[02:12] <lamont> gcc-opt: forcing -O1
[03:03] <jbailey> lamont: Can you tell me if hppa has succesfully built klibc?
[03:05] <lamont> on debian?
[03:05] <lamont> or ubuntu?
[03:05] <lamont> latest, or ever?
[03:06] <lamont> make[1] : Entering directory `/build/buildd/klibc-1.0.14'
[03:06] <lamont> MCONFIG:90: klibc/arch/parisc64/MCONFIG: No such file or directory
[03:06] <lamont> make[1] : *** No rule to make target `klibc/arch/parisc64/MCONFIG'.  Stop.
[03:07] <lamont> fix your uname -m crap to strip the 64...
[03:09] <jbailey> Okay.
[03:10] <jbailey> Just that initramfs-tools 0.11 works for booting off of a harddrive, software raid1 and nfsroot now.
[03:11] <jbailey> Is 2.6.13 really does drop devfs, then this will be it from then on. =)
[03:11] <jbailey> s/Is/If/
[03:14] <jbailey> Should I map hppa* to hppa, or just hppa64 ?
[03:16] <lamont> parisc* -> hppa
[03:16] <lamont> jbailey: way cool.
[03:17] <jbailey> Hmm.
[03:17] <lamont> and yeah, good that it's working that well.
[03:17] <lamont> well, let me rephrase that...
[03:17] <jbailey> No, the kernel considers your arch to be "parisc"
[03:17] <lamont> kernel knows it as parisc or parisc64, depending on which kernel is running
[03:17] <lamont> and in fact, has an ifdef that appends the 64...
[03:17] <lamont> user space knows it as hppa
[03:18] <lamont> so whichever is appropriate for your environment...
[03:18] <jbailey> klibc is using this to grab the kernel headers.
[03:18] <jbailey> for syscalls and such.
[03:18] <lamont> hrm...
[03:18] <lamont> package names have hppa{,64}{,-smp} versions, iirc
[03:18] <jbailey> But it looks like there isn't really a separate parisc64 header set.
[03:19] <lamont> there isn't - it's purely config diffs
[03:19] <jbailey> I'm going to go with s/parisc*/parisc/ then.
[03:19] <lamont> sounds correct for kernel space.
[03:19] <jbailey> Have a good one, I'll push this to upload now.
[06:12] <lamont> hrmpf.  hppa is at 56.5%, sparc at 57.9% (installed)
[06:27] <fabbione> morning
[06:27] <fabbione> lamont: :)
[06:42] <lamont> fabbione: but my builds go faster now, since I'm forcing -O1 for a max optimization level in g{cc,++}-4.0.
[06:43] <lamont> otoh, ccache is slowly flushing
[06:43] <fabbione> i am not using ccache on the external buildd
[06:44] <fabbione> and i leave standard opts
[06:44] <fabbione> but well.. i am ok at this speed :)
[06:45] <lamont> dpkg: error processing tetex-bin (--configure):
[06:45] <lamont>  subprocess post-installation script returned error exit status 30
[06:45] <lamont> stupid tetex
[06:46] <lamont> gonna have to fix that tomorrow
[06:46] <fabbione> eh good idea :)
[06:47] <lamont> already done, it appears
[06:47] <lamont> tex/tetex-bin_2.0.2-30ubuntu3: Needs-Build [optional:out-of-date] 
[06:49] <lamont> and depwait gtk+2.0 and linux-source-2.6.12 on the newer tetex-bin.  sigh
[06:50] <fabbione> oh l-s-2.6.12 is FTBFS on sparc :)
[06:50] <fabbione> i need to fix that ;)
[06:50] <lamont> eventually, at least, eh?
[06:50] <fabbione> probably, maybe, mostlikely :)
[06:51] <lamont> the -O1 is a workaround for a g*-4.0 regression in the optimizer.  annoying
[06:51] <fabbione> ah
[06:52] <lamont> I'd rather be able to leave it at 0..3
[06:52] <fabbione> i wonder if that also fixes some of my FTBFS
[06:52] <lamont> might... although it's an hppa-specific regression
[06:52] <fabbione> ah ok
[06:53] <lamont> it would appear that some code misbehaves when it compares two function pointers for equality and gets told 'not equal', when in fact they are.
[06:53] <fabbione> ah
[06:56] <fabbione> YEAH
[06:56] <fabbione> .12 is out
[06:58] <lamont> so how soon will we upload 2.6.12-1.1?
[06:59] <fabbione> that will be mostlikely 2.6.12-2.1
[06:59] <fabbione> or do you truely believe that from rc6 to final they did not break the ABI?
[07:00] <lamont> ah, we started being good and tracking, eh?
[07:00] <fabbione> yes.. because we moved .12 in main yesterday :(
[07:00] <lamont> we really should deliver the abi files as part of one (or each) of the kernel packages
[07:00] <lamont> rather than scaping them from the log files
[07:00] <fabbione> yeah that's in the spec..
[07:01] <fabbione> we need to think deeper about it :)
[07:02] <lamont> yeah - the proper place to put it would be linux-source, but that's arch: all --> no can do
[07:03] <lamont> maybe add a package that is nothing but the abi spec files?  then you can install that and linux-source-2.6.NN, and have everything you need to build a kernel properly
[07:05] <lamont> g'night
[07:05] <fabbione> re
[07:05] <fabbione> there is the b-d problem to do that
[07:05] <lamont> hrm.. /me needs to deal with gcc-opt it appears.
[07:06] <fabbione> good night
[07:06] <fabbione> what's wrong with it?
[07:06] <fabbione> it's a global problem or local?
[07:06] <lamont> -O gets misparsed
[07:06] <lamont> only if you restrict to < 2
[07:06] <lamont> is it a problem, that is
[07:06] <fabbione> ah ok
[07:07] <lamont>                     else if (c=='\0')
[07:07] <lamont>                         OptLevel=1;
[07:07] <lamont> DOH!!!
[07:07] <lamont> -O is == -O2, yes?"
[07:07] <fabbione> whoops
[07:08] <lamont> that hurts
[07:14] <lamont> gcc -O
[07:14] <lamont> gcc-opt: forcing -O1
[07:14] <lamont> gcc-4.0.gcc-opt: no input files
[07:14] <lamont> much better
[07:58] <fabbione> r/sc
[07:58] <fabbione> meh jb is not around
[02:43] <fabbione> Total 1594 package(s)
[02:43] <fabbione> Needs-Build :)
[02:44] <fabbione> Total 621 package(s)
[02:44] <fabbione> Building
[02:44] <fabbione> -20 in the buildd
[02:44] <fabbione> 601 in DepWait/FTBFS
[02:44] <fabbione> not really encouraging
[02:44] <fabbione> but most of them are missing B-D on X and DepWait
[03:10] <fabbione> hey jbailey 
[03:17] <jbailey> Fabio!
[03:17] <fabbione> jbailey !
[03:17] <jbailey> Tell me about klibc on sparc.
[03:17] <fabbione> i did prepare a chroot with the new libc on sparc
[03:18] <jbailey> I want it working on sparc and hppa today. =)
[03:18] <jbailey> Ooo ooo ooo!
[03:18] <fabbione> klibc is FTBFS
[03:18] <fabbione> i think i told you already
[03:18] <fabbione> now.. i have the chroot.. what do you want me to test with it?
[03:20] <jbailey> Assume since you can get into it that shell and ls and friends work.
[03:21] <jbailey> +I
[03:21] <fabbione> yeah they do
[03:22] <jbailey> *lol* alright then. =)
[03:22] <jbailey> I've only been awake for 6 or 7 minutes. =)
[03:22] <fabbione> that's fine :)
[03:22] <jbailey> Did you seriously just upload?
[03:23] <fabbione> no.. i moved it to the upload dir...
[03:23] <doko> s/g/k/
[03:23] <fabbione> jbailey: i did test apt and such
[03:23] <jbailey> t would be nice if you could test something that uses getpid
[03:23] <jbailey> Like top
[03:23] <fabbione> hmm
[03:23] <fabbione> sure
[03:23] <jbailey> REmember that this instance has the known clone bug in it.
[03:24] <jbailey> Next upload will have that fixed (and hopefully some langpack love for pitti)
[03:24] <fabbione> i trash langpacks
[03:24] <fabbione> there is no way LP can access external buildds
[03:24] <jbailey> Sounds like a bug.
[03:25] <fabbione> doko: what are you going to upload?
[03:25] <fabbione> jbailey: LP is designed to access only DC buildds
[03:25] <jbailey> Then it's broken.
[03:25] <jbailey> Becuase ports will never be DC.
[03:26] <fabbione> jbailey: rule is: given port foo with a community, Mark will buy buildds to put at the datacenter
[03:26] <fabbione> sparc and hppa are 2 exceptions
[03:26] <fabbione> very special cased because it's lamont and me running them
[03:27] <jbailey> How big does the community need to be? =)
[03:27] <fabbione> jbailey: at least 50/100 users?
[03:27] <jbailey> Oh, that's easy enough.
[03:27] <jbailey> Hell, we have that many Hurd users. =)
[03:27] <fabbione> we can start faking email addresses and domains to mail bomb Mark asking for them :)
[03:28] <doko> fabbione: 4.0.1
[03:28] <fabbione> but hounestly.. i did stop Mark from buying sparc buildd's
[03:28] <fabbione> doko: dude.. get it right or i might knock on your door this evening
[03:29] <jbailey> AH did 4.0.1 release?  I'm a couple days behind on the toolchain lists.
[03:29] <doko> RC2
[03:29] <jbailey> BUt initramfs now works for the three cases that I claim it should.
[03:29] <jbailey> head is in stage 3 now, right?
[03:30] <fabbione> jbailey: top seems to work fine
[03:30] <fabbione> anything fancy you want me to test?
[03:31] <jbailey> You don't have a spare machine, right?
[03:31] <fabbione> nope
[03:31] <jbailey> Oh well.  I would suggest booting with it usually.
[03:31] <jbailey> Just because the change is so drastic.
[03:32] <jbailey> GA and upload then. =)
[03:32] <fabbione> yeah who cares :)
[03:32] <fabbione> there are probably 4 sparc users out there
[03:34] <fabbione> jbailey: in case it works.. you have to add a changelog entry for libc in debian:
[03:34] <fabbione> "NTPL on sparc is dedicated to that Kamikaze of fabbione"
[03:35] <fabbione> ;)
[03:35] <jbailey> It won't happen in Debian until the sparc porters say clearly that they want it to happen.
[03:35] <fabbione> ahaha
[03:35] <fabbione> that will never happen
[03:35] <fabbione> no.. really..
[03:35] <jbailey> Exactly.
[03:35] <fabbione> they can't even agree on what kind of toilet papers to use...
[03:35] <fabbione> generally speaking i mean
[03:35] <jbailey> Right.
[03:36] <fabbione> Benc and davem are pretty sane tho
[03:36] <jbailey> benc's last directive was 'don't to nptl'
[03:36] <jbailey> A quick poll of the other Debian glibc folks say that they are not in a rush to move away from LinuxThreads.
[03:36] <jbailey> So I do not expect to roll out nptl in Debian the way I have in Ubuntu.
[03:37] <fabbione> you can still use a very good argument
[03:37] <fabbione> "I overabused my Ubuntu powers to get it tested on 6 arches and it works...."
[03:37] <jbailey> 5.  Hppa isn't there yet. =)
[03:37] <fabbione> ".. but it's your call to decide"
[03:37] <fabbione> ah right.. hppa is even behind on the buildd!
[03:37] <jbailey> Right, and drow said "No.  Ubuntu may be able to get away with it, but Debian has a different situation"
[03:38] <fabbione> for once i am satisfied for not being the slowest arch :)))
[03:38] <jbailey> LOL
[03:40] <doko> ohh, I didn't know that infinity did join the porting team ;-P
[03:41] <fabbione> doko: no. i got the second buildd running
[03:41] <doko> no m68buntu :-/
[03:42] <fabbione> probably.. if we start m68buntu i will add 2 buildds here :)
[03:42] <fabbione> anyway it's saturday.. the sun is shining.. my wife has backed a cake
[03:42] <fabbione> cya later fellas
[03:43] <fabbione> Uploading via ftp glibc_2.3.5-1ubuntu7_sparc.changes: done.
[03:43] <fabbione> Successfully uploaded packages.
[03:43] <fabbione> ENJOY THE RIDE
[03:43] <jbailey> fabbione: Lovely.  Unblock glibc from the buildd then, please? =)
[03:43] <fabbione> i did already
[03:43] <fabbione> ;)
[03:43] <fabbione> i am off
[03:44] <jbailey> c'ya Fabio!
[10:24] <doko> lamont, infinity: please could you look at the build failure of db4.2 on i386? gij-4.0 is supposed to install the /usr/bin/java symlink as an alternative