[05:32] <fabbione> morning
[05:44] <lamont> jbailey: any word on: /usr/include/linux/socket.h:10: error: redefinition of 'struct sockaddr'
[05:50] <jbailey> Arch and/or package, please?
[05:50] <jbailey> My brain's been a bit full lately.
[07:55] <elmo> jbailey/fabbione: the --as-needed stuff is a mess
[07:55] <jbailey> elmo: From the bit that I posted for him?
[07:55] <jbailey> Or do you mean in the archive?
[07:55] <elmo> it has a touch-all testsuite update, which depends on an earlier touch-all testsuite update which isn't easily backportable
[07:56] <jbailey> Right, the testsuite will suck incredibly.
[07:56] <jbailey> I think the best bet for Debian is a CVS update, the best for Ubuntu (hppa, sparc) is to ignore the testsuite and pray.
[07:56] <elmo> well, I'm leaving ubuntu up to you guys
[07:57] <elmo> a CVS update for Debian isn't a horrible idea actually
[07:57] <elmo> let's ask drow
[07:58] <jbailey> elmo: Where are you asking?
[07:59] <elmo> privmsg
[07:59] <elmo> old cabal habits die hard
[08:00] <jbailey> *lol*
[08:05] <fabbione> elmo/jb: i just finished to build binutils...
[08:06] <fabbione> but i am too tired right now to do sensible tests
[08:06] <fabbione> will do it tomorrow morning
[08:11] <jbailey> fabbione: Can you humour me and do hello world at least?
[08:11] <fabbione> jbailey: sure...
[08:14] <fabbione> what was the bug number for the binutils?
[08:14] <elmo> #320697:
[08:14] <fabbione> no .. in ubuntu
[08:14] <jbailey> 12822
[08:15] <fabbione> there.. 
[08:15] <fabbione> thanks
[08:17] <fabbione> jbailey: you know what?
[08:18] <fabbione>  grep PROCEDURE_LINKAGE_TABLE *
[08:18] <fabbione> Binary file libGL.so.1.2 matches
[08:18] <fabbione> sparcbuildd@vultus5:~$ gcc hello.o -Wl,--as-needed /usr/lib/libGL.so.1.2
[08:18] <fabbione> sparcbuildd@vultus5:~$ ./a.out 
[08:18] <fabbione> hello world!
[08:19] <jbailey> Did that used to fail?
[08:19] <fabbione> sparcbuildd@vultus5:~$ gcc hello.o -Wl,--as-needed /usr/lib/libGL.so.1.2
[08:19] <fabbione> sparcbuildd@vultus5:~$ ./a.out 
[08:19] <fabbione> hello world!
[08:19] <fabbione> sparcbuildd@vultus5:~$ ldd a.out 
[08:19] <fabbione>         libc.so.6 => /lib/libc.so.6 (0x70030000)
[08:19] <fabbione>         /lib/ld-linux.so.2 (0x70000000)
[08:19] <fabbione> jbailey: yeps :)
[08:19] <fabbione> it works..
[08:19] <jbailey> And it's Modra that rocks.  Glad we don't have to invalidate the archive while we're at it, though.
[08:19] <fabbione> and hello world comes out :)
[08:19] <jbailey> Have a good sleep, Fabio. =)
[08:19] <fabbione> yeah i am in the need of one
[08:19] <fabbione> i think i am getting a big flu
[08:19] <jbailey> fabbione: I'll confirm what to do with doko and try to get you a binutils update.
[08:19] <jbailey> doko: ^^^
[08:20] <fabbione> jbailey: that would be cool
[08:20] <fabbione> at least to get main done for breezy will rock
[08:20] <jbailey> lamont-away: ping?
[08:20] <fabbione> i am running breezy atm on my buildd.. a suicide :)
[08:21] <jbailey> It's a good workout.
[08:21] <lamont> jbailey: 
[08:21] <jbailey> Tests glibc and such nicely.
[08:21] <jbailey> lamont: Hey.  Will you have time to test this binutils to see if it's enough love for HPPA as well?
[08:22] <lamont> it was building when I left the house, checking on its status now
[08:22] <fabbione> jbailey: btw.. i did also install an initramfs kernel on sparc...
[08:22] <lamont> I wish expect worked correctly
[08:22] <jbailey> Ah lovely. =)
[08:22] <fabbione> i will let you know tomorrow if it boots
[08:22] <jbailey> fabbione: Boots?
[08:23] <lamont> jbailey: once this build finishes, I'll upgrade said chroot and try building openexr, since that was the simplest failure case before
[08:23] <doko> jbailey: applying that for sparc and hppa conditionally? saves us the testsuite results for the other archs ...
[08:23] <fabbione> jbailey: i don't think we did ever test initramfs on sparc...
[08:23] <fabbione> jbailey: so this will be a test to do :)
[08:23] <jbailey> doko: Right.
[08:30] <lamont> jbailey: on the binutils upload, care to make it say build-depends: ... expect-tcl8.3 [hppa] , expect (>= 5.32.2-1) [!hppa]  ...
[08:30] <jbailey> tcl8.4 sucking wind for you?
[08:31] <jbailey> Actually, if this fixes your problem, we'll be disabling the testsuite for hppa anyway, so it doesn't much matter.
[08:31] <lamont> jbailey: rather, the kernel's failure to deliver SIGCLD under some particular and unknown conditions, is.
[08:31] <lamont> ah, true
[08:31] <jbailey> And tcl8.3 solves this?
[08:34] <lamont> jbailey: uh, yes.
[08:34] <lamont> and no, I don't know why.
[08:34] <lamont> if I did. we could fix the kernel bug.
[08:35] <lamont> which is affectionately (and incorrectly) known as "the expect bug"
[08:35] <lamont> see the gcc-4.0 build-deps... doko was the one who figured out the hack (and created expect-tcl8.3)
[08:37] <jbailey> We should have a race, who gets working Signals first, the Hurd or parisc-linux? =)
[09:56] <lamont> jbailey: test build running now
[09:56] <jbailey> lamont: Sorry, I don't bump it until I'm ready to upload.  This one unconditionally applies it everywhere.
[09:57] <lamont> right
[09:57] <jbailey> It's really one of those "please only add me by hand in a chroot and you should have to think reasonably hard about doing this"
[09:57] <lamont> otoh, making it -3ubuntu1.0 would be a clue...
[09:57] <lamont> heh
[10:12] <lamont> jbailey: SCORE!!!!
[10:12] <lamont> checking for strerror... yes
[10:12] <lamont> checking for compress in -lz... yes
[10:12] <lamont> checking for ios_base support in C++ standard library... yes
[10:20] <jbailey> lamont: Izgood?
[10:24] <lamont> sim
[10:25] <jbailey> doko: Awake?
[10:25] <jbailey> doko: It seems sparc and hppa both appreciate this binutils better than a swift kick in the teeth.
[10:26] <jbailey> doko: So I'd like your final blessing to upload with this patch only applying for hppa and sparc, and with the testsuite disabled on those archs.
[10:26] <doko> jbailey: sounds fine.
[10:26] <doko> maybe after the release we should change the name from binutils-static to ld-static
[10:27] <jbailey> I named it that in case it needed to grow more than just ld.
[10:27] <doko> then let's watch how it grows in breezy+1 :-)
[10:44] <lamont> so then.  jbailey: yes, I'm happy with the new binutils
[10:46] <lamont> jbailey: how soon should I expect an upload?
[10:47] <jbailey> I'm trying to decide if I want to pick a quote for the top of the changelog file...
[10:57] <jbailey> Feh, my mind is blanking on how to do ifneq ($(DEB_HOST_ARCH),sparc) or ifneq ($(DEB_HOST_ARCH),hppa)
[10:57] <jbailey> Do I have to nest them?
[11:02] <jbailey> Solved a different way.
[11:08] <jbailey> elmo: Around?
[11:26] <lamont> ifeq ($(arch),$(findstring $(arch),hppa sparc))
[11:26] <jbailey> lamont: What's the best way to bootstrap a breezy chroot these days for hppa from a running sid system?
[11:26] <lamont> I _think_ you can just debootstrap it...  but no.
[11:26] <lamont> libgcc2 isn't in main, so life sucks
[11:26] <jbailey> Well, the ubuntu debootstrap scripts aren't in Debian, are they?
[11:26] <lamont> debootstrap hoary from people.ubuntu.com/~lamont/ubuntu-hppa/tree and then dist-upgrade
[11:26] <jbailey> libgcc2? =(
[11:27] <jbailey> You have an abi event with libgcc?
[11:27] <jbailey> s/have/had/
[11:27] <lamont> well, doko did.
[11:27] <lamont> http://people.ubuntu.com/~lamont/hoary.buildd
[11:28] <lamont> debootstrap with that script
[11:30] <jbailey> When this is done, can I just dist-upgrade off of archive.ubuntu.com?
[11:30] <lamont> no.  ports.ubuntu.com
[11:30] <lamont> deb ports.ubuntu.com/ubuntu-ports breezy main restricted ....
[11:30] <lamont> deb-src archive.ubuntu.com/ubuntu breezy main restricted ....
[11:31] <lamont> hppa-hacks/expect-tcl8.3-dev_5.43.0-2_hppa.deb
[11:31] <lamont> hppa-hacks/expect-tcl8.3_5.43.0-2_hppa.deb
[11:31] <lamont> hppa-hacks/expectk-tk8.3_5.43.0-2_hppa.deb
[11:31] <lamont> hppa-hacks/gcc-3.4-hppa64_3.4.4-5ubuntu1_hppa.deb
[11:31] <lamont> hppa-hacks/libgcc2_4.0.1-4ubuntu4_hppa.deb
[11:31] <lamont> hppa-hacks/palo_1.9_hppa.deb
[11:31] <lamont> that's what I have in a local repository which sources.list.main includes
[11:31] <jbailey> Err..
[11:32] <lamont> those are copies/symlinks to the universe packages
[11:32] <jbailey> Okay, so if I add universe, will I be fine?
[11:32] <lamont> right
[11:32] <jbailey> Oh good. =)
[11:32] <jbailey> Aside from the kernel not booting, is ia64 releasing with Breezy?
[11:32] <lamont> note that if you have universe debs in sources.list, and you build a main package, you may get incorrectly-satisfied build-depends...
[11:32] <lamont> ia64 is tracking breezy nicely
[11:33] <jbailey> Right.  At this point I just want to make sure that the binutils package actually applies the patch on hppa correctly.
[11:33] <lamont> except for that little kernel issue
[11:33] <jbailey> The testsuite disabling I've tested.
[11:33] <lamont> ok
[11:33] <jbailey> Hmm.  debootstrap seems stuck on I: Checking component main on http://people.ubuntu.com/~lamont/ubuntu-hppa/tree...
[11:34] <lamont> it should be happy...
[11:34] <lamont> you told it hoary, yes
[11:34] <lamont> ?
[11:35] <jbailey> debootstrap --variant=buildd hoary hoary http://people.ubuntu.com/~lamont/ubuntu-hppa/tree
[11:35] <jbailey> And I updated the hoary.buildd from the location you have.
[11:35] <jbailey> s/have/game/
[11:35] <jbailey> GAVE
[11:35] <jbailey> DAMMIT
[11:35] <jbailey> ROAR
[11:35] <jbailey> don't mind me.
[11:36] <lamont> lol
[11:41] <jbailey> Oh, it's done something useful.
[11:41] <jbailey> How nice of it.
[11:49] <elmo> jbailey: ?
[11:49] <jbailey> You don't have lvm2 or klibc-utils on hppa yet? =( 
[11:49] <jbailey> hmm
[11:50] <jbailey> elmo: I've added some stuff to make it possible to specify per-arch which ones should disable the testsuite, and also to make it honour DEB_BUILD_OPTIONS=nocheck
[11:50] <jbailey> elmo: I'd like you to consider these for Debian as well.  Will you take a look at the Ubuntu packae that I upload, or do you need a patch in the BTS?
[11:50] <elmo> jbailey: azeem made a patch for the latter already which is in the BTS
[11:51] <elmo> I'm sort of disinclined to make it easy to disable the test suite per-arch, but I'll see what it looks like
[11:51] <elmo> vorlon + drow approved binutils cvs for sid, btw
[11:51] <jbailey> Well, I wrapped the testsuite stuff in ifeq ($(CHECK),), and at the top I set CHECK based on DEB_BUILD_OPTIONS
[11:52] <jbailey> the sideeffect is that you can trivial define CHECK whenever you'd like for other reasons.
[11:53] <jbailey> cool on the cvs binutils for sid.
[11:53] <jbailey> It'll be nice to get this cleaned up everywhere.
[12:00] <jbailey> lamont, elmo: While you're both hereish, any luck on hppa build logs?