On 3/9/21 8:14 AM, shwaresyst via Libc-alpha wrote:
The question becomes whether glibc is in violation of POSIX for having made the change,
I don't see how that would be. Apps must define _SC_SIGSTKSZ_SOURCE or _GNU_SOURCE to get the new API, which means the apps do not want strict POSIX conformance anyway.
or whether POSIX needs to be amended to allow SIGSTKSZ to be non-preprocessor-safe and/or non-constant.
That would be a good idea, yes.
I've now seen reports of several projects failing to build when using glibc with this change included.
Yes. I just now checked, and Emacs appears to have this problem so I installed a patch[1] to Emacs, which should fix it. I'm not surprised that other packages have similar issues, and would need similar patches. Gnulib was fixed to avoid this problem in October - before that, Gnulib didn't even conform to POSIX, because it used SIGSTKSZ in #if. The Gnulib patch[2] illustrates other portability messes in this area.
[1] https://git.savannah.gnu.org/cgit/emacs.git/commit/?id=f97e07ea807cc6d38774a3888a15091b20645ac6 [2] https://git.savannah.gnu.org/cgit/gnulib.git/commit/?id=f9e2b20a12a230efa30f1d479563ae07d276a94b
