Since x86 and x64-64 were merged in 2.6.24 (and just called x86), klibc can't figure out what it should be doing when emerged, and so fails to build. This means that fresh installs (such as the one I was attempting to perform on my new laptop) will fail; and/or users cannot use uvesafb. A check through the ebuild code shows that it's looking for an amd64 kernel to return ARCH="x86_64" or similar, but of course that won't happen since it's now just called x86. Current stable and the latest ~arch both fail identically, and the ~arch version tries to pull in an older 2.6.23 kernel too. This is a blocker because it means one can't make uvesafb work, as directed in the instructions on spock's devspace: http://dev.gentoo.org/~spock/projects/uvesafb/ The error at merge time is as follows: * Found sources for kernel version: * 2.6.24-gentoo >>> Unpacking source... >>> Unpacking klibc-1.5.tar.bz2 to /var/tmp/portage/dev-libs/klibc-1.5/work * CTARGET = x86_64-pc-linux-gnu * Your kernel sources are not configured for your chosen arch! * (KERNEL_ARCH="x86", ARCH="x86_64") * * ERROR: dev-libs/klibc-1.5 failed. * Call stack: * ebuild.sh, line 1701: Called dyn_unpack * ebuild.sh, line 817: Called qa_call 'src_unpack' * ebuild.sh, line 44: Called src_unpack * klibc-1.5.ebuild, line 82: Called die * The specific snippet of code: * die "Your kernel sources are not configured for your chosen arch!" * The die message: * Your kernel sources are not configured for your chosen arch!
Forgot to add blocker for the regression tracker.
Oops, didn't test the versions thoroughly enough. This bug should be "please stabilize klibc-1.5.8", as stable 1.5 utterly fails. 1.5.8 does work for kernel 2.6.24; sorry for the noise. Should remain a blocker for 2.6.24 stabilization? Retitled. :)
yes i don't see why this should not be a blocker Robin, any reason why 1.5.8 should not go stable?
+1 from me, let's get it stable.
archs please do your thing
x86 stable.
ppc stable
amd64 stable
