Hi,

As we discussed a bunch in #gentoo-toolchain over the last few weeks,
GCC 14 is now in ~arch. We've spent the last > 2 years working on this
and fixing lots of packages, but kind of reached the limit of how much a
handful of people can do.

This is kind of the analogue of the Clang post I made a while ago [0].

It makes a lot of dangerous, broken C constructs errors by default. We
have a writeup at https://wiki.gentoo.org/wiki/Modern_C_porting. See
also https://gcc.gnu.org/gcc-14/porting_to.html.

Advice:
* Please ask us in #gentoo-toolchain if you need help understanding how
to fix a bug. You may have users submitting patches, but please ask if
you're unsure if they're correct.
* Please only add casts if you're sure they're right. Otherwise, they're
just suppressing it.
* Report upstream immediately if you see an issue, even if you don't yet
have time to try fix it yourself. This means upstream can work on it in
the background and other distros and folks may end up sharing their
patches.
* Similarly, it's paramount that patches get submitted upstream so they
can be reviewed and other distros can benefit. We have benefitted a lot
from others doing that, and vice-versa.
* The code was broken before (badly) and this just surfaces
that. -Wno-error=* and friends are not an option. The wiki page covers
that but I can explain it more if needed.

The tracker bugs are:
* https://bugs.gentoo.org/870412 (c99-porting)
* https://bugs.gentoo.org/906027 (implicit-in-configure)
* https://bugs.gentoo.org/914580 (gcc-14, general; only use for other issues)

The implicit-in-configure issues are more important than c99-porting because of the risk
of a misconfigured build.

[0] https://public-inbox.gentoo.org/gentoo-dev/87tu4c8z49.fsf@mop.mail-host-address-is-not-set/

thanks,
sam