Matthew Burgess <[EMAIL PROTECTED]> wrote: > Theodoros V. Kalamatianos wrote: >> So glibc-2.5 does not have this issue ? > > I've just run a build against glibc-2.5 (vanilla upstream sources) and > not seen any test failures. This is running under Ubuntu's > linux-2.6.15-27 kernel, in case it's a combination of kernel+libc that > is triggering your failure.
FYI, it was most definitely a glibc problem, fixed sometime before 2.5. Strace showed that the fchownat+AT_SYMLINK_NOFOLLOW call was mistakenly mapped (by glibc) to a chown call, and we all know that the chown syscall follows symlinks. _______________________________________________ Bug-coreutils mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-coreutils
