For people who heard about Zig just recently:
- Zig is not 2x faster than Rust, despite what recent benchmarks might lead you to believe.
- You won't find many Zig jobs for a few years still, despite the Tiobe stuff.
- Don't join to the Zig community just to rant about Rust.
The Zig project is globally distributed and "work from home" by necessity, while at the same time being one of the most prolific and impactful projects currently active.
But to be fair he was CEO of an advertisement company, can't expect him to know too much about technology.
Former Google CEO, Eric Schmidt, told Stanford students earlier this week that Open AI is beating Google in artificial intelligence because “Google decided that work-life balance and going home early and working from home was more important than winning”😬
People on HN are hating on Casey's video about performance while their company is spending double digits % of their revenue on AWS bills to serve 3 forms and 1 database view which took them 6 months to develop using the latest agile methodology.
It also crashes every 24h.
Please don't go around suggesting to projects to rewrite their code in Zig. It's cringe as fuck.
Yes, even if they're currently in the process of choosing a new language to rewrite the project in.
Advocating for mandatory full memory safety is like advocating for abstinence as the only form of contraception.
It would totally work if not for all the sexy intrusive data structures and slutty C APIs out there.
20% faster compiler boot times on windows by removing dependency on shell32.dll and ole32.dll
60% faster for builds of Zig sans LLVM, which currently is not that interesting but will be eventually once LLVM becomes optional
Most of the Rust drama that I know about is based on the fact that the project is at a point in its life where most people don't feel they can assume good intentions from the other party anymore when shit happens.
Ziglings has become so powerful that they don't need GitHub Stars anymore (they already found Yoshi at the top of the castle), so they moved to Codeberg & have a new fancy website.
Announcing the 2024 Zig Software Foundation financial report and fundraiser!
We're also going to have a Zig SHOWTIME episode where Andrew will present the 2024 roadmap for Zig!
Zig is the 6th most used language on Compiler Explorer (Godbolt).
Frankly, I find this stat significantly more interesting than the number of Zig-related questions on Stack Overflow.
Zig as a Multi-OS Build System
Another great interview with
@krisajenkins
, the best host I've ever had the pleasure to meet. Of particular interest for those who suffered the pain of failing to build C/C++ dependencies in the past.
Bun stages of denial
- It's not faster than node
- How can it be faster than industry standard tools w/ VC backing
- Ok it is faster but it doesn't matter because network latency (LMAO)
- How could it take this long for npm to fetch my deps
- *updates CV to add Bun next to Node*
Rule
#1
when organizing events: if you tell a speaker that their talk was accepted, you don't walk it back. If you tell a speaker that their talk is going to be the keynote, you don't walk that back either.
I haven't been organizing events for that long but even I know that.
Some people were complaining about unused variable errors, Andrew heard your complaints and added unmutated 'var' errors.
The breakages and hard errors will continue until morale improves.
So I have a local fork of Go that can build CGo projects that depend directly on Zig files. Maybe I should call it ZGo :^)
See screenshots attached: first one is a ZGo project, second one is the one-line diff to the Go compiler to make this work.
The best thing about Zig's async/await is that (once the compiler restores support for it) people won't have to stress over it: you want it? It's there; you don't want it? It won't annoy you.
If I ever see somebody become some kind of Zig extremist, I will do everything in my power to push that person away from the community.
Not going to look away while that person relentlessly attacks other languages because they don't fully espouse the same philosophy as Zig.
Want to improve as a programmer?
Stop staring at programming language benchmarks while debating which is the fastest with other people who have no idea how programs get executed by the machine.
Thanks to , Zig now has bindings to all win32 APIs based on Microsoft's projection metadata, and we're even listed on the official repository!
Awesome work!
Say what you will about the EU but now iPhones will have usbc ports and app sideloading thanks to EU regulations.
Legally mandated design improvements.
I warmly recommend checking out the results of the wasm4 game jam, it got lots of cool submissions.
It's also extremely pleasing to see how many how those were made in
@ziglang
. Hopefully one day I too will have the opportunity of trying my hand at making games :^)
Memory safe programs are nice, but correct programs are better. If you don't know what I mean, take a look at
@TigerBeetleDB
's design documents. Memory safety is a itty bitty tiny aspect of correctness.
It took a while, but with the addition of TLS and an HTTP client to the standard library, the floodgates to webdev in Zig are about to burst open. I guess I'll have to finish all the remaining TODOs to my Redis client soon.
A build of Zig's stdlib docs based on the new Autodoc implementation has been published on the official website. Still *a lot* to do, but you can now see the full API of ArrayList, AutoHashMap and other generic data structures. This was buggy before.
So, C got `defer`, but they went for Go's version.
In Go (and now C), defer runs at the end of the function. Place a defer in a loop's body and the language will have to allocate to keep track of everything.
In Zig, defer runs at the scope level, no allocation required, ever.
Nobody uses Zig in production and yet every other day I chat privately with people who say that their startup is using Zig one way or another. Oh well :^)
Every Open Source project is different, starting from the governance, up to the community. In this article I give you an overview of
@ziglang
, both as a project and a community to help you decide if we're the right fit for you.
People who think that Zig is a dumbed down language are the same ones that think that a professional barber's razor is also dumbed down because it only has one blade instead of 6, a swiveling head and a battery-powered vibration function.
People today don't know yet how much duplicated effort will be prevented by colorblind async in Zig.
One single, readable codebase for both blocking and evented I/O, instead of spawning the *-async version of every project.
How to make people stop using Discord: make a better chat app. It's that simple. No, don't be a "nice dev" and shame users for not using a chat application that better aligns to your ideology.
Just make a better one. It's that simple.
After getting a taste of fully instrumented fuzzing, I can't stop doing it. I feel such a bozo for having written unit tests for minor things in the past.
New blog post about how I think we can improve systems programming.
Fair warning: I'm not a hardcore systems programmer, but that's actually why I think I'm qualified to have this opinion.
There's been a bunch of discussion about a design flaw in Zig related to fn arguments (pass by value vs ptr).
Martin Wickham has given a great talk on the subject at SYCL23 delineating both the problem and how the core team plans to find a solution.
Introducing Oven: the company behind Bun.
Oven has raised $7m led by
@buckymoore
at Kleiner Perkins with participation from
@rauchg
@ycombinator
and more
The recording from Mitchell's talk is out! This is the first Zig SHOWTIME episode I've run in a long time, it was interesting going back to my Zig roots :^)
I've created a small Zig package to make it easier to get started fuzzing your Zig/C/C++ programs with AFL++.
Yes you can use this even if you only have C/C++ code!
Some people complain about Zig's @ syntax but man knowing without a doubt when something is a builtin vs an imported symbol (fully qualified btw) is amazing.
When I read code in other languages I always have myself questions that in Zig you simply don't need to ask yourself.
New post on my personal blog, co-authored with
@andy_kelley
, about the upcoming Zig self-hosted compiler that will feature a very nice quality-of-life improvement for iterative development: incremental compilation with in-place binary patching!
"Why I rewrote my Rust keyboard firmware in Zig" by
@lynaghk
"Perhaps, but Zig makes a compelling case that — at least for my pandemic-hobby-project keyboard firmware — I get by with far fewer concepts."
@ThePrimeagen
There's a big difference between the Rust project and the Foundation, with the latter doing sus stuff in my opinion. Too long a list for a tweet, but just to name one: look at what happened to the last ED & who controlled the Foundation in the meantime (chair + board seats)
Just received confirmation for a talk for the upcoming Zig meetup that might interest some of you:
"How Zig is used at Uber"
More info about the meetup:
A lot of the lust for metaprogramming that devs feel is because they subconsciously hope to save themselves some typing.
If you're an iron-smith you gotta get hammering, and if you're a software engineer you gotta get typing.
Reserve metaprogramming for when it's truly needed.
Not to complain, but most of "Zig needs better docs" comments are in fact "systems programming needs better learning materials" comments.
Zig does need better docs, but pointers, stack vs heap, C ABI, etc, are not exactly concepts unique to Zig.
That said, we'll get there :^)