text stringlengths 0 1.36k |
|---|
• Glitch's "raise your hand" feature for getting help with coding problems |
• Importance of having a support system in online education |
• Examples of automated tools and bots used to aid learning, such as RepoRobot |
• Problems with success and the need to address barriers to entry in programming |
• Discussion of Glitch as an educational tool that streamlines the development environment setup process |
• The importance of measuring the effectiveness of tools like Glitch in helping people learn and be successful |
• Comparison of frameworks and tools for developers, including Electron apps like Tad, and their ability to simplify tasks and provide offline functionality |
• Critique of other tools and platforms, such as Google Docs, for lacking offline capabilities |
• Discussion of Tad, a tool for simple CSV file manipulation |
• Origins of the name "Tad" discussed and debated |
• Showcase of Tad's features, including command-line interface and GitHub web page |
• Mikeal Rogers' announcement about leaving Node.js Foundation |
• Personal picks from participants, including Alex Sexton's self-promotion of Stripe products |
• Excitement about Stripe's new product and the designers involved |
• Acknowledgement of designers Benjamin De Cock and Philipp Antoni |
• Conclusion of the episode, thanking attendees and encouraging ratings |
• Mention of Tushy, a sponsor suggestion |
**Mikeal Rogers:** Hey, welcome to JS Party, where it's a party every week with JavaScript. I'm Mikeal Rogers... |
**Alex Sexton:** I'm Alex Sexton... |
**Jessica Lord:** I'm Jessica Lord |
**Mikeal Rogers:** Jessica is filling in for Rachel this week, because she's at the conference. It's awesome. |
**Jessica Lord:** I've already failed! \[laughter\] |
**Mikeal Rogers:** We're lucky enough to have her back. And our first topic today is the release of Node.js version 8 - not to be confused with V8 the engine - and to talk about that we've got a guest on, James Snell. Why don't you say hello, James? |
**James Snell:** Hey, how's it going, everybody? |
**Mikeal Rogers:** Alright... |
**Alex Sexton:** No one answered him... \[laughter\] |
**James Snell:** It's going well, James... Really well. \[laughter\] |
**Jessica Lord:** Doing well... |
**James Snell:** That's good! |
**Mikeal Rogers:** Yeah, so James is actually now the director of the TSC; he sits on the board of directors for the Node.js Foundation... But why don't you tell us a bit about what is in Node.js version 8? |
**James Snell:** Oh, there's lots of stuff in version 8. |
**Mikeal Rogers:** We're gonna need more details than that. \[laughter\] |
**Alex Sexton:** Alright, we'll see you next week. It was nice talking to you all. \[laughter\] |
**James Snell:** As I said, it's one of the largest major releases we've had in quite a while. |
**Alex Sexton:** I mean, 8 is at least two bigger than 6, and it's way more than 4. |
**James Snell:** Oh yeah, it's twice as big as four. I think there were somewhere around just over 150 semver major commits in this one, so it's actually pretty significant. A number of the high points is we have what should be the final API for async hooks. They're still experimental, but that's there... There is the ... |
We have some other things like WHATWG URL implementation; it's not officially supported, no longer experimental... And a number of other things, but those are probably the most significant new things. |
**Mikeal Rogers:** Dig into that N-API thing a little bit, because it's like behind a flag, and stuff like that. Tell us a bit more about that in detail. |
**James Snell:** N-API - the whole purpose of it is to allow native add-on developers to write to an API that is going to remain stable, that is guaranteed to remain stable across Node versions. |
\[03:58\] Previously, if you were writing a native add-on, you were writing to the V8 API, so you were writing to the native abstractions for Node, the nan tool. The problem with those is that every time V8 changes its API, you end up having to recompile. Every time you install a new version of Node - a new major versi... |
The new N-API is really designed to make it so that a native add-on can be written once and used across multiple versions of Node without recompiling on a single system. It's even possible to write a module that works both against the V8 version of Node and the Chakra Core version of Node... So not just different versi... |
The idea is to reduce maintenance costs and development costs over time, and make it significantly easier to actually build and deploy and use these native add-ons. It's pretty exciting. |
**Jessica Lord:** That's interesting... I have a question - maybe it's unrelated. I wonder if this will help Electron development at all. |
**James Snell:** It should. I'm not very familiar with what native modules Electron is using, but it should. Anything that interfaces with the native layer should be helped significantly by this. It should allow those modules to become more stable over time, and it should allow the upgrade experience to be far less irr... |
**Mikeal Rogers:** I think the big thing with Electron too isn't that it is using native modules, it's that when you use native modules in your Electron app, they need to be pointed at the version of Node that goes into the Electron app, not the one that you're using on the command line. So yeah, this will definitely h... |
**Jessica Lord:** But it's also that Electron sometimes breaks ABI because it has to patch Node to work with that V8, since V8's moving faster. I think that's also a problem. Maybe if ABI is getting more stable in general, maybe it helps everybody. |
**James Snell:** Yeah, that's the goal. The whole reason N-API exists is because a year ago we started taking a look at what we were gonna do with Node Chakra Core and how we were going to enable the module ecosystem to remain stable as we go through this effort of enabling other VMs. So it's just the problems of diffe... |
I think as we go down this path - it's still very early - it will make things like Electron development easier. We have ways to go. |
**Mikeal Rogers:** It says in the release notes here that you did a bunch of good stuff around Promises. Can you tell me a bit about that? |
**James Snell:** The key thing with Promises -- there's two things in particular. There's this new util.promisify() which will take generally any callback style async API - so if you think of like fs.readFile, right? |
**Alex Sexton:** That error-first thing... |
**James Snell:** Yeah, anything that takes the error first and then some set of arguments afterwards - it will take that and automatically wrap it in a function that is Promise-enabled. If you do util.promisify() fs.readFile, then you'll get a Promise back -- a function that returns a Promise back that you can use with... |
\[08:00\] It works probably for 80%-90% of the functions, the callback-receiving functions in Node. For the ones that don't quite fit the pattern, like fs.exists, for instance, that doesn't have that error-first pattern, there is a way of customizing the source function, so that you can customize the way that that Prom... |
The other thing we did with Promises is we made it so that for a native Promise it is now domain-aware. The domains module has been deprecated for a while, but it's still used; there's still lots of people that use it. Now inside your cache you can actually access the current domain; it's relatively minor, but it is im... |
**Mikeal Rogers:** Did that change go in, that everybody was freaking out about, about the throwing on uncaught rejection, that whole thing? |
**James Snell:** No, that has not landed yet. The reason for that is we still have not settled on any kind of consensus on what that action should be. The biggest thing that we've identified right now is that when an unhandled rejection occurs, that there should at least be a process warning emitted that has the origin... |
The challenge with anything that we do here is that the Promises model really leaves it up to the host environment to figure something out, and any path we go down doesn't quite fit well within the model. We either end up with a high number of false positives and rejections that are handled later, and we don't quite kn... |
So it's really difficult to figure those things out and see what we need to do there. At this point we've settled on "Let's not do anything yet, until we figure it out more." |
**Mikeal Rogers:** So pulling us out of the weeds a little bit here... I've seen a lot of people throw around this V8 TurboFan+Ignition thing in the compiler pipeline, and it gets thrown around a lot, and it's like a magic performance sauce that everybody's kind of throwing everything -- can you explain what this actua... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.