Zig is a nice language, but those articles always compare new languages against old-style C without use of modern tools. Most of the listed advantages you can have in C as well if you are willing to adopt modern tooling and style: signed overflow checking: sanitizers, errors you can not ignore: [[nodiscard]], slice type which know their size: variably-modified types or span types.
Defer will also appear in C compilers soon, but personally I still think goto-error handling is clearer.
(Rust has an advantage for temporal memory safety, but this does apply to Zig.)
It seems that part of what Zig is trying to do, though, is create a pit of success around these practices. So the response might be something like, “Yes you can do these things in modern C, but doing it consistently requires more discipline than we’d like.”
Yes sir, with the right kind of discipline even assembly starts to look like luxury. With a heart that holds no fear, one can code in straight 0's and 1's.\s
These days I wonder what is harder:
1. problems of collective action or
2. communicating that reducing everything to collective action is the real issue.
IMO you cannot fairly compare static analyzers/sanitizers with features baked into the language/compiler.
Though I've noticed the same trend when reading a famous Rust primer for C++ developers. Like there are obviously so many good and valid arguments when comparing the two languages that would've gotten the point across. Yet the authors felt the need to push it one step further, where the comparison no longer seems fair. Like comparing Rust with C++17 when C++20 provides the same exact feature that Rust claims to provide.
I'm very certain you can statically analyze zig's IR for safety without language change. And zig makes it easier to get consistency than C because you are steered away from the mishmash of malloc/jemalloc/pick-your-allocator-function. (The analyzer can say "all bets are off if you don't use the allocator interface") same with the new stdlib io interface for async-shaped analysis on aliasing analysis
The obvious answer is that I don't want to pay for these checks at runtime and can catch most of these bugs at test time. There are definitely use cases where that's acceptable.
As soon as you have dependencies you are dealing with however the authors decided to build things.
Part of why languages exist is just to get a new ecosystem.
At this point, you could fork C/C++, only remove features, give it a new name, a new package service, and maybe fork the standard library, and you could end up with something pretty competitive against Zig or Rust.
But Zig is expressly not trying to bootstrap a completely new ecosystem. Integrating with - and depending on - existing C and C++ code is a core use case. So much so that they’ve adopted “maintain it with Zig” as a motto that intentionally puns on “rewrite it in Rust”.
I don't see why this requires not having dependencies.
Yes, new languages try to create a new ecosystem. I think this is a bad thing - we already have too many of such - often somewhat isolated - language ecosystems. But I also do not think we need new ecosystems, when we can evolve the existing one. I have seen C evolve quite a bit over time, so the idea that we have to create new ecosystems to solve problems is wrong in my opinion. C evolves very slowly, which means that some things need more time, but I also think that it is generally a very good thing that it evolves slowly.
Zig's tooling is very good, so is the cross-compatibility story and tagged unions.
But while you can have some/most of the benefits Zig has from modern C, I think an underrated benefit is the language culture. As a different language it was the opportunity for a reset, e.g. it's idiomatic Zig to pass allocators to functions, meaning every library in Zig will work this way.
You could do that in C, but none or very few libraries do that so whenever you reach out to third-party dependencies (which is not a lot, in C as in Zig) you have to fight against C's traditional patterns.
Because some of those tools exist since 1979, and it keeps being a quixotic battle to make corporations adopt them as standard, until cybersecurity legislation started to take off.
I’ve settled on a dialect of C++ for systems programming after using C for a long time. It’s mostly C-like and I disable exceptions, RTTI, etc. but I do use templates, overloads, methods, lambdas, etc.
I have my own object system with constructor methods that all return a status code so you don’t need exceptions, my own stdlib for containers, buffered I/O, networking, hashing, math, threading, async, etc.
C++ gives you all the features to effectively build your own language, and you can use first class tooling since C++ is widely supported. LLMs also make it much easier to build out and test an entire stdlib on your own.
Agreed. I'd not ditch the STL completely, but only use its freestanding subset. Implementations contain some very fast SIMD-accelerated code[1]. +The STL works without exceptions.
I've essentially done much the same. One of the more under-rated features of C++ is being able to build a new language environment based on a different set of composable idioms from the low-level primitives on up.
C++ is as much a tool for building nice systems DSLs as it is a systems language. The heavy investment in compile-time expressiveness has really paid off for the language over time.
Also, Zig is as at least as expressive as C++ (but without the implicitness and general language complexity that makes reviewing some sensitive code, where you need to see exactly what operations are performed, hard), and it supports arenas better than C/C++/Rust.
Seems like a long way to walk though. Retrofitting this will take time (if it'll ever complete).
Zig made it a core part of the language and library. I'll assume Rust will make it farther than C++ ever made it, but just take a look at how retrofitting custom allocators is going there.
Seems like it will be useful for use-cases like embedded where you're restricted to certain parts of the library anyway. But the battle is kind of already lost for support at the level of Zig or Odin. Even if they retrofitted support for the whole standard library, they'll never be able retrofit all crates or make people even bother adding support for all new crates.
> Seems like it will be useful for use-cases like embedded where you're restricted to certain parts of the library anyway. But the battle is kind of already lost for support at the level of Zig or Odin.
Odin in particular does a simply terrible job here. Take the most common need for allocations, a growable array type, in Rust Vec<T> and in Odin dynamic[T]
In Rust we only pay extra to store the allocator metadata if there is allocator metadata. If we're content using the general purpose allocator Global with global metadata it's free. In Odin we always need to pay to store an "allocator proc" function pointer and a (possibly unused) metadata pointer, so that's 2x machine word worse for Odin. We pay this even if we never allocate, and indeed even if we can't allocate, whereas of course in Rust that's free.
But then when we do need to allocate, in Rust that's a static dispatch unless our allocator specifically wants dynamic dispatch. So again for Global it's just a static dispatch. But in Odin it's always a slower dynamic dispatch.
The upside? To the extent there is one it's probably the New Jersey style, ie it was simpler to implement Odin. Which is a benefit to Ginger Bill, but not to you as an Odin programmer.
Rust's Allocator APIs have been around since 2016, less than a year after 1.0, they're just now stabilized.
The stdlib types that allocate have been generic (with default generic parameters) over the underlying Allocator used for a very long time. All widely-used data structure libraries where it's feasible/reasonable to use other allocators also are already generic over the Allocator.
It won't "take time to retrofit" because the language and ecosystem have already been prepared for it for 10 years. For context, async syntax was introduced a few years after that and if you just started learning the language today you wouldn't notice it wasn't there in 1.0.
Seems like I've misread the recent Rust version announcement:
> This release brings about the stabilization of the core parts of allocator_api, enabling library authors to define custom allocators that can be used together with certain standard library types - currently Vec and Box. These are enabled through the newly-stabilized Allocator trait.
That made it sound like it was rather limited currently.
I asked an LLM (because it's outside my expertise) and it said:
> It is fair to say that the heaviest architectural lifting is already done: adding a defaulted type parameter (A: Allocator = Global) preserves backward compatibility for existing code, and the core data structures already store and thread the allocator internally. However, plumbing that generic parameter through String, HashMap's std wrapper, and hundreds of standard library methods still represents substantial retrofitting work rather than a completed state.
Looks like The patch notes lead me a bit astray. But it's also not at an 'already all there' state either.
Although it's true that all of Rust's standard library collections are generic over Allocator, Rust 1.100 only stabilizes useful allocator APIs for Box and Vec
If you want to use your MyCustomAllocator with a HashMap or a BinaryHeap or you've lost your mind and want a LinkedList none of that was stabilized yet.
Personally I am slightly interested in HashMap, VecDeque and String but I think Vec is much more heavily used than any of those and will illustrate just how thin the pickings are in this space for real.
In critical low-level code, any kind of overloading - operator or otherwise - is something you'd rather not have.
By expressivity I don't mean what characters you use in the conrecte syntax, but the number of lines needed to express an algorithm (up to a small constant). C doesn't have C++'s expressivity because it lacks templates, and while you can sort of express that kind of thing with macros, it's not all the way there. Zig, however, gets you all the way there (you might need to write matrix_plus instead of +, but the expressivity is the same).
BS. Using operator overloading with SIMD makes it easier to translate algorithms into code. I wrote my own U8x32 wrappers and a + b is way better than a.add(b).
I'm not disputing that some people like user-controlled operator overloading (Zig does have + etc. for vectors), but clearly others prefer no user-controlled overloading of any kind (don't know what the LLMs prefer, if they have any preferences at all). In the safety-critical space where I once worked, where people could get hurt or die if there's a stack overflow, the more common preference was against (user-controlled) overloads or any implicitness. That's why we need options.
C++ destroyed the good will here by simultaneously insisting that you're only supposed to use this feature to give types their natural operators (makes sense) and then inventing implausible new "natural" operators so that Bjarne can show off his new toy with the I/O streams.
I actually don't like Rust's choice to implement + for Strings but most people seem to accept that, however C++ goes ludicrously far beyond "I don't like it but some people do" and straight off the cliff into "Nobody thinks that's OK" leaving you to wonder why there was effort to implement this bad idea. For example C++ can overload the short-circuiting boolean operators... but the overloaded versions don't short circuit. So now you've made things dramatically worse!
Many people are against operator overloading, imo my only issue with C++ operator overloading is that you're allowed to overload the the assignment operator. From what I can see this is the root of many issues
I’ve also experienced some unnecessary shenanigans related to blurring the distinction between primitive operations, and stuff that happens dynamically at run time such as method calls.
For a somewhat different example of the same general idea, look at how Java handles primitives and their boxed counterparts. = on two ints always compares their value. = on two Integers compares their pointers. But it uses flyweight instances for values up to 256, so in Java, whether two Integers with the same numeric value are equal according to the = operator depends on what that value is.
Then throw autoboxing and generics into the mix, and it suddenly becomes very easy to do refactorings that work fine in the unit tests (which tend to use small numbers) but fails in confusing ways in production.
This isn’t an example of operator overloading, of course. But I hope it more simply illustrates why, at least in certain problem domains, there’s value in being able to visually tell when surprising runtime shenanigans are and are not possible. And also to make it impossible to cross that boundary without touching all code that falls within the blast radius of that change.
Is there a way to support varying the meaning of "+" etc without introducing a lot of complexity? I suppose if Zig had nested functions, and the ability to define a nested function called "+", then you could have name shadowing instead of "overloading".
OCaml basically gives you operator overloading this way without actually doing overloading. I'm wondering why we can't have these OCaml-isms in other languages - without needing to use OCaml - which sadly comes with other commitments.
All these new languages are unreadable relative to C. If, like me, you like explicit stuff like for (size_t i = 0; i <= n; i++) total += a[i]; all these new languages won't do it.
Funnily, Zig does allow doing it explicitly like that, just need use while rather for:
var i: usize = 0;
while (i <= n) : (i += 1) {
total += a[i];
}
That said, perhaps Odin, for which simplicity and readability is guiding principle in its design, is more up your alley. In Odin, can be read:
for i := 0; i <= n; i += 1 {
total += a[i]
}
Exactly same style can be used in Go & V (which is/was? quite infamous here in HN). But in all languages, for this example at least, a foreach-like iteration, i.e.
I don't like this explicitness, because very often it's a lie. In the times of PDP-11 that would compile into a literal counter and literal ADD, DEC, and BNE instructions. Now it will most often compile into some vectorized semi-unrolled form with seven-letter instructions like VPHADDD which add many integers at once, etc.
If I want specific instructions to be run in a specific order, like touching hardware registers, I reach for intrinsics or assembly. If I want an array summed, I prefer to express the intent: total = the_array.sum(); I leave the implementation to the compiler / stdlib.
I honestly don't think total = a.sum() is any less explicit than for (size_t i = 0; i <= n; i++) total += a[i]. If anything, a.sum() is more explicit as at least it expresses the programmer's intention explicitly. The loop is one more layer between the reader and the writer's intention, without faithfully reflecting what happens on the bare metal. So it's the worst of both worlds.
God I loved playing around with D back in the day when I was in high school. It seemed like a no-brainer that it was going to be huge. It's such a shame it never took off. Incredible language.
It's so damn nice, its a shame it's never taken off the way it could. I think there's still time to make it attractive to companies, but with people's increasing use of AI I don't think anyone even cares what language their model writes in, look at the insane uptick in Rust programming.
I try D almost every year, but I've never really gotten a working IDE environment with it. I'm not sure if it's issues with being on Windows or just the LSP not being that great. I really wish D was adopted as widely as Go.
It’s pretty obvious why they didn’t, they’re GC by default. That’s already enough to put people off, even if it’s “not a problem” in reality, people will just avoid it for systems programming if they hear that, they will assume all the good features need the GC
Nim is stack allocated by default, and you annotate types explicitly with `ref` to use the GC. Even then, the compiler elides reference counting using borrow checks, you can choose different GC options, and the GC isn't stop the world like most. You don't need to touch reference counting if you don't want to, and Nim fits embedded dev pretty well for that reason.
I like both of those, I like the idea of Zig, it does seem to be striving to be pure in some areas where C is "king" that other languages gloss over, but Zig is always changing, and its got some elitist people in its ranks that left me not wanting to bother with it. Imagine a programming language maintained and used by geohot clones.
I've been toying with the idea of learning zig and weirdly enough I was looking at this guys book. I don't have any experience with low level languages. The closest I've gotten is Go.
I was debating between C and Zig, but I still don't really know what I want to choose. Feels like C is probably the right choice given the languages age and it could technically be more useful (not that I'd be looking for a job doing C work).
C is still supreme for working on the majority for a Linux/BSD system stack. It also imparts skill that transcend to most other spaces where you may eventually need an FFI to interact with a library anyway or understand how it actually runs on a system.
It depends for what you want it. If you're going to do an API, well, Golang is more practical here in my opinion. For TUI apps, I would dare to say Rust was more pleasant than Golang. My experience with Zig was limited, I would not say is not a good fit.
I think you should know C if you want to do low level stuff. It will help understand why zig (or rust) is built the way it is and what is useful about it, too.
I would say from a pedagogical perspective, learning C is hugely useful for understanding how programming works at a much lower level.
This has the benefit of you learning about manual memory management and how to deal with all the issues that come with such a rules free language like C that will basically let you do anything. This understanding means lot of the constraints and features that safer languages like Zig and Rust add start to make a lot more sense. For example, going from Python to Rust, the borrow checker will probably seem rather arbitrary, but going from C to Rust, you will find the borrow checker is just automating a lot of checks you were already doing yourself.
It's okay for me to declare a macro with `__attribute__((warn_unused_result))` for GCC and Clang, and `[[nodiscard]]` when the C code is later compiled with C++ (or nothing).
Rust will take over all these niche languages anyway. Yes, Rust is a viable replacement for C and C++ in the long term. Zig is just trying to be a middle ground, which we've established doesn't work. Zig is particularly irrelevant compared to Rust in the LLM era.
AI slop. Every article on this site has "The honest tradeoff" as a header. Looks like it was generated to promote a book which is no doubt also just slop.
> Every article on this site has "The honest tradeoff" as a header.
This is a hallucination. I'm sorry. This is what we call these kind of statements these days. Unless you can prove your humanhood, that's how it is now.
I wrote recently about a weird issue I encountered with Zig, though to be fair it was related to using Zig to compile C code, humorously in a Rust project (which touches on all of the languages mentioned in this post).
While I did just dismiss it as an edge issue, it turns out that this same rather clumsy sort of mistake has happened repeatedly in Zig's pure rewrite of MUSL. And despite being an absolutely catastrophically bad problem, by my judgment, the problem is still in main at this moment, across a wide number of musl function rewrites.
Which gives me pause about the whole zig project. If you're rewriting a critical library, making such simple but consequential mistakes shouldn't happen, and if they do happen it should be a red alert type situation.
When working with Zig, you are a beta tester. The development process is unfortunately not great, so bugs slip in fairly easily. You need to be ready for it and test your application properly. Large Zig projects don't target the latest version. That said, if you can deal with that, you get a great language that is fun to use and very practical. I think it's worth it, even if it can be frustrating at times.
Zig can be a hobby/learning project (in which case these elementary mistakes, and the lax development process that allowed them, are totally acceptable), or it can be a project that targets critical applications (like TigerBeetle, which handles REAL MONEY), but it can't be both.
Personally I think it's not that catastrophic for zig < 1.0. Just a performance regression, and one can always link to other libc implementations, or is their own implementation enforced now?
If errors like this are acceptable because of the version number of an otherwise decade old project, then that's ample evidence of why Zig is not ready and should not be used for any software that needs surety.
You know that GNU libc to this day essentially calls strlen() on every call to sscanf(), which is pretty much the same issue? And that this behaviour, present in many other implementations of standard C library (including Microsoft's), was responsible for e.g. slow JSON parsing in some very popular online games [0]? Does that mean glibc is still not production-ready?
sscanf is quadratic in the worst case, and has been flagged along with the bulk of the rest of the libc string support as being hazardous or potentially harmful. Its continued existence is for compatibility, but new projects should not use it.
First of all, sscanf doesn't need to be quadratic in the worst case; it absolutely can work without finding the NUL-terminator first. Second, I'd like to see that list of "flagged along with the bulk of the rest of the libc string support as being hazardous or potentially harmful" functions. Does this list include strchr()? Because if yes, your original argument fails to stand.
A version number starting with 0 is not just evidence, but a direct statement by the maintainer that software is unstable/pre-release no matter how long it has been in development.
It isn't my thing though. It's quite literally, by choice, their thing. They are the ones who decided that rewriting musl in zig was something they need to do and was such a priority. Only it's a serious regression.
But I mean, the six people in a trenchcoat comment is basically precisely what should give people pause. In my project I simply eliminated zig from the stack entirely because even if I worked around the problems, what comes next? What's the next showstopper that gets waved away as oh well six people in a trenchcoat...
Some people (not its creators) talk about Zig as if it's anywhere near production ready. It's fine to be excited about it (there's a lot to be excited about), but we gotta be realistic.
This submission is advocating Zig over C for system programming. I pointed out a pretty significant reason why maybe you should pause that decision for a bit, and the response have been much more significant reasons why you should be wary.
Now it's "FUD", also it's a beta and you're a beta tester, and it's 0.x, and look they disclaim that you shouldn't trust it, and barely anyone works on it, and... By the advocates for the language, it should be noted.
This is a bizarre thread. These very comments are precisely why it should give pause (which isn't remotely "FUD", and what an utterly asinine claim). Zig looks neat and it might get there, but this absurd apologism and goalpost moving is not convincing to adopt the language.
Again, in a submission advocating using Zig for system programming.
llm_nerd: "I found a bug. This gives me pause about the whole zig project"
I don't see how the latter is a fair reaction given the former. Especially since they only recently started working on the libc port (although it had been talked about for much longer)
They have been cooking for over ten years. It is more than fair to say at this point that they are not capable of shipping a stable, reliable, fast and efficient compiler. It is fundamentally flawed and I could easily write thousands of words explaining why from the architectural design to the development process.
Zig hides behind its pre 1.0 status to justify poor decision making while simultaneously actively trying to attract attention through marketing. That attention translates to resources and influence that are undeserved on the technical merits.
The dudes name is "llm_nerd" its not worth engaging lol. These are the same people that saw the bun rewrite in rust break tls and fail open and shrugged.
C is far behind the sate-of-art in C and Go families. True, but not bad.
It is what it is, but can work on what it already has, while keeping its tradition.
Small changes, add common extensions, conservative additions, improve coherence and correctness of the standard while keeping easy to implement and small.
No crazy things.
Keep in mind that it is standardized and targets things others do not target.
I don’t get it, C arrays can be used as slices, no?? Given a pointer and a length, which is how C does any array, works just as well (I mean if you consider arrays without a length a fine thing) as a slice!
"C is not going away and I still read a lot of it, because the UNIX source code and most system libraries are written in C and their man pages document them with C prototypes. But for new code I choose Zig."
Has this aged well though?
The TechBros pour in tons of money into "agentic programming" or "agentic workflows" right now. You can see that all over the place; OmarchySlop and other AI-controlled projects, where Claude is now the main designer. So in such a scenario, where does Zig really fit in? The "but for new code I choose Zig" is not a really compelling argument IMO. It reminds me of the issues that D has ("we are the better C++!!!" or even Java too). It's interesting that so many languages try to replace C. All failed, IMO. And C is still a pretty bad language. For instance, I think almost everyone would concede that crates in Rust are useful or good; C++ kind of submitted to this and has their own variant now as far as I know. C does not - or perhaps it has but almost nobody would know it.
I kind of want a language that really sits like in both camps: compiled, and scripting use. But each language seems to just fail hard in one area. Including syntax-wise. Complexity-wise too; I find Rust annoyingly complicated. Go is in many ways simpler than Rust, but it sucks in other ways. Of course there are always trade-offs either way, but I kind of want a language that can offer insane speed while having a great syntax (crystal also fails here; it is fast but the syntax is pretty bad. Why is "alias xyz abc" not working? It's so easy to support. But anyway I could go on for years here.)
My goto language for both fast and still scriptable is Dart. They invested a lot in UX and performance in the last few years and it shows. Run as a script or compile to a small binary that runs pretty fast, similar to Go performance as as as I have bothered to check. Best DX out there even if you don’t care about Flutter.
Their rewrites of all security critical libraries while also refusing to use AI which at the moment is surfacing huge vulnerabilities all over the place.
Zig is a nice language, but those articles always compare new languages against old-style C without use of modern tools. Most of the listed advantages you can have in C as well if you are willing to adopt modern tooling and style: signed overflow checking: sanitizers, errors you can not ignore: [[nodiscard]], slice type which know their size: variably-modified types or span types.
Defer will also appear in C compilers soon, but personally I still think goto-error handling is clearer.
(Rust has an advantage for temporal memory safety, but this does apply to Zig.)
It seems that part of what Zig is trying to do, though, is create a pit of success around these practices. So the response might be something like, “Yes you can do these things in modern C, but doing it consistently requires more discipline than we’d like.”
https://blog.codinghorror.com/falling-into-the-pit-of-succes...
I do not think it needs a lot of discipline. It needs the willingness to do it, but this is also true for adopting a different language.
Yes sir, with the right kind of discipline even assembly starts to look like luxury. With a heart that holds no fear, one can code in straight 0's and 1's.\s
These days I wonder what is harder: 1. problems of collective action or 2. communicating that reducing everything to collective action is the real issue.
If your heart holds no fear then shouldn’t that read 01 or 10?
might as well use rust and fall into a much wider and deeper pit of success
IMO you cannot fairly compare static analyzers/sanitizers with features baked into the language/compiler.
Though I've noticed the same trend when reading a famous Rust primer for C++ developers. Like there are obviously so many good and valid arguments when comparing the two languages that would've gotten the point across. Yet the authors felt the need to push it one step further, where the comparison no longer seems fair. Like comparing Rust with C++17 when C++20 provides the same exact feature that Rust claims to provide.
I'm very certain you can statically analyze zig's IR for safety without language change. And zig makes it easier to get consistency than C because you are steered away from the mishmash of malloc/jemalloc/pick-your-allocator-function. (The analyzer can say "all bets are off if you don't use the allocator interface") same with the new stdlib io interface for async-shaped analysis on aliasing analysis
github.com/ityonemo/clr
Where do you think the static analyzers are baked into for C?
I meant static analyzers like clang-tidy and cppcheck. `-fanalyze` I suppose is a fair comparison, but `fsanitize` is not since it's not even static
Those aren't static analyzers, they're dynamic analyzers; they only catch issues at runtime.
GCC's -fanalyze, and clang's `--analyze` both run static analyzers.
If this is such a good default, why make me type `-fsanatize=address? Why isn't this the default when not doing "release mode"?
The obvious answer is that I don't want to pay for these checks at runtime and can catch most of these bugs at test time. There are definitely use cases where that's acceptable.
It adds nontrivial runtime cost. It’s also not a perfect mitigation (there are plenty of unsound patterns that ASan, UBSan, and TSan won’t catch).
That works as long as you have zero dependencies.
As soon as you have dependencies you are dealing with however the authors decided to build things.
Part of why languages exist is just to get a new ecosystem.
At this point, you could fork C/C++, only remove features, give it a new name, a new package service, and maybe fork the standard library, and you could end up with something pretty competitive against Zig or Rust.
But Zig is expressly not trying to bootstrap a completely new ecosystem. Integrating with - and depending on - existing C and C++ code is a core use case. So much so that they’ve adopted “maintain it with Zig” as a motto that intentionally puns on “rewrite it in Rust”.
I don't see why this requires not having dependencies.
Yes, new languages try to create a new ecosystem. I think this is a bad thing - we already have too many of such - often somewhat isolated - language ecosystems. But I also do not think we need new ecosystems, when we can evolve the existing one. I have seen C evolve quite a bit over time, so the idea that we have to create new ecosystems to solve problems is wrong in my opinion. C evolves very slowly, which means that some things need more time, but I also think that it is generally a very good thing that it evolves slowly.
> I also think that it is generally a very good thing that it evolves slowly
This is why new ecosystems starts. Not everyone is ok with the slow pace
Zig's tooling is very good, so is the cross-compatibility story and tagged unions.
But while you can have some/most of the benefits Zig has from modern C, I think an underrated benefit is the language culture. As a different language it was the opportunity for a reset, e.g. it's idiomatic Zig to pass allocators to functions, meaning every library in Zig will work this way.
You could do that in C, but none or very few libraries do that so whenever you reach out to third-party dependencies (which is not a lot, in C as in Zig) you have to fight against C's traditional patterns.
Because some of those tools exist since 1979, and it keeps being a quixotic battle to make corporations adopt them as standard, until cybersecurity legislation started to take off.
There's a big difference between opt-in and opt-out features.
comptime instead of preprocessor macros is pretty big though…
I mostly agree, but as usual there are the classical catches:
1. yeah you can do that, but we should really do some (most? all?) of the things you mention (as well as other safeguards) mandatory
2. your code might be alright, but what about your dependencies ?
I’ve settled on a dialect of C++ for systems programming after using C for a long time. It’s mostly C-like and I disable exceptions, RTTI, etc. but I do use templates, overloads, methods, lambdas, etc.
I have my own object system with constructor methods that all return a status code so you don’t need exceptions, my own stdlib for containers, buffered I/O, networking, hashing, math, threading, async, etc.
C++ gives you all the features to effectively build your own language, and you can use first class tooling since C++ is widely supported. LLMs also make it much easier to build out and test an entire stdlib on your own.
This is the way, take what suits your use case, using C++ doesn't mean taking the whole feature list.
Agreed. I'd not ditch the STL completely, but only use its freestanding subset. Implementations contain some very fast SIMD-accelerated code[1]. +The STL works without exceptions.
[1]: https://github.com/microsoft/STL/blob/main/stl/src/vector_al...
I've essentially done much the same. One of the more under-rated features of C++ is being able to build a new language environment based on a different set of composable idioms from the low-level primitives on up.
C++ is as much a tool for building nice systems DSLs as it is a systems language. The heavy investment in compile-time expressiveness has really paid off for the language over time.
What are you using for a build system?
Also, Zig is as at least as expressive as C++ (but without the implicitness and general language complexity that makes reviewing some sensitive code, where you need to see exactly what operations are performed, hard), and it supports arenas better than C/C++/Rust.
To anyone interested in arena allocators.
Rust just stabilized its `Allocator` API in version 1.100.0. Definitely worth checking out!
Seems like a long way to walk though. Retrofitting this will take time (if it'll ever complete). Zig made it a core part of the language and library. I'll assume Rust will make it farther than C++ ever made it, but just take a look at how retrofitting custom allocators is going there.
Seems like it will be useful for use-cases like embedded where you're restricted to certain parts of the library anyway. But the battle is kind of already lost for support at the level of Zig or Odin. Even if they retrofitted support for the whole standard library, they'll never be able retrofit all crates or make people even bother adding support for all new crates.
> Seems like it will be useful for use-cases like embedded where you're restricted to certain parts of the library anyway. But the battle is kind of already lost for support at the level of Zig or Odin.
Odin in particular does a simply terrible job here. Take the most common need for allocations, a growable array type, in Rust Vec<T> and in Odin dynamic[T]
In Rust we only pay extra to store the allocator metadata if there is allocator metadata. If we're content using the general purpose allocator Global with global metadata it's free. In Odin we always need to pay to store an "allocator proc" function pointer and a (possibly unused) metadata pointer, so that's 2x machine word worse for Odin. We pay this even if we never allocate, and indeed even if we can't allocate, whereas of course in Rust that's free.
But then when we do need to allocate, in Rust that's a static dispatch unless our allocator specifically wants dynamic dispatch. So again for Global it's just a static dispatch. But in Odin it's always a slower dynamic dispatch.
The upside? To the extent there is one it's probably the New Jersey style, ie it was simpler to implement Odin. Which is a benefit to Ginger Bill, but not to you as an Odin programmer.
Rust's Allocator APIs have been around since 2016, less than a year after 1.0, they're just now stabilized.
The stdlib types that allocate have been generic (with default generic parameters) over the underlying Allocator used for a very long time. All widely-used data structure libraries where it's feasible/reasonable to use other allocators also are already generic over the Allocator.
It won't "take time to retrofit" because the language and ecosystem have already been prepared for it for 10 years. For context, async syntax was introduced a few years after that and if you just started learning the language today you wouldn't notice it wasn't there in 1.0.
Seems like I've misread the recent Rust version announcement:
> This release brings about the stabilization of the core parts of allocator_api, enabling library authors to define custom allocators that can be used together with certain standard library types - currently Vec and Box. These are enabled through the newly-stabilized Allocator trait.
That made it sound like it was rather limited currently.
I asked an LLM (because it's outside my expertise) and it said:
> It is fair to say that the heaviest architectural lifting is already done: adding a defaulted type parameter (A: Allocator = Global) preserves backward compatibility for existing code, and the core data structures already store and thread the allocator internally. However, plumbing that generic parameter through String, HashMap's std wrapper, and hundreds of standard library methods still represents substantial retrofitting work rather than a completed state.
Looks like The patch notes lead me a bit astray. But it's also not at an 'already all there' state either.
Although it's true that all of Rust's standard library collections are generic over Allocator, Rust 1.100 only stabilizes useful allocator APIs for Box and Vec
If you want to use your MyCustomAllocator with a HashMap or a BinaryHeap or you've lost your mind and want a LinkedList none of that was stabilized yet.
Personally I am slightly interested in HashMap, VecDeque and String but I think Vec is much more heavily used than any of those and will illustrate just how thin the pickings are in this space for real.
show me operator overloading in Zig
In critical low-level code, any kind of overloading - operator or otherwise - is something you'd rather not have.
By expressivity I don't mean what characters you use in the conrecte syntax, but the number of lines needed to express an algorithm (up to a small constant). C doesn't have C++'s expressivity because it lacks templates, and while you can sort of express that kind of thing with macros, it's not all the way there. Zig, however, gets you all the way there (you might need to write matrix_plus instead of +, but the expressivity is the same).
BS. Using operator overloading with SIMD makes it easier to translate algorithms into code. I wrote my own U8x32 wrappers and a + b is way better than a.add(b).
I'm not disputing that some people like user-controlled operator overloading (Zig does have + etc. for vectors), but clearly others prefer no user-controlled overloading of any kind (don't know what the LLMs prefer, if they have any preferences at all). In the safety-critical space where I once worked, where people could get hurt or die if there's a stack overflow, the more common preference was against (user-controlled) overloads or any implicitness. That's why we need options.
Fair, but don't conflate custom operators with overloading just because C++ happened to have one very specific way of implementing custom operators.
I'm not talking about custom operators. I'm talking about the ability for programs to overload either operators or other symbols (subroutine names).
The example you choose is rather telling, though. It indicates that the use case you have in mind is not at all the one that Zig is targeting.
A more humble critique would be to just acknowledge that you don’t seem to be a member of Zig’s target audience, and move on.
You can do a+b in Zig with SIMD vectors. No operator overloading required
Sure. But doesn't it still lack autodiff, quaternions, etc?
C++ destroyed the good will here by simultaneously insisting that you're only supposed to use this feature to give types their natural operators (makes sense) and then inventing implausible new "natural" operators so that Bjarne can show off his new toy with the I/O streams.
I actually don't like Rust's choice to implement + for Strings but most people seem to accept that, however C++ goes ludicrously far beyond "I don't like it but some people do" and straight off the cliff into "Nobody thinks that's OK" leaving you to wonder why there was effort to implement this bad idea. For example C++ can overload the short-circuiting boolean operators... but the overloaded versions don't short circuit. So now you've made things dramatically worse!
Many people are against operator overloading, imo my only issue with C++ operator overloading is that you're allowed to overload the the assignment operator. From what I can see this is the root of many issues
I’ve also experienced some unnecessary shenanigans related to blurring the distinction between primitive operations, and stuff that happens dynamically at run time such as method calls.
For a somewhat different example of the same general idea, look at how Java handles primitives and their boxed counterparts. = on two ints always compares their value. = on two Integers compares their pointers. But it uses flyweight instances for values up to 256, so in Java, whether two Integers with the same numeric value are equal according to the = operator depends on what that value is.
Then throw autoboxing and generics into the mix, and it suddenly becomes very easy to do refactorings that work fine in the unit tests (which tend to use small numbers) but fails in confusing ways in production.
This isn’t an example of operator overloading, of course. But I hope it more simply illustrates why, at least in certain problem domains, there’s value in being able to visually tell when surprising runtime shenanigans are and are not possible. And also to make it impossible to cross that boundary without touching all code that falls within the blast radius of that change.
overloading the comma is also a special flavor. everyone should try it once.
Is there a way to support varying the meaning of "+" etc without introducing a lot of complexity? I suppose if Zig had nested functions, and the ability to define a nested function called "+", then you could have name shadowing instead of "overloading".
OCaml basically gives you operator overloading this way without actually doing overloading. I'm wondering why we can't have these OCaml-isms in other languages - without needing to use OCaml - which sadly comes with other commitments.
please no
Yes! And exceptions and inheritance and RTTI too. Bring back all the mistakes! Let's party like it's c++99
All these new languages are unreadable relative to C. If, like me, you like explicit stuff like for (size_t i = 0; i <= n; i++) total += a[i]; all these new languages won't do it.
Bad ergonomics in loops is questionable design, but that loop in particular isn't better than total = reduce(a, +, 0).
Funnily, Zig does allow doing it explicitly like that, just need use while rather for:
That said, perhaps Odin, for which simplicity and readability is guiding principle in its design, is more up your alley. In Odin, can be read: Exactly same style can be used in Go & V (which is/was? quite infamous here in HN). But in all languages, for this example at least, a foreach-like iteration, i.e. would be preferred.I don't like this explicitness, because very often it's a lie. In the times of PDP-11 that would compile into a literal counter and literal ADD, DEC, and BNE instructions. Now it will most often compile into some vectorized semi-unrolled form with seven-letter instructions like VPHADDD which add many integers at once, etc.
If I want specific instructions to be run in a specific order, like touching hardware registers, I reach for intrinsics or assembly. If I want an array summed, I prefer to express the intent: total = the_array.sum(); I leave the implementation to the compiler / stdlib.
I feel like every time someone tries to tell me how good C is they immediately end up reminding me of https://queue.acm.org/doi/10.1145/3212477.3212479
I honestly don't think total = a.sum() is any less explicit than for (size_t i = 0; i <= n; i++) total += a[i]. If anything, a.sum() is more explicit as at least it expresses the programmer's intention explicitly. The loop is one more layer between the reader and the writer's intention, without faithfully reflecting what happens on the bare metal. So it's the worst of both worlds.
D and Nim have had the same features for years btw.
God I loved playing around with D back in the day when I was in high school. It seemed like a no-brainer that it was going to be huge. It's such a shame it never took off. Incredible language.
It's so damn nice, its a shame it's never taken off the way it could. I think there's still time to make it attractive to companies, but with people's increasing use of AI I don't think anyone even cares what language their model writes in, look at the insane uptick in Rust programming.
I try D almost every year, but I've never really gotten a working IDE environment with it. I'm not sure if it's issues with being on Windows or just the LSP not being that great. I really wish D was adopted as widely as Go.
Which IDE are you trying to setup with D? I've not had any issues... You may try DlangIDE possibly?
https://github.com/buggins/dlangide
https://github.com/nim-lang/nimony
Check out Nimony effectively Nim 3.0 from the author. It's a top to bottom rethinking of Nim with all the lessons learned. I'm so hyped about it tbh
https://nim-lang.org/araq/nimony.html
It’s pretty obvious why they didn’t, they’re GC by default. That’s already enough to put people off, even if it’s “not a problem” in reality, people will just avoid it for systems programming if they hear that, they will assume all the good features need the GC
Nim is stack allocated by default, and you annotate types explicitly with `ref` to use the GC. Even then, the compiler elides reference counting using borrow checks, you can choose different GC options, and the GC isn't stop the world like most. You don't need to touch reference counting if you don't want to, and Nim fits embedded dev pretty well for that reason.
I like both of those, I like the idea of Zig, it does seem to be striving to be pure in some areas where C is "king" that other languages gloss over, but Zig is always changing, and its got some elitist people in its ranks that left me not wanting to bother with it. Imagine a programming language maintained and used by geohot clones.
I've been toying with the idea of learning zig and weirdly enough I was looking at this guys book. I don't have any experience with low level languages. The closest I've gotten is Go.
I was debating between C and Zig, but I still don't really know what I want to choose. Feels like C is probably the right choice given the languages age and it could technically be more useful (not that I'd be looking for a job doing C work).
For those that know both, what would you pick?
C is still supreme for working on the majority for a Linux/BSD system stack. It also imparts skill that transcend to most other spaces where you may eventually need an FFI to interact with a library anyway or understand how it actually runs on a system.
It depends for what you want it. If you're going to do an API, well, Golang is more practical here in my opinion. For TUI apps, I would dare to say Rust was more pleasant than Golang. My experience with Zig was limited, I would not say is not a good fit.
I think you should know C if you want to do low level stuff. It will help understand why zig (or rust) is built the way it is and what is useful about it, too.
I would say from a pedagogical perspective, learning C is hugely useful for understanding how programming works at a much lower level.
This has the benefit of you learning about manual memory management and how to deal with all the issues that come with such a rules free language like C that will basically let you do anything. This understanding means lot of the constraints and features that safer languages like Zig and Rust add start to make a lot more sense. For example, going from Python to Rust, the borrow checker will probably seem rather arbitrary, but going from C to Rust, you will find the borrow checker is just automating a lot of checks you were already doing yourself.
TLDR: Learn C. I would recommend the 15122 CMU course as a pretty good foundation. You can find it online here https://www.cs.cmu.edu/~15122/schedule.shtml .
> Errors you cannot ignore
It's okay for me to declare a macro with `__attribute__((warn_unused_result))` for GCC and Clang, and `[[nodiscard]]` when the C code is later compiled with C++ (or nothing).
[[nodiscard]] is in C23.
Rust will take over all these niche languages anyway. Yes, Rust is a viable replacement for C and C++ in the long term. Zig is just trying to be a middle ground, which we've established doesn't work. Zig is particularly irrelevant compared to Rust in the LLM era.
Is Zig stable yet? No.
I like Zig but having to fix a bunch of version issues each time I touch an old zig project of mine really put a damper on my enthusiasm.
I like the idea of Zig too - "better C" sounds just up my valley - but the actual implementation leaves a lot to be desired.
to save your enthusiasm, tell your clanker to fix version issues instead
Heretic! Exile the clanker enthusiast for heresey!
AI slop. Every article on this site has "The honest tradeoff" as a header. Looks like it was generated to promote a book which is no doubt also just slop.
Pangram says 99% AI.
https://www.pangram.com/history/5f3d749c-713c-447a-82cc-d3a6...
I don't know if this Zig over C article is generated, but the Zig over Rust one [0] is painfully AI-written.
I'm not ideologically against AI, but I'd certainly not buy a book from someone who needs to use AI to write an article like that...
[0]: https://www.mtsoukalos.eu/2026/04/why-zig-over-rust-for-syst...
How can you tell? I just scanned it but it doesn't feel like being generated.
Doesn't read AI generated at all.
The prosody and the headline style is a dead giveaway but compared to other AI generated text this article is far more tolerable.
Meanwhile no one's pointing out any technical errors.
> Every article on this site has "The honest tradeoff" as a header.
This is a hallucination. I'm sorry. This is what we call these kind of statements these days. Unless you can prove your humanhood, that's how it is now.
I wrote recently about a weird issue I encountered with Zig, though to be fair it was related to using Zig to compile C code, humorously in a Rust project (which touches on all of the languages mentioned in this post).
https://dennisforbes.ca/blog/microblog/2026/10/the_accidenta...
While I did just dismiss it as an edge issue, it turns out that this same rather clumsy sort of mistake has happened repeatedly in Zig's pure rewrite of MUSL. And despite being an absolutely catastrophically bad problem, by my judgment, the problem is still in main at this moment, across a wide number of musl function rewrites.
Which gives me pause about the whole zig project. If you're rewriting a critical library, making such simple but consequential mistakes shouldn't happen, and if they do happen it should be a red alert type situation.
When working with Zig, you are a beta tester. The development process is unfortunately not great, so bugs slip in fairly easily. You need to be ready for it and test your application properly. Large Zig projects don't target the latest version. That said, if you can deal with that, you get a great language that is fun to use and very practical. I think it's worth it, even if it can be frustrating at times.
I agree Zig can be fun but for real projects people should stick with Rust or even D which is just as fun!
Zig can be a hobby/learning project (in which case these elementary mistakes, and the lax development process that allowed them, are totally acceptable), or it can be a project that targets critical applications (like TigerBeetle, which handles REAL MONEY), but it can't be both.
Why not adding the link to the issue in the blog post?
https://codeberg.org/ziglang/zig/issues/37402
Looks like it is being worked on.
Personally I think it's not that catastrophic for zig < 1.0. Just a performance regression, and one can always link to other libc implementations, or is their own implementation enforced now?
If errors like this are acceptable because of the version number of an otherwise decade old project, then that's ample evidence of why Zig is not ready and should not be used for any software that needs surety.
You know that GNU libc to this day essentially calls strlen() on every call to sscanf(), which is pretty much the same issue? And that this behaviour, present in many other implementations of standard C library (including Microsoft's), was responsible for e.g. slow JSON parsing in some very popular online games [0]? Does that mean glibc is still not production-ready?
[0] https://news.ycombinator.com/item?id=26296339
sscanf is quadratic in the worst case, and has been flagged along with the bulk of the rest of the libc string support as being hazardous or potentially harmful. Its continued existence is for compatibility, but new projects should not use it.
First of all, sscanf doesn't need to be quadratic in the worst case; it absolutely can work without finding the NUL-terminator first. Second, I'd like to see that list of "flagged along with the bulk of the rest of the libc string support as being hazardous or potentially harmful" functions. Does this list include strchr()? Because if yes, your original argument fails to stand.
https://blog.jetbrains.com/blog/2026/06/05/why-zig-isn-t-1-0...
I did not say it's acceptable, just not catastrophic, especially since workarounds exist.
I've used zig in similar capacity in production in the past and personally I would take the slowdown just to avoid dealing with MSVC :)
When you're setting expectations about a project, it's never a bad idea to listen to the creators themselves.
https://ziglang.org/download/0.17.0/release-notes.html
> Zig has known bugs, miscompilations, and regressions.
> Even with Zig 0.17.x, working on a non-trivial project using Zig may require participating in the development process.
> When Zig reaches 1.0.0, Tier 1 support will gain a bug policy as an additional requirement.
A version number starting with 0 is not just evidence, but a direct statement by the maintainer that software is unstable/pre-release no matter how long it has been in development.
strchr has no time complexity guarantees, by the way. This behavior is strictly correct.
The best kind of correct!
> Which gives me pause about the whole zig project
The whole zig project is like six people in a trenchcoat. You can't expect them to prioritize your one thing after 3 weeks
Did you make a PR?
I'm not the poster, but I found the related issue:
https://codeberg.org/ziglang/zig/issues/37402
It isn't my thing though. It's quite literally, by choice, their thing. They are the ones who decided that rewriting musl in zig was something they need to do and was such a priority. Only it's a serious regression.
But I mean, the six people in a trenchcoat comment is basically precisely what should give people pause. In my project I simply eliminated zig from the stack entirely because even if I worked around the problems, what comes next? What's the next showstopper that gets waved away as oh well six people in a trenchcoat...
Eh, whatever.
It's a 0.x project, let them cook. Building things takes time.
https://ziglang.org/documentation/master/std/#std.mem.findSe... didn't exist then, now it does.
And if i may quote from https://ziglang.org/download/0.17.0/release-notes.html:
> Zig has known bugs, miscompilations, and regressions.
> Even with Zig 0.17.x, working on a non-trivial project using Zig may require participating in the development process.
> When Zig reaches 1.0.0, Tier 1 support will gain a bug policy as an additional requirement.
If you're not willing to participate in the development process, then you should wait or use something else.
Isn't that exactly what they're saying?
Some people (not its creators) talk about Zig as if it's anywhere near production ready. It's fine to be excited about it (there's a lot to be excited about), but we gotta be realistic.
No, the framing "give pause" is the sort of framing that people create FUD create
This submission is advocating Zig over C for system programming. I pointed out a pretty significant reason why maybe you should pause that decision for a bit, and the response have been much more significant reasons why you should be wary.
Now it's "FUD", also it's a beta and you're a beta tester, and it's 0.x, and look they disclaim that you shouldn't trust it, and barely anyone works on it, and... By the advocates for the language, it should be noted.
This is a bizarre thread. These very comments are precisely why it should give pause (which isn't remotely "FUD", and what an utterly asinine claim). Zig looks neat and it might get there, but this absurd apologism and goalpost moving is not convincing to adopt the language.
Again, in a submission advocating using Zig for system programming.
Zig team: "Zig is 0.x and has known bugs"
llm_nerd: "I found a bug. This gives me pause about the whole zig project"
I don't see how the latter is a fair reaction given the former. Especially since they only recently started working on the libc port (although it had been talked about for much longer)
They have been cooking for over ten years. It is more than fair to say at this point that they are not capable of shipping a stable, reliable, fast and efficient compiler. It is fundamentally flawed and I could easily write thousands of words explaining why from the architectural design to the development process.
Zig hides behind its pre 1.0 status to justify poor decision making while simultaneously actively trying to attract attention through marketing. That attention translates to resources and influence that are undeserved on the technical merits.
> I could easily write thousands of words explaining why from the architectural design to the development process.
Alright then go for it, hotshot. We're waiting.
It has been 0.x for 10y+. At this point won't be off if considered 0ver (https://0ver.org) and permanently beta.
The dudes name is "llm_nerd" its not worth engaging lol. These are the same people that saw the bun rewrite in rust break tls and fail open and shrugged.
If their resources are so limited, why have they decided that rewriting libc from scratch is a top priority?
Andrew Kelley had a comment re this bug over at lobste.rs:
> I made the judgement call to let this slide in the review. I'm still not sure whether it was the right or wrong call.
https://lobste.rs/s/lh1j0e/shlemiel_painter_is_back_how_zig_...
C is far behind the sate-of-art in C and Go families. True, but not bad.
It is what it is, but can work on what it already has, while keeping its tradition. Small changes, add common extensions, conservative additions, improve coherence and correctness of the standard while keeping easy to implement and small.
No crazy things.
Keep in mind that it is standardized and targets things others do not target.
Yeah
The lack of a native slice is the most glaring C fault I think exists
(And of course the committees won't implement it, and of course C is going to remain broken)
You could help? https://www.open-std.org/jtc1/sc22/wg14/www/contributing.htm...
But I am also quite happy about how far one already gets with variably modified types and a custom span type: https://godbolt.org/z/rPen5Movd
I don’t get it, C arrays can be used as slices, no?? Given a pointer and a length, which is how C does any array, works just as well (I mean if you consider arrays without a length a fine thing) as a slice!
"C is not going away and I still read a lot of it, because the UNIX source code and most system libraries are written in C and their man pages document them with C prototypes. But for new code I choose Zig."
Has this aged well though?
The TechBros pour in tons of money into "agentic programming" or "agentic workflows" right now. You can see that all over the place; OmarchySlop and other AI-controlled projects, where Claude is now the main designer. So in such a scenario, where does Zig really fit in? The "but for new code I choose Zig" is not a really compelling argument IMO. It reminds me of the issues that D has ("we are the better C++!!!" or even Java too). It's interesting that so many languages try to replace C. All failed, IMO. And C is still a pretty bad language. For instance, I think almost everyone would concede that crates in Rust are useful or good; C++ kind of submitted to this and has their own variant now as far as I know. C does not - or perhaps it has but almost nobody would know it.
I kind of want a language that really sits like in both camps: compiled, and scripting use. But each language seems to just fail hard in one area. Including syntax-wise. Complexity-wise too; I find Rust annoyingly complicated. Go is in many ways simpler than Rust, but it sucks in other ways. Of course there are always trade-offs either way, but I kind of want a language that can offer insane speed while having a great syntax (crystal also fails here; it is fast but the syntax is pretty bad. Why is "alias xyz abc" not working? It's so easy to support. But anyway I could go on for years here.)
My goto language for both fast and still scriptable is Dart. They invested a lot in UX and performance in the last few years and it shows. Run as a script or compile to a small binary that runs pretty fast, similar to Go performance as as as I have bothered to check. Best DX out there even if you don’t care about Flutter.
Check out Mojo, a compiled language by the creator of LLVM with Python syntax.
Zig will never be successful.
Why so?
Their rewrites of all security critical libraries while also refusing to use AI which at the moment is surfacing huge vulnerabilities all over the place.
And of course it needs to get to stable first