[06:16] <fabbione> morning
[06:50] <jbailey> Fabio!
[06:51] <fabbione> hey jbailey 
[06:51] <fabbione> jbailey: apparently the kernel errors on sparc (2.6.12) is something to do with the toolchain
[06:51] <jbailey> Joy
[06:51] <fabbione> jbailey: in debian the same kernel (or almost) builds fine
[06:51] <fabbione> so i am building now in a hoary chroot to see
[06:52] <jbailey> Both using gcc-3.4?
[06:52] <jbailey> Or just using the usual default?
[06:52] <fabbione> both using gcc-3.4
[06:52] <jbailey> Odd
[06:52] <jbailey> It shouldn't be any of the glibc bits that I've done.
[06:52] <fabbione> oh you mean in debian?
[06:52] <fabbione> debian was the default
[06:52] <jbailey> So from 3.3 to 3.4
[06:52] <fabbione> for hoary/breezy i am using gcc-3.4
[06:52] <fabbione> but
[06:52] <fabbione> i did try in breezy with gcc-3.3
[06:53] <fabbione> same results
[06:53] <jbailey> Hmm, I won't have much time to help you this week.
[06:53] <fabbione> don't worry
[06:53] <jbailey> I forgot that gcc summit was this week, so I'll be away a good chunk of the week.
[06:53] <fabbione> i am only trying to figure out what is causing the breakage
[06:53] <jbailey> But I'll have my laptop.
[06:54] <fabbione> once we know what does it, we will have a better idea of what to hunt down
[06:54] <fabbione> right now there are too many vars in place
[07:27] <fabbione> infinity: i think i have almost done with ghc6 :)
[07:32] <fabbione> oh crap
[07:32] <fabbione> battle star concordia is down
[07:46] <fabbione> doko: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=309317
[08:46] <fabbione> jbailey: confirmed.. kernel on sparc is a toolchain problem. i am going to update one bit at a time to see what make it fail
[10:37] <doko> fabbione: ocaml: we have to try it again with 3.4 again, when the X headers are in place
[10:37] <fabbione> ok
[10:38] <fabbione> doko: i was only reporting svenl "let's bitch fabbione for doko's stuff" rant on #d-kernel
[11:01] <fabbione> oh fun
[11:02] <fabbione> jbailey: you are not going to like this... the kernel link failures happen with the new glibc :)
[11:02] <fabbione> jbailey: clean hoary -> kernel ok
[11:02] <fabbione> jbailey: hoary + breezy glibc -> failure
[11:02] <fabbione> jbailey: please fix libc6. kthxbye
[11:02] <fabbione> (oh btw.. it does the same with or without NTPL
[11:02] <fabbione> )
[01:13] <jbailey> fabbione: Joy.  I need more details.  Like why is your kernel linking with glibc? =)
[01:14] <fabbione> jbailey: pleasure :)
[01:14] <Mithrandir> jbailey: I guess makedepend and such is linking with glibc?
[01:14] <fabbione> probably all the tools are....
[01:15] <doko> Mithrandir: could you have a look at updating amd64-libs to be able to build the biarch compiler again?
[01:15] <fabbione> doko: don't even think about uploading gcc-4
[01:15] <doko> fabbione: no, 3.4 first ;)
[01:15] <fabbione> or i am going to send you a dead horse :)
[01:15] <Mithrandir> doko: could you get jbailey to do it?  He's, afaik, working on some stuff for glibc to make it go away.
[01:15] <fabbione> 3.3 and 3.4 are ok
[01:15] <fabbione> 4.0 is insane to build
[01:16] <doko> Mithrandir: promised, I'll pester him ;)
[01:18] <jbailey> doko: I still need your help on that to get the includedir working right.
[01:19] <jbailey> The problem is still that glibc doesn't consider i386 and amd64 biarch to one another, so I need to be able to have different headers looked at for -m32 and -m64
[01:20] <jbailey> I think Tollef's multiarch patch for just the cppFOO.c file
[01:20] <jbailey> (looking up the actual file name)
[01:20] <doko> ahh, ok
[01:21] <jbailey> cppdefault.c
[01:22] <doko> so let's experiment with 3.4 first, and break the kernel builds
[01:22] <Mithrandir> jbailey: actually not, it doesn't look at switches, it just adds to the list of include files, IIRC.
[01:22] <Mithrandir> jbailey: but with a nice set of #ifdef __i386__ and __amd64__ it could work fine.
[01:23] <jbailey> I'm not really interested in screwing with the glibc headers in ways that they won't accept for the 2.3 branch if I can avoid it.
[01:23] <jbailey> Since the goal is to have the multiarch patch in anyway....
[01:23] <jbailey> Basically, I think I just need to move bits/ and the kernel headers into the multiarch include directories, and I should be fine.
[01:30] <karlheg> Right, no need for that.  Was trying to get xchat to open a separate channel for us, but there isn't really any need.
[01:31] <karlheg> I've looked over initramfs-tools and am making some changes.
[01:31] <jbailey> karlheg: #ubuntu-kernel is a better choice. =)
[01:31] <karlheg> I have it (untested) able to support module arguements, so far, and now I'm looking at hook scripts that run when mkinitramfs is run.
[01:32] <karlheg> Ok.
[04:22] <jbailey> lamont: I'll be in a car with Carlos for 7 hours tomorrow.  I imagine hppa hacking will come up at some point.  Anything in particular you want covered? =)
[04:23] <lamont> hrm... lets see.  Fixing the current 4.0 ICE's would be cool.
[04:24] <lamont> working TLS would be cool
[04:49] <lamont> oh, and (of course) getting a gcc-4.0 with the optimizer regression fixed.
[04:49] <jbailey> And of course to finish his NM. =)
[05:20] <elmo> concordia is back, for those who care
[05:34] <jbailey> elmo: Tx.
[10:49] <lamont> gcj: : No such file or directory
[10:49] <lamont> make[5] : *** [gnu/java/security/x509/ext/IssuerAlternativeNames.lo]  Error 1
[10:49] <lamont> doko: is that expected  for -9ubuntu1?
[10:50] <doko> hmm, interesting ...
[10:52] <lamont> that's about 4.5 hours into the build, -9ubuntu2 is 2.5 hours into its build
[10:55] <doko> that won't change
[10:57] <doko> lamont: hppa?
[10:59] <lamont> yes
[10:59] <doko> fix it ;)
[11:00] <lamont> (missing build-dep?)
[11:01] <lamont> build-depends on gcj, but does not declare such a build-dep.
[11:02] <lamont> doko: fix that.  kthxbye
[11:03] <doko> heh, hppa does have java starting with 3.3 :-)
[11:03] <lamont> well, okj
[11:04] <lamont> hrm.. not in ubuntu main, though. :-)
[11:04] <lamont> anyrate, it runs the gcj command, which winds up being 4.0 these days.
[11:12] <doko> ehh, it doesn't run gcj from the build tree?