interface99 VS c2rust

There is a project that achieves something like this, and quite nicely I think, worth checking it https://github.com/Hirrolot/interface99

I agree; I would especially not recommend abusing macros throughout an application codebase too much. Conceptually, Metalang99 is more of a (sub)language than a library, which also adds some entry barrier. Ideally, I see the application of Metalang99 being used "behind the scenes", e.g., encapsulated in separate code files/libraries such as Datatype99 and Interface99. This is what I (mostly) do in SmolRTSP.

There are also some differences. Golang, for example, can resolve interface methods at run-time, whereas Interface99 constructs virtual tables statically. Interface99 allows default implementations; Golang doesn't. And, of course, Interface99 mandates placing impl(MyIface, MyType), whereas Golang uses a.k.a. duck typing for interfaces (interface implementations are indistinguishable from ordinary methods). Also, when you would use embedding in Golang, such as this:

I think your github repo link is broken. Here's a working link.

I've done something similar with Datatype99 and Interface99. They are like a macro eDSL that compiles to C. The first one features algebraic data types, the second one features interfaces.

If you need a bit more high-level constructions, you can use the Datatype99 and Interface99 libraries. The former provides polymorphism over data, the latter -- over behaviour (I am the creator of these libraries).

Is it possible that you could benefit from "a better C"? If so, I'd love to hear your take on Crystal, and the recent announce on Interface99.

There is a full example code: https://github.com/Hirrolot/interface99/blob/master/examples/state.c.

A recent practical example of the former: the fish shell re-wrote incrementally from C++ to Rust, and is almost finished https://github.com/fish-shell/fish-shell/discussions/10123

An example of the latter: c2rust, which is a work in progress but is very impressive https://github.com/immunant/c2rust

It currently translates into unsafe Rust, but the strategy is to separate the "compile C to unsafe Rust" steps and the "compile unsafe Rust to safe Rust" steps. As I see it, as it makes the overall task simpler, allows for more user freedom, and makes the latter potentially useful even for non-transpiled code. https://immunant.com/blog/2023/03/lifting/

But not all transpilers are between languages where at least one of them is designed to be transpiled. For example, c2rust can transpile, as the name suggests, C to (ugly, unsafe) Rust. A while ago there was a Java -> C compiler in GCC (GCJ), but it's pretty out of date now.

I do not know about your specific issue but you may be interested by https://github.com/immunant/c2rust

You should also consider using C2Rust (they're even working on C -> safe Rust translation)

> The date at the bottom of the article is 2022-06-13. Has there been further progress?

The article links to their github repo:

https://github.com/immunant/c2rust

There's commits in the last hour, so at least some signal of life.

This is arguably already the state of things.

Rust might get compiled down through MIR, down through LLVM IR, down to assembly or wasm... which then might be JIT or AOT (re)compiled into other bytecodes... which might perhaps be decompiled back up to C... and C might be retranslated back to horrific unsafe-spamming Rust by the likes of https://c2rust.com/. We've come full circle!

The main issue is that retranslating high level languages into other high level languages isn't something that there's actually a lot of demand for, especially commercially, especially given the N x M translation matrix going on. So a lot of the projects "stabilize" (get abandoned). And automatically translating between the idioms of those languages gets even nastier in terms of matrix bloat.

Well, you've got stuff like MSIL and JVM bytecodes which are higher level, and preserve more type information, and can be compiled to / decompiled from while still preserving more structure, but they still form competing incompatible ecosystems.

That's the wrong direction. What's needed are intelligent converters which convert less-strict languages to more-strict ones.

Non-intelligent converters just make a mess. Here's c2rust.[1]

Classic C++ to modern C++, plus a compiler flag to lock out all the old unsafe stuff, would be an achievement.

[1] https://c2rust.com/

Well, technically, it's not hard to build such data structures. If you are willing to liberally use raw pointers, UnsafeCell, MaybeUninit and ManuallyDrop, then you can more-or-less write C-equivalent code in unsafe Rust. (there are even transpilers from C to Rust)

/uj This transpiles from C to unsafe Rust using an existing tool, then strips the unsafe keyword from the generated function signatures