On Wed, Jan 29, 2025 at 12:20:51AM +0100, Dimitry Andric wrote: > On 28 Jan 2025, at 19:32, Jessica Clarke wrote: > > On 22 Jun 2023, at 22:11, Dimitry Andric wrote: > >> > >> clang: re-downgrade implicit int/function declarations to warning only > >> > >> [...] > >> As noted in the upstream discussion, there are many programs that fail > >> to configure or build correctly, if these warnings are turned into > >> errors by default. > >> > >> Note that most affected programs in ports are relatively old, and are > >> unlikely to be fixed by actually adjusting their declarations, but by > >> compiling with -std=gnu89, which downgrades the errors back to warning > >> again. Lots of tedious work for very little gain. > > > > This remains as a downstream divergence 1.5 years later, but I do not > > think we should be carrying this indefinitely. Software needs to build > > with upstream LLVM, and the fact that many ports do not is now biting > > us for CheriBSD where we use our own external toolchain that does not > > have such a diff. In my opinion this should be reverted for FreeBSD 15 > > and it's long time for ports people to step up or have the broken ports > > removed; we can't keep going on like this with abandonware holding back > > forward progress. > > While I agree, at the least a full exp-run should be done, and even then > you can't be entirely sure if ports that build are built the same as > before, for example due to bad configure tests. So this is potentially a > massive undertaking.
Can we actually do an exp-run to identify affected ports and at least get the idea of the fallout size? I've recently hunted one such subtle bug exposed by GCC 14 so yeah I fear there'd be a lot, but we need to start with something. I presume fixing it won't be particularly hard, just tedious. ./danfe
