taken from another issue, before I was able to build a VM myself. (EDIT: I've uploaded my own screenshot sans the JSON and errors, to make things clearer.) I'm going to try to bisect systemd soon, but it'll probably take a while, especially given how many patches there are that could fail to apply :P\nOK, maybe the stuff is a red herring? I d the entire call and got the following log: <details<summarystrace output</summary</detailsCan't interpret this now, but maybe it will give somebody an idea. Most interesting part to me:\nAlright, thanks to my Google-fu, I found a diff that fixes the issue: According to , one of the articles that turned up, this might be caused by udev missing?"}
{"_id":"q-en-nixpkgs-e343aa7af631e8be150ece63f24883052fabd603cc832f17353af3be637aeccf","text":"Describe the bug When trying to start a dynamic playlist, the user is greeted by an error banner stating the following: To Reproduce Steps to reproduce the behavior: Cantata dynamic playlist to play dynamic playlist Expected behavior I expected Cantata to start the dynamic playlist. Screenshots ! Additional context This is likely fallout from perl's removal in However, this exposed something impure about Cantata's packaging, so it's a good thing! I tried simply patching the shebang to point to (as well as changing this condition to never return true: ), but to no avail. EDIT: Even adding to my does nothing -- when I click on the play button by the dynamic playlist, it turns to the red stop button, and then back to the play button without ever playing or queuing any songs. Seems like this might need more work. Notify maintainers Metadata system: host os: multi-user?: sandbox: version: channels(root): nixpkgs: Maintainer information:\nIf you have the time to give it go, I think this should do the trick: And then add to .\nThanks for the tip. It was slightly more involved than that, but fixed in this PR:"}
{"_id":"q-en-nixpkgs-cae653329cee70589ac1abd889ec2c8a6bbfa11f1e7f3ca3dfb934cfd77994f5","text":"Describe the bug It was left out being implemented in the initial PR , but it's a very useful feature for nixos development and endusers. I do understand it's nothing can't do, but it's a useful wrapper. cc\nSee for an attempt at adding support for this."}
{"_id":"q-en-nixpkgs-c6a5080795ada8614fa7625dc1ecff9144ec3f7b01ede0b31b6fafdd90257df5","text":"ping It was agreed that should be removed before . It was generally accepted that an option isn't a great idea and was to the module temporarily for ease of testing. It is important to remove this because other people have seen this option and started believing this is an accepted design pattern for NixOS modules - we don't want people getting the wrong idea and having more options like this slip in. can you please file a PR removing this option and the entry from the release notes?"}
{"_id":"q-en-nixpkgs-08945925a12e2f8aa47f671456344524326fda83c9d089d4ca6ccb7a90ed2874","text":"Describe the bug I am trying to install (cli version) on macos from the unstable channel but I a getting the following error: To Reproduce Steps to reproduce the behavior: Expected behavior joplin should be installed correctly Screenshots N/A Additional context Notify maintainers Unfortunaltely maintainers are not specified :( Metadata system: host os: multi-user?: sandbox: version: channels(dbarrosop): nixpkgs:\nI marked this as stale due to inactivity. &rarr;"}
{"_id":"q-en-nixpkgs-a3902d4ff4e5ac8373f8cbe0fbb4e7c701ee97e3e901d107f97bc4992d66afec","text":"Describe the bug The file has this function: Since this is called for every dependency, we get quadratic complexity. This may cause a non-trivial slowdown - building digikam I noticed it takes a few seconds just to get to , but there might be other causes. It also causes huge amounts of debug output when you add a to (about 600 MB for digikam). We should use an associative array for this like stdenv does (see ). Notify maintainers\nWe also need to do this in the setup hook for paths seen by .\nI marked this as stale due to inactivity. &rarr;\nNot stale.\nFixed in ."}
{"_id":"q-en-nixpkgs-750837f36f6294d741910c981ce27a6baf95ddf4ee062f15ecac0c961cc5562c","text":", CVSSv3=9.8 (nixos-20.03, nixos-20.09, nixos-unstable) Scanned versions: nixos-20.03: ; nixos-20.09: ; nixos-unstable: . Cc"}
{"_id":"q-en-nixpkgs-77e3b298b9ddd944cfbb0d11dbc8f12d7e5198833a7c2d1060230388f78429e8","text":"Describe the bug According to and , noto-fonts-cjk should have 14 ttc files, but noto-fonts-cjk in nixpkgs downloads from , this zip files doesn't include serif fonts. To Reproduce Steps to reproduce the behavior: -env -iA -list : family | grep 'Noto Serif CJK' Expected behavior It should output: Notify maintainers Metadata Please run and paste the result.\nThe nix file should download 14 .ttc files and install to target directory in a loop, it's a simple logic, but sorry I'm not familiar with Nix language yet. Mention more people involved in this file since 2020: hope this gets fixed soon, thanks!\nAs I recall, the issue with packaging it initially is there was no Super OTC version of the noto serif cjk font available. There still isn't a Super OTC in the repository. There is https://noto-website- from , but that URL doesn't look stable. I guess for now you could use , it is 99% the same font.\n, thanks for your quick response! Although is almost same, fontconfig can’t assure it will alway be selected for document using , there are a few other popular cjk font such as Arphic PR UMing & UKai, WenQuanYi, GNU freefont, GNU unifont. Like Debian, Gentoo also packages separate ttc files. FreeBSD even packages separate otf font, I guess this can save some memory and time for applications to load a font subset. It’s a pity google doesn’t upload a single super otc file for noto serif cjk, but I guess the huge noto sans zip is uploaded to git by mistake, it’s kept there just to not break download url. So I suggest just packaging 14 separate ttc file, it’s convenient enough compared to those separate otf files. Another reminder, it’s better to also include a fontconfig configuration snippet like Debian and Gentoo.\nArch also packages separate ttc and include a fontconfig configuration file.\nI marked this as stale due to inactivity. &rarr;\nStill important to me."}