text stringlengths 0 2.35k |
|---|
**Jerod Santo:** Right. |
**Kris Brandow:** It is a very consistent language. It has a nice set of features. It's kind of the opposite of German, or something, where it's just like "Why do I have nine different versions of "the"? It's kind of the opposite of that. |
**Jerod Santo:** Right. I think it's like the XKCD, there's too many specifications. What we need is one more specification that fixes all the problems with the other ones. And then there's one more specification... |
**Kris Brandow:** Yeah, yeah. |
**Jerod Santo:** N+1. |
**Kris Brandow:** That's where Log15 got its name. The popular Go logging package Log15 got its name from that, from the XKCD comic. |
**Jerod Santo:** Wow. |
**Kris Brandow:** That's why it's called Log15, because there are 14 specs; we need to make a new one, and now there are 15. |
**Jerod Santo:** Oh, it's the 15th logger. |
**Kris Brandow:** Yeah. |
**Jerod Santo:** That's cute, I like that. |
**Johnny Boursiquot:** That's a resume-driven project right there. \[laughter\] |
**Kris Brandow:** "I wrote a logger." Okay, well - I guess my opinion is not unpopular... Or I guess it's neither. |
**Jerod Santo:** We'll see. We'll have to vote it out. We'll have people -- I don't know if gophers are gonna... Is it a chance for gophers to dunk on Rust? It might go popular just because of that. |
**Kris Brandow:** Hm, yeah... |
**Ian Lopshire:** If I saw this on Twitter, I'd click "Yes, I agree", but I don't have a reason. \[laughter\] |
**Jerod Santo:** That's how most of the voting goes, I'm sure... |
**Kris Brandow:** I told my friend that I was talking to, I was like "This is gonna be my unpopular opinion in the next episode", so... I'm sticking to that. |
**Jerod Santo:** There you go, friend. |
**Kris Brandow:** Alright... Well, this has been a fun episode, hopefully not too meta for our listeners out there, and you've learned a bit about the myth of incremental progress... I feel like it's buried somewhere in the content of this episode... But yeah, so thank you for joining me, Johnny and Jerod, and welcome,... |
**Jerod Santo:** Thanks for having us. |
**Ian Lopshire:** Yeah, it was fun. |
• Strings.cut function for cutting strings in two |
• New IP address representation in Go (net.ip) developed by Tailscale |
• Comparison and performance advantages of net.ip over byte slices |
• Module workspaces (Michael Matloob's work on the Go Tools team) |
• Discussion about other features in Go 1.18 aside from fuzzing and generics |
• Backwards compatibility with API changes |
• New data type for improved performance |
• Helpers to switch between old and new types |
• vcs build stamping to track version control system information |
• buildinfo function to report dependencies and module versions |
• Support for multiple version control systems (Git, Mercurial, Bazaar, Subversion, Fossil) |
• Automatic inclusion of version control system metadata in binaries |
• The pronunciation of "1.18" and whether it is a decimal number |
• Changes to the `go fmt` tool, including its ability to format files in parallel and how it now works similarly to the tool with no space |
• Speed improvements when formatting large repositories |
• Formatting code and how some developers use editors to format files as they save them |
• Using formatting tools as a feedback loop for writing correct syntax and avoiding errors |
• The pacer redesign in the garbage collector, which is an area of interest but not an area of expertise for the speakers |
• The garbage collector (GC) pacer in Go is being redesigned to improve its performance in edge cases. |
• The original GC pacer was designed long ago and has accumulated quirks over time. |
• A redesign is underway to address these issues and improve the GC pacer's performance. |
• The issue number for more information on the redesign is #44167. |
• Go already had support for the M1 chip, but required additional work to get binaries working properly. |
• Newer x86-64 machines are also getting improvements in terms of binary optimization. |
• Go performance optimization techniques |
• ARM64 and x86-64 CPU architecture differences |
• Go AMD64 versions and their instruction sets (v1-v4) |
• Template functionality in Go (new features for control flow and short-circuiting) |
• Break and continue statements in templates |
• Go workspaces and multi-module repositories |
• Overuse of modules in projects and the new approach to using modules |
• The upcoming Go 1.18 release and its features |
• Introducing go.work files as a new concept in Go command |
• go.work files allow multiple modules to be used within a single workspace |
• go.work files have a similar syntax to go.mod files and include a "use" directive to specify directories and modules |
• Replaces can still be used, but using the go.work file is preferred |
• The introduction of go.work files requires changes in tools that interact with the Go command |
• There was an experiment phase where users could try out go.work files with a development version of the Go command |
• Discussion about code generation and reflection in Go programming |
• Opinions on code generation: some think it should be avoided due to added developer friction, increased build size, and build time |
• Bringing back the "try" proposal for error handling in Go |
• Discussion of why the try proposal was unpopular and potential issues with its implementation |
• The importance of proper error handling in Go code |
• Potential benefits of try, including encouraging better error handling practices and reducing the need for shortcuts like "if err != nil {return err;}" |
• Concerns about adding new features to Go due to backwards-compatibility promises and maintenance requirements |
• Package management and its stability |
• Challenges of contributing to JavaScript projects due to constantly evolving language features |
• Difficulty in creating a new JavaScript interpreter engine without significant resources (e.g., huge corporation) |
• Discussion of the limitations on individual developers pursuing their own JavaScript engines |
**Mat Ryer:** Hello, and welcome to Go Time. I'm Mat Ryer, and today we're talking about the other features in Go 1.18. Now listen here, you... Come here, come here. Go 1.18 has got two great, big features that everyone's talking about. I mean, everybody. All the popular people are talking about it. Everyone. Well, we'... |
In fact, we've got a new rule - if they are mentioned by anyone, even accidentally, unfortunately you will be immediately booted from the podcast. So please bear that in mind. No talk of those two subjects. Pinky promise? That's a pinky promise from me, and we'll find out if we're also gonna get a pinky promise from ou... |
**Daniel Martí:** Hello. Nice to be back, and nice to bring my technical problems along with me. |
**Mat Ryer:** \[laughs\] Your technical problems are, like you, always welcome, Daniel. Daniel's been using and contributing to Go for quite a few years now, and you've actually written a few tools as well, like the stricter `go fmt`, and what could be described as the opposite, a Go code obfuscator. They're interestin... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.