[01:29] <lamont> stage2/xgcc -Bstage2/ -B/usr/hppa-linux/bin/ -c   -O2 -DIN_GCC   -W -Wall
[01:29] <lamont> +-Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes        -DHAVE_CONFIG_H
[01:29] <lamont> +-I. -Iada -I../../src/gcc -I../../src/gcc/ada -I../../src/gcc/../include
[01:29] <lamont> +-I../../src/gcc/../libcpp/include  ../../src/gcc/ada/decl.c -o ada/decl.o
[01:29] <lamont> make[4] : *** [ada/decl.o]  Segmentation fault
[01:29] <lamont> make[4] : *** Waiting for unfinished jobs....
[01:30] <doko> we don't support hppa ;-?
[01:31] <lamont> feh
[01:31] <doko> lamont: I did hear rumors you would work for a company producing hppa stuff ...
[01:31] <lamont> -7ubuntu7 built just fine though...
[01:32] <doko> maybe fabbione did change the kernel ...
[01:34] <lamont> doko: not on the machine where it's building... same kernel
[01:35] <doko> ohh, looks like hppa is catching up
[01:35] <doko> starting a 4.0 build as well
[01:36] <lamont> yeah - I dragged home a nearly-current source archive friday.. it's been building since saturday
[01:36] <lamont> or friday night..
[01:36] <lamont> Total 870 package(s) in state Installed.
[01:36] <lamont> Total 4295 package(s) in state Needs-Build.
[01:36] <lamont> still have a goodly distance to go
[01:37] <lamont> note also that ~lamont/buildLogs/Lists/breezy.all.hppa isn't really accurate, other than wrt needs-build and installed
[01:37] <lamont> er... needs-build could be any non-installed state, that is.. :-(
[01:38] <lamont> same is true for breezy.all.sparc: non-data-center architectures don't have w-b access, so installed is the only state that w-b accurately reflects
[01:38] <doko> hmm, packages.u.c doesn' list sparc ...
[01:38] <doko> fabbione: ^^^
[01:38] <lamont> does it list hppa?
[01:38] <doko> no
[01:38] <doko> but ia64
[01:38] <lamont> ia64 is probably way wrong
[01:39] <lamont> since it's split off to the same home as hppa/sparc...
[01:39] <lamont> guess it depends on the code running there..
[01:39] <lamont>  ~lamont/buildLogs/Lists/breezy.all.$ARCH is gospel for data-center architectures, and install-state-accurate for non-DC architectures
[01:50] <doko> E: Package gawk has no installation candidate
[01:52] <doko> E: Package libmpfr-dev has no installation candidate
[01:54] <doko> E: Package libart-2.0-dev has no installation candidate
[01:58] <doko> lamont: how do you build that?
[02:04] <doko> anyway, going to bed now
[03:59] <lamont> doko: clearly I have missed uploading some stuff...
[04:00] <lamont> adding people.u.c/~lamont/ubuntu-hppa hoary main to sources.list may help
[04:36] <lamont> doko: if you're still short packages after adding the hoary repository, let me know - I'll see about making them exist
[06:07] <fabbione> morning
[06:49] <infinity> Afternoon.
[07:02] <desrt> word.
[10:23] <fabbione> doko: gcc-3.4 on ppc is SERIOUSLY broken
[10:23] <fabbione> it keeps segfaulting at very high ratio
[11:37] <doko> morning
[11:37] <doko> fabbione: which version did you use before?
[11:38] <fabbione> the one before the ppc64 changes
[11:40] <jbailey> Heya Matthias.
[11:42] <doko> hi, already, or still awake? :)
[11:49] <jbailey> Sort of both.  It's humidity season in Toronto, so I will sleep like crap for the next 2 weeks before getting used to it.
[11:49] <jbailey> I watched the clock crawl from 4am to 5am and decided that I may as well work and nap mid-day
[12:17] <doko> fabbione: do you see the segfaults only when compiling kernels?
[12:23] <fabbione> doko: difficult to say.. i didn't build anything different
[12:24] <doko> and I only did build compilers ;)
[12:25] <fabbione> doko: gcc doesn't take 10 hours to build :)
[03:54] <fabbione> doko: gcc-4 is still FTBFS here with that timeout thingy
[03:54] <fabbione> i will try to build it manually
[04:10] <Mithrandir> I wonder if multiarch would make sense for doing ABI transitions like we're doing for gcc4.
[04:10] <jbailey> Mithrandir: taggart suggested that when I was chatting with him about it before.
[04:11] <Mithrandir> the only issue is we end up with a bloody long search path for ld at some point.
[04:16] <jbailey> fabbione: l?
[04:41] <fabbione> jbailey: i need to go out.. i will be back later
[05:26] <jbailey> doko: There?
[05:27] <doko> yes
[05:27] <jbailey> doko: Are the multiarch patches applied to gcc-4?  I'm pulling it down, but it'll be a couple minutes more
[05:28] <jbailey> ISTR there being stuff about multiarch in rules.defs
[05:30] <doko> multiarch-include.dpatch is included, but not applied
[05:30] <doko> no, no build support yet
[05:33] <Mithrandir> doko: would you care to pull in the changes from tfheen@idi.ntnu.no--2005/gcc-debian--multiarch--3.4 ?
[05:34] <jbailey> Mithrandir: It looks like the patch that's already included gets rid of lib64 support at the same time.
[05:35] <Mithrandir> well, lib64 is just a symlink on amd64, so.
[05:35] <jbailey> Oh, I see.  i386/t-linux64
[05:37] <doko> Mithrandir: ok
[05:37] <Mithrandir> doko: for both ubuntu and Debian or just Ubuntu?
[05:38] <jbailey> Mithrandir: I suspect that it shouldn't remove existing /lib32 /lib64 support, but just add the multilib directories.  
[05:38] <doko> Mithrandir: it's the same archive
[05:38] <Mithrandir> jbailey: possibly, yes.
[05:51] <fabbione> jbailey: re
[05:51] <fabbione> i have only 10 minutes but i would love to spend them here for you :)
[05:54] <doko> fabbione: i386 kernel build did fail ...
[05:54] <fabbione> doko: read #u-k :)
[05:54] <fabbione> you are too slow
[05:55] <doko> so ppc64 does work now?
[05:55] <fabbione> doko: not in 1.2
[05:55] <fabbione> doko: ocalm is FTBFS :)
[05:56] <doko> xorg "reorganization" ...
[05:56] <fabbione> doko: i know :)
[05:56] <fabbione> jbailey: where are you?
[05:57] <fabbione> doko: you will be able to use ppc64 tomorrow
[05:57] <fabbione> doko: or....
[05:57] <jbailey> fabbione: Sorry, was in another window.
[05:57] <fabbione> doko: grab the latest 1.3 image from http://people.ubuntu.com/~fabbione/ 
[05:57] <infinity> fabbione : Were you happy with that buildd result of linux-source on ppc, or did you still want me to build it on my machine?
[05:57] <fabbione> infinity: i didn't check them yet. i will have to
[05:58] <fabbione> infinity: i386 has a really weird FTBFS
[05:58] <fabbione> it wasn't there yesterday
[05:58] <fabbione> but it was building with the old gcc
[05:58] <fabbione> so i will need to investigate that too
[05:58] <fabbione> jbailey: what can i do for you?
[05:59] <jbailey> Mithrandir, doko: Would you consider a hack for now that added the multiarch includes now, but didn't touch /libFOO?
[05:59] <jbailey> fabbione: Erm.  What was it about...  bugger.
[05:59] <Mithrandir> jbailey: yes, sure.
[05:59] <Mithrandir> jbailey: also, we need to fix binutils.
[05:59] <fabbione> jbailey: dunno.. you pinged me before
[06:00] <jbailey> fabbione: Right, and trying to remember what the !@#$ it's for?
[06:00] <jbailey> Mithrandir: Even for just the include hack?
[06:00] <Mithrandir> jbailey: that's a start, yes.
[06:00] <infinity> fabbione : Alright, well I'll build it anyway, just for kicks.
[06:00] <fabbione> infinity: i am happy with the build. i think davis is the box having problems
[06:01] <infinity> Oh.  Then I won't build it. :)
[06:01] <jbailey> Mithrandir: Sorry, EPARSE, I think.  What needs hacking in binutils for adding the include hack?
[06:01] <elmo> fabbione: did you say you wanted me to not NEW ppc kernel?
[06:01] <fabbione> elmo: just wrote it in #u-d :)
[06:01] <jbailey> fabbione: Right!
[06:01] <jbailey> fabbione: I remember.  make-kpkg lovel for initramfs.
[06:01] <Mithrandir> jbailey: no, but you need to fix binutils for looking for the libs in the right place.
[06:01] <fabbione> elmo: i386 is FTBFS so don't wait for it. i will fix it tomorrow
[06:01] <jbailey> fabbione: When can I book time with you on that?  My perl is teh suck.
[06:01] <fabbione> elmo: given that it did build for me
[06:02] <fabbione> jbailey: tomorrow?
[06:02] <fabbione> jbailey: if you wake up early like today we will have plenty of time to work together
[06:02] <jbailey> Mithrandir: Right, except that I don't care about the libs right now.  libs are already sufficiently handled with /lib{32,64,}
[06:02] <fabbione> jbailey: otherwise i can schedule a late evening session with you
[06:02] <jbailey> Mithrandir: But right now, it assumes a common set of headers for those, which is not true in a number of cases.
[06:02] <Mithrandir> jbailey: hm, ok, sure.
[06:03] <Mithrandir> jbailey: ack, go ahead.
[06:03] <jbailey> fabbione: 'kay.  Probably near the end of your day then tomorrow?
[06:03] <fabbione> elmo: btw the source for 1.2 and 1.3 will be the same. only a few config tweaks to make ppc64 booting.
[06:03] <jbailey> fabbione: Your start-of-day is usually my hint that I should be asleep. =)
[06:03] <fabbione> jbailey: well today you woke up early and that was kind of ideal
[06:04] <jbailey> fabbione: I'm hoping not to repeat that.  5am doesn't agree with me.
[06:04] <fabbione> jbailey: ok. let's do this way
[06:04] <fabbione> if i don't see you around 9 UTC (that should be your 5 am)
[06:04] <jbailey> fabbione: My usual show-up time is 11GMT.  Will you still be around?
[06:05] <fabbione> i will shift the working hours for the evening
[06:05] <fabbione> 11UTC is 1pm here so yes
[06:05] <jbailey> Cool.  Wasn't sure, since you appear to start at 4UTC. =)
[06:05] <fabbione> i do start at 4UTC
[06:05] <jbailey> Sick&Wrong(tm)
[06:05] <fabbione> but i don't stop before 14 or so
[06:05] <fabbione> well it's 6am here
[06:05] <fabbione> i need to wake up to kick my wife out of the bed
[06:07] <fabbione> jbailey: so i guess from 1pm it's fine than
[06:07] <fabbione> let's just start on it as soon as you wake up
[06:07] <fabbione> so we don't need to end the day in 36 hours :)
[06:07] <jbailey> doko: Do you keep your source in something I can look at?  From what's in gcc-4/debian/patches/multiarch-include.dpatch, it looks like all we need is the gcc/Makefile.in patch, and the gcc/cppdefault.c patch.
[06:07] <jbailey> fabbione: Cool, thanks.
[06:07] <fabbione> ok cya tomorrow than
[06:07] <fabbione> bye bye guys
[06:07] <jbailey> g'd evening Fabio!
[06:14] <doko> jbailey: svn.debian.org
[06:14] <jbailey> doko: Cool, I think I have access there.  Is there an ubuntu branch?
[06:14] <doko> no
[06:19] <jbailey> doko: Have you imported Tollef's new change?
[06:19] <jbailey> doko: Or should I start from his patch?  I think even editting what's in there now is probably fine.
[06:24] <doko> jbailey: no, not yet. better to start from his patch
[06:24] <jbailey> doko: Cool, thanks.
[07:14] <doko> lamont, lamont_r, infinity: please could you update dep-waits on the buildds from chinstrap:~doko/frozenapps.txt?
[07:17] <\sh> doko: streber ;)
[07:21] <fabbione> elmo: can you please dist-upgrade breezy and breezy-i386 on concordia?
[07:21] <fabbione> elmo: with the old gcc i can't reproduce the i386 FTBFS
[07:22] <elmo> eh, old?
[07:22] <elmo> I dist-upgraded both, this morning?
[07:23] <fabbione> dpkg -p kernel-package
[07:23] <fabbione> Version: 8.135ubuntu2
[07:23] <fabbione> Filename: pool/main/k/kernel-package/kernel-package_8.132ubuntu2_all.deb
[07:23] <fabbione> either the version of the filename are bongs :)
[07:24] <fabbione> hmm gcc -v says it's ok...
[07:24] <fabbione> wtf
[07:24] <fabbione> elmo: nevermind... 
[07:24] <fabbione> they are ok
[08:18] <\sh> guys...fast question fast answer ;)
[08:19] <\sh> a new debian source package for libX with new name libX-123 (not the version) 
[08:19] <\sh> ubuntu package is libX-111 (not the version)
[08:19] <\sh> do we need to put the "c2" in the packagename, or do we only sync with debian?
[08:25] <Kamion> should add c2 since Debian is using the old ABI; otherwise we become binary-incompatible with Debian in a painful way
[08:28] <doko> Kamion: why do you care at this point? this is a temporary thing.
[08:29] <Kamion> binary-incompatibility is bad, temporary or not
[08:29] <Kamion> it's not hard to add c2, and it doesn't cause problems
[08:31] <doko> well, ok, we would need to add a conflict with the old ubuntu c2 package as well, so, yes, one more modifictation
[08:34] <\sh> Kamion: so u want to have the c2 
[08:34] <Kamion> yes
[08:34] <\sh> ok :) small change
[08:35] <\sh> and C/R also to debian package?#
[08:35] <\sh> not only to old ubuntu package, but also to new debian package?
[08:36] <doko> \sh, you did say, there was no old ubuntu package? or old-ubuntu == old-debian?
[08:37] <\sh> ok again :)
[08:37] <\sh> Old ubuntu package: libcommoncpp2-1.0-1.0,13-5ubuntu1
[08:37] <\sh> sorry
[08:38] <\sh> Old ubuntu package: libcommoncpp2-1.0c102-1.0.13-5ubuntu1
[08:38] <\sh> New Debian package: libcommoncpp2-1.3-1.3.10-3
[08:39] <\sh> so C/R adjustments to libcommoncpp2-1.0c102 (= 1.0.13-5ubuntu1) 
[08:39] <Kamion> C/R?
[08:39] <doko> so new ubuntu will be libcommoncpp2-1.3c2 C/R against libcommoncpp2-1.0c102 and libcommoncpp2-1.3
[08:39] <\sh> Conflicts/Replaces
[08:39] <\sh> doko: ok...now we have it :)
[08:40] <\sh> doko: for "libcommoncpp2-1.0-0, libcommoncpp2-1.0-0c102 (=1.3.10-1), libcommoncpp2-1.0-0c102 (= 1.0.13-5ubuntu1), libcommoncpp2-1.3"
[08:40] <\sh> can i say also "libcommoncpp2-1.0-0, libcommoncpp2-1.0-0c102 (>=1.3.10-1), libcommoncpp2-1.3" ?
[08:42] <\sh> forget it
[08:42] <\sh> different version
[08:42] <doko> just drop the version thing ...
[08:42] <\sh> ok...sorted out actually..thx guys :)
[08:50] <\sh> doko: if you have time, please have a look http://bugzilla.ubuntu.com/show_bug.cgi?id=11542 so i can upload this lib
[09:52] <elmo> doko: what's this 32 gcj for?
[09:54] <doko> OOo2 runtime on amd64, we won't need it, when we have multiarch
[09:54] <elmo> hmm, are you splitting all libs out of ia32-libs?
[09:55] <doko> well, probably jbailes is splitting out glibc, not sure, what remains in there ... maybe ncurses?
[09:55] <elmo> X? :P
[09:56] <doko> wasn't that in ia32-libs-openoffice.org, or was this gtk?
[09:56] <jbailey> Should be ncurses and zlib, I think.
[09:56] <doko> no, I don't go any further at the moment
[09:56] <jbailey> Might be one other.  I think they're largely there for LSB requirements.
[09:56] <elmo> well, yeah, okay, I mean, are you planning to split out all of ia32-libs-oo.o and ia32-libs
[09:56] <elmo> I'm just curious as to why we're doing even these?
[09:57] <Mithrandir> there's a bunch of gtk stuff in ia32-libs-gtk
[09:57] <doko> my original goal was to get rid of OOo.amd64 and build it natively on amd64 using gcc -m32
[09:59] <Mithrandir> jbailey: the set of libraries in ia32-libs is based on "somebody wishlisted this", more or less.
[10:00] <jbailey> Right.  I'm less worried about ia32-libs.  amd64-libs is easy enough to get rid of completely.
[10:02] <Mithrandir> ok, good.
[10:03] <Mithrandir> amd64-libs felt like a gross hack to me.
[10:03] <Mithrandir> even grosser than ia32-libs
[10:03] <elmo> Mithrandir: now you know how we feel about multiarch
[10:03] <elmo> \o/
[10:03] <Mithrandir> elmo: "not happy, but less bad than the alternatives" :-)
[10:04] <Mithrandir> elmo: apart from the "ew multiarch", do you have any thoughts on the binutils patch? http://arch.err.no/index.cgi/tfheen@idi.ntnu.no--2005/binutils--multiarch--0--patch-1?cmd=cs_new&file=debian/patches/124_multiarch.dpatch
[10:05] <Mithrandir> (hooaray for long urls)
[10:11] <elmo> what's it do?
[10:11] <elmo> the description of the patch could use some love
[10:11] <Mithrandir> makes ld look in lib/i386-linux in addition to lib
[10:11] <Mithrandir> hm, agreed.
[10:13] <jbailey> Mithrandir: I trust that it does it to whatever GNU triple it was configured with, right?
[10:14] <Mithrandir> jbailey: not triplet, but yes.  Adding i386-linux every time would be quite silly. :-)
[10:28] <doko> Mithrandir: do you want i386-linux, not i486-linux-gnu (as the new dpkg will print it)?
[11:05] <\sh> hmmm
[11:06] <\sh> can someone check the buildd where /home/shermann/pbuilder/result/breezy/libbonobomm1.3-9c2_1.3.8-2.2ubuntu1_i386.deb is?
[11:07] <\sh> without the homedir stuff ;)