Comment by anqurvanillapy | Hacker News Reader

Comment by anqurvanillapy | Hacker News Reader

Parent Full thread

> There are approaches e.g. Zig.

Yes! Zig has done a great job on many C-related stuff, e.g. they've already made it possible to cross-compile C/C++ projects with Zig toolchain years ago. But I'm still quite stupidly obsessed with source-level compatibility with C, don't know if it's good, but things like "Zig uses `0xAA` on debugging undefined memory, not C's traditional `0xCC` byte" make me feel Zig is not "bare-bone" enough to the C world.

> Micron and Oberon+ programming language.

They look absolutely cool to me! The syntax looks inspired from Lua (`end` marker) and OCaml (`of` keyword), CMIIW. The features are pretty nice too. I would look into the design of generic modules and inheritance more, since I'm not sure what a good extendability feature would look like for the C users.

Well BTW, I found there's only one following in your GitHub profile and it's Haoran Xu. Any story in here lol? He's just such a genius making a better LuaJIT, a baseline Python JIT and a better Python interepreter all happen in real life.

> The syntax looks inspired from Lua (`end` marker) and OCaml (`of` keyword), CMIIW

Oberon+ and Micron are mostly derived from Wirth's Oberon and Pascal lineage. Lua inherited many syntax features from Modula-2 (yet another Wirth language), and also OCaml (accidentally?) shares some keywords with Pascal. If you are interested in even more Lua similarities, have a look at https://github.com/rochus-keller/Luon, which I published recently, but which compiles to LuaJIT and thus serves different use-cases than C.

> I would look into the design of generic modules

I found generic modules to be a good compromise with simplicity in mind; here is an article about some of the motivations and findings: https://oberon-lang.github.io/2021/07/17/considering-generic...

> Haoran Xu, making a better LuaJIT

You mean this project: https://github.com/luajit-remake/luajit-remake? This is a very interesting project and as it seems development continues after a break for a year.

> source-level compatibility with C

not sure if this is exactly what you meant, but in Zig you can #include a C header and then "just" invoke the function. no special FFI syntax or typecasting (except rich enums and strings). it can produce compatible ASTs for C and Zig.

Notable approaches to compatibility with C might be: 1) LLVM, like Rust and Zig did (Zig stopped using it in 2023), since LLVM IR is good for being compatible and optimizing. 2) Other backends like libgccjit, I mentioned this because rustc has a libgccjit backend and its author who is a libgccjit maintainer loves the simplicity of it, one could also think of it as a programmable GCC. 3) Code generation directly to C, that's what Koka does.

So I was talking about the direct C codegen approach, but there's still much of some mess like one needs to choose a C standard and knows how to verify the generated code.

Reply on news.ycombinator.com