Hacker News: fuhsnn

New comment by fuhsnn in "QBE – Compiler Back End"

fuhsnn — Fri, 08 May 2026 17:48:57 +0000

The page says "aims to provide 70%" though, funny how words spread.

New comment by fuhsnn in "QBE – Compiler Back End"

fuhsnn — Fri, 08 May 2026 16:52:12 +0000

There is a library conversion fork: https://github.com/sgraham/sqbe.

IMO when the intended usage is AOT with an external assembler, which is another subprocess, text-based IO is actually the more natural approach.

New comment by fuhsnn in "Nobody Reviews Compiler Output"

fuhsnn — Thu, 07 May 2026 20:05:42 +0000

Some of us do spend hours on godbolt.org tweaking code like it was game character build.

New comment by fuhsnn in "CJIT: C, Just in Time"

fuhsnn — Wed, 29 Apr 2026 04:12:18 +0000

That read like a supply-chain attack gold mine if you ask me.

New comment by fuhsnn in "CJIT: C, Just in Time"

fuhsnn — Wed, 29 Apr 2026 04:08:28 +0000

rcc[1] is another real-JIT C compiler. antcc[2] and xcc[3] are worth mentioning for being fast-enough to run C like scripting.

[1] https://github.com/dstogov/rcc

[2] https://codeberg.org/lsof/antcc

[3] https://github.com/tyfkda/xcc

New comment by fuhsnn in "Writing a C Compiler, in Zig (2025)"

fuhsnn — Thu, 23 Apr 2026 11:25:53 +0000

Looking at the repo, the author seemed a little fed up [1] with the nature of lower level language and quitted.

[1] https://github.com/asibahi/paella/blob/main/writeup/c19.md#u...

New comment by fuhsnn in "Kefir C17/C23 Compiler"

fuhsnn — Tue, 21 Apr 2026 18:14:02 +0000

Yeah, to get arcane macro tricks right I probably spent more than a full month total fighting with linked lists, never want to touch that part again.

New comment by fuhsnn in "Kefir C17/C23 Compiler"

fuhsnn — Tue, 21 Apr 2026 08:17:18 +0000

Can't speak for kefir; slimcc has been `make unittest`ing Neovim v0.10.4 with no source modification in a Debian VM (so it's pretty portable, thanks!)

On tcc, the most common dealbreaker is no thread local support, having independent assembler/linker is a wonderful feat but not being feature parity with binutils could lead to build failure, contributors generally less willing to support the correct behavior for ABI, C standards and GCC extensions.

I hope these don't sound like a diss, if I had been better at reading its coding style I would definitely try to contribute to mob branch.

New comment by fuhsnn in "Kefir C17/C23 Compiler"

fuhsnn — Tue, 21 Apr 2026 06:03:03 +0000

c-testsuite itself was curated from simple-cc and tinycc's test files, the latter originated from picoc.

New comment by fuhsnn in "Kefir C17/C23 Compiler"

fuhsnn — Tue, 21 Apr 2026 02:48:45 +0000

I have found portability bugs in many projects with slimcc just because it exposed different preprocessor defines, some gate critical __attribute__ behind __GNUC__ check, some have buggy fallback to __builtin functions or VLA that nobody noticed in years, these could have been avoided with just an automatic build job in the CI with kefir or slimcc, (tcc is awesome but less suited for being a drop-in).

It is also important to have more independent implementations of the C standard, not only to sort out dark corners in the specification (current WG14 have been doing great), but to prevent it turning into GCC-Clang power struggle.

New comment by fuhsnn in "Kefir C17/C23 Compiler"

fuhsnn — Tue, 21 Apr 2026 01:47:48 +0000

Congrats for the new optimizing pipeline, and thanks for the acknowledgment! It's nice to have company in the non-__GNUC__ camp.

New comment by fuhsnn in "SDL bans AI-written commits"

fuhsnn — Thu, 16 Apr 2026 11:17:44 +0000

TinyCC's mob branch on repo.or.cz just got trolled with AI commits today. Nowhere is safe it seems.

New comment by fuhsnn in "Retrofitting JIT Compilers into C Interpreters"

fuhsnn — Thu, 16 Apr 2026 00:27:43 +0000

I don't fully grasp it either, the most appropriate analogy I can think of is like how OpenMP turns #pragma annotated loops into multi-threading, this work turns bytecode interpreting loops into JIT VM.

New comment by fuhsnn in "Retrofitting JIT Compilers into C Interpreters"

fuhsnn — Wed, 15 Apr 2026 23:35:59 +0000

Took me a while to figure out whether it's interpreters for C programs or if there's a particular class of interpreters called "C". Turns out it's about interpreters implemented in C that they use modified LLVM to do the retrofitting, but couldn't it be applicable for other languages with LLVM IR, or other switch-in-a-loop patterns in C?

New comment by fuhsnn in "Just Enough Chimera Linux"

fuhsnn — Tue, 14 Apr 2026 03:57:05 +0000

The website you linked says mallocng?

New comment by fuhsnn in "Just Enough Chimera Linux"

fuhsnn — Tue, 14 Apr 2026 00:14:15 +0000

Don't get mimalloc and mallocng mixed up though, completely different animals.

New comment by fuhsnn in "C89cc.sh – standalone C89/ELF64 compiler in pure portable shell"

fuhsnn — Fri, 03 Apr 2026 08:35:20 +0000

Thanks, that actually look like a very solid baseline to start things with. Are you aware of onramp[1]? They use a custom VM to base compiler and shell on, it's extra steps, but could be more flexible long term.

[1] https://github.com/ludocode/onramp

New comment by fuhsnn in "C89cc.sh – standalone C89/ELF64 compiler in pure portable shell"

fuhsnn — Fri, 03 Apr 2026 05:41:01 +0000

Well, I happen to have been recreationally maintaining a hobbyist C compiler for three years, adding tests is part of the fun.

New comment by fuhsnn in "C89cc.sh – standalone C89/ELF64 compiler in pure portable shell"

fuhsnn — Fri, 03 Apr 2026 03:21:35 +0000

> It does make it a little hard to understand

Or much easier to backdoor...

New comment by fuhsnn in "C89cc.sh – standalone C89/ELF64 compiler in pure portable shell"

fuhsnn — Fri, 03 Apr 2026 03:17:25 +0000

Don't understand why you were downvoted. An untested C compiler is simply worthless.