For as long as I can remember, the multilib builds I've been doing
(x86_64 and ppc64) produce an appalling number of failures in the
32-bit glibc tests, starting with some sort of problem with
de_DE.UTF-8 text in catgets, and then going rapidly downhill.
The 64-bit multilib test results are reasonably clean, as are the
corresponding tests for i686 and ppc.
I was hoping to compare the builds on ppc64 today, but starting
from 32-bit userspace with a 32-bit kernel personality ('linux32') I
can't build the 64-bit glibc, so I'm not going to get anywhere
fast. (Gotta remember to alter the 'boot or chroot' for ppc64,
the 64-bit kernel with 32-bit personality had no problems executing
either version of glibc, it's just the glibc configury that gets
confused and tries to include the powerpc32 directories with -m64).
So, in the meantime has anybody got any ideas for what we might be
failing to do, or doing wrongly, on multilib which causes these
failures ?
The only thing I can say for certain is that repeating configure -
make - check for 32-bit glibc after it has been installed does not
change the results, so it isn't a question of "the second
build/test uses something installed by the first install".
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
_______________________________________________
Clfs-dev mailing list
Clfs-dev@lists.cross-lfs.org
http://ninja.linux-phreak.biz/mailman/listinfo/clfs-dev
