https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108694

--- Comment #10 from Mark Wielaard <mark at gcc dot gnu.org> ---
(In reply to Florian Weimer from comment #9)
> If we have documentation suggesting to use those options in the Fedora
> context or elsewhere, we should fix them. These options just aren't that
> useful, and I don't think it's appropriate to guide developers in this
> direction. Even warning about old-style function definitions by default in
> GCC 15 is a bit dubious.

As far as I can see we have been using -Werror=old-style-definition since 2015.
And it seemed to be recommended in Modern C guides from Gentoo and Fedora since
2023. So it doesn't seem a very recent thing.

> We got lucky and there is no conflict between old-style definitions and
> unnamed function parameters. I don't think old-style function definitions
> are particular bug-prone after requiring type declarations for all
> parameters. So there is no fundamental reason to rewrite existing code to
> avoid them.

This isn't really about rewriting existing code, the code was already rewritten
years ago, the -Werror=old-style-definition was added to make sure no old style
definitions were newly introduced. It is about not accidentially adding a C23
feature. Why not simply add this to -Wold-style-definition if e.g.
-Wc11-c23-compat is also given?

Reply via email to