On Dec 22, 2024, at 07:03, Jason Bacon <jtocino%gmx.com@localhost> wrote:
Is this patch fixing a critical problem on Darwin-x86? If so, perhaps
it could be limited to that platform by using PATCHFILES, similar to the
aarch64 patches it collides with.
Yes. The package (or any of gcc13/14) will not build without the fix.
I feel we are facing a trade-off between which of the two Darwin platforms is to be supported in the immediate short term of this pkgsrc release.
It seems that the only way to do both is to follow your suggestion of using PATCHFILES and hosting a second patchset. Another, perhaps less long-term hacky, solution would be to include a special patch file in FILESDIR and a special conditional patch command in the Makefile. Those could be removed once the bigger issue of two conflicting patch sets is addressed in a better way, and there would be no need for long-term hosting of a separate patch file as it would all be self-contained within the package.
This is messing with the release process, though, and I am leaving the call to others. Postponing until after the release leaves Darwin x86 broken, but that may be an appropriate cost.
Cheers,
Brook