Cmacro: Lisp Macros for C

[thesephist]

https://github.com/eudoxia0/cmacro

[zetalyrae]

That's me! I was really naive about parsing back then and the "parser" is really not stellar.

[informatimago]

Or elpp: https://github.com/informatimago/elpp/blob/master/elpp.el

[slaymaker1907 → informatimago]

I think this is doing something more sophisticated since it seems to understand C syntax. I don’t think it’s hygienic though which is a shame.

[smegsicle → slaymaker1907]

hygienic macros are a fun intellectual curiosity but using regular macros with gensym as needed is more practical

[shirleyquirk → smegsicle]

Oh man I couldn't disagree more. Having used Nim's hygienic templates, I strongly believe the escape hatch should be the other way round. Hygienic by default, tag individual identifiers as being 'dirty' to inject them as is into the surrounding scope.

Edit: actually, I realize I don't know whether you meant 'macro' in the C sense or the Lisp sense. I was thinking C

[bitwize → slaymaker1907]

When people say "Lisp macros" they usually mean Common Lisp defmacro-flavored ones. Any problems introduced by lack of hygiene can be papered over with separate variable and function namespaces and gensyms.

[slaymaker1907 → bitwize]

Not really because a huge feature of hygiene is accurate symbol provenance for syntax error reporting. syntax-case (syntax-parse for Racket) lets you do anything a Common Lisp like macro system can do but is still hygienic by default.

[marcodiego]

The example using qsort illustrate how good was the lambda proposal for c23 that was rejected.

[FrozenVoid → marcodiego]

Lambdas can still be written using GCC extensions. https://github.com/FrozenVoid/C-headers/blob/main/lambda.h

[beeforpork → FrozenVoid]

Wow, returning a local function from a ({...}) block is sick. I like that.