Re: Unfortunate interaction of glibc-2.44 and gnulib regarding posix_spawn_file_actions_addchdir => infinite recursion
Sam James <[email protected]> Sun, 26 Jul 2026 14:53:34 +0100
| Newsgroups | gmane.comp.lib.gnulib.bugs,gmane.comp.lib.glibc.alpha |
|---|---|
| Organization | Gentoo |
| Message-ID | <[email protected]> |
Collin Funk <[email protected]> writes: > Xi Ruoyao <[email protected]> writes: > >> It can be worked around with >> ac_cv_func_posix_spawn_file_actions_addchdir=yes (which basically tells >> the configure system we have a posix_spawn_file_actions_addchdir, some >> kind of a "fact" with glibc 2.44). >> >> Thoughts on fix this? > > Does the attached patch work? In a container, it looks like the symbols > are defined correctly, but I haven't tested it more than that yet: Your patch doesn't look like it marks it as 'posix', so would it show up using the instructions for posix-modules at https://sourceware.org/glibc/wiki/Testing/Gnulib? At the moment, RMs are instructed to use that guide which is why this problem didn't show up. Should we suggest using `all-modules` instead of `posix-modules`? I think so given it'll catch more changes, and any differences should be investigated anyway. sam
signature.asc
(application/pgp-signature, 418 B)
-----BEGIN PGP SIGNATURE----- iQEBBAEWCgCpFiEEJaa7iN2bdkxrVUHCc4QJ9SDfkZAFAmpmEV8bFIAAAAAABAAO bWFudTIsMi41KzEuMTIsMiwyXxSAAAAAAC4AKGlzc3Vlci1mcHJAbm90YXRpb25z Lm9wZW5wZ3AuZmlmdGhob3JzZW1hbi5uZXQyNUE2QkI4OEREOUI3NjRDNkI1NTQx QzI3Mzg0MDlGNTIwREY5MTkwDxxzYW1AZ2VudG9vLm9yZwAKCRBzhAn1IN+RkMJf AP9lpVUrMveCTe3LbNTvppbjHJNV1gCwc2glj07Px9wJZgD/XE3zqXPXzEeHgouQ VaapJIWXPBTENpKa0E1PNPjswQ0= =5b5Q -----END PGP SIGNATURE-----