The new 6680_all_glibc-mips-updates.patch in glibc-2.4-patches-1.12.tar.bz2 breaks glibc builds on MIPS. This patch should be applied only to the trunk glibc checkouts after 2006-05-04. See http://sources.redhat.com/ml/glibc-cvs/2006-q2/msg00437.html Please remove this patch from all glibc *release* ebuilds!
Well, there is a reason that none of this stuff is unmasked for mips. We do not support glibc-2.4 on mips. This version requires gcc-4.1.0 to be keyworded and umasked and that the nptl USE flag be unmasked on mips, neither of which has yet happened. Therefore, we really can't support glibc-2.4 bugs on mips at this time. That said, I have been working on making sure all of this works so that when the time comes, the unmasking should go smoothly. I do know that the next glibc-2.4 patchset will work for mips (I just built it in two chroots yesterday). Vapier probably just hasn't pushed out the new patchball yet. If you are really impatient, you could grab the new patch from viewcvs and roll your own patchball.
binutils-2.16.92 x gcc-4.1.0 x glibc-2.4 nptlonly builds fine *without* this 6680_all_glibc-mips-updates patch. the point is not unmasking or keywords or USE flags, the point is i'm pissed off because you try to apply a patch *which is necessary for the trunk* to a release! if you apply this, you break things. is it that hard to understand? you never read libc-ports and glibc-cvs mailing lists?
(In reply to comment #2) > binutils-2.16.92 x gcc-4.1.0 x glibc-2.4 nptlonly builds fine *without* this > 6680_all_glibc-mips-updates patch. Irrelevant. Besides, as I already said, this is fixed in the next patch tarball which will be pushed out *very* soon. > > the point is not unmasking or keywords or USE flags, the point is i'm pissed > off because you try to apply a patch *which is necessary for the trunk* to a > release! Uhh, this is precisely the point. You are using stuff that is *not supported* by the Gentoo/MIPS team. Therefore, any breakage you suffer by trying to rice out your system before we are confident that this stuff is ready for end users is your problem. Or, if you want to me to put it more bluntly, tough shit. > if you apply this, you break things. is it that hard to understand? > you never read libc-ports and glibc-cvs mailing lists? No, it isn't hard to understand. Is it also hard to understand that we have already fixed it, and by reopening this bug and making an ass of yourself, you aren't helping yourself any?
6680_all_glibc-mips-updates.patch in no way whatsoever requires the getaddrinfo.c changes ... if you think that, maybe you're the one who needs to go back and review the libc mailing lists
