=== BenM is now known as bhome [12:31] jbailey: DUDE [12:32] elmo: do you want to demote 3.3 from main to universe? [12:36] elmo: eh? [12:37] jbailey: are you going to undo that CDBS brain damage where the build-depends get generated ont he fly? [12:37] doko: a bunch of stuff still uses it? [12:38] elmo: What jrg and I talked about for cdbs2 (since it's too late to undo it for cdbs1) is that we'll still provide the facility, but that it will be wired such that if debian/control is out of date at build time, it'll just fail the build. [12:38] well, yes. so we'll have 3 in main and the 4th in universe [12:39] elmo: cdbs is in a good position to know information for calculating build-deps, but it needs to really not happen at buildd time. [12:40] jbailey: why's it too late? :P [12:40] elmo: Things are using it. [12:40] SATAN WORSHIPPERS [12:40] elmo: dh_make --cdbs for some reason uses it by default. =( [12:40] Just insanity. [12:40] ARGH [12:41] those people are THROWING BABY JESUS ONTO A PYRE AND LAUGHING ALL THE WHILE\ [12:41] Wow, elmo is channelling a mix of mjg59 and overfiend. === jbailey runs. [12:42] jbailey: why is it _ever_ adding build-essential tho? [12:42] ignoring for one moment the whole auto-update insanity === jbailey glances at the ubuntulog bot and sensors what he was going to say. [12:44] nfi [12:44] Wasn't my hack. =( It went quite a bit further than I expected it would, even under Robert Millan's care. === jbailey wanders off to the gym singing "Sam Hall" [12:54] jbailey: censors [12:55] sensoring is something much more touchy-feely === bsleep is now known as BenM [04:18] lamont: Right. I get that wrong because of a church conference I went to called 'unsensored' as a pun. [04:20] heh === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain [06:40] morning === Riddell [jr@muse.19inch.net] has joined #ubuntu-toolchain === BenM is now known as bzzz === chmj [~d3vic3@dumbledore.hbd.com] has joined #ubuntu-toolchain [11:07] infinity/lamont: d-i daily auto-builds are being flaky [11:08] as in, not turning up for a random 2/4 arches [11:11] elmo: did you install the ppc64 kernel on any of the buildd?? [11:11] fabbione: not yet - do you need me to? [11:12] elmo: perfect. no don't :) [11:12] i found a regression in building ppc kernels on ppc64 running machines [11:12] hehe [11:12] elmo: somehow it breaks headers build [11:13] and i didn't have the time to track it down yet [11:13] it took a while to understand why davis was FTBFS while it was ok on the buildd :) [11:14] elmo: if you have time i need a dist-upgrade on breezy/{concordia,davis,hally} and breezy-i386/concordia to get all the latest kernel-packages, gcc-3.4 [11:14] btw has davis' stability continued? [11:14] elmo: yup.. not a single problem other than that regression [11:15] woo [11:15] elmo: i pushed it up to -j300 with no problems [11:15] the ppc scheduler >> * [11:15] and you know i am good at pushing machine.. given nagios waking you up everytime i upload :P [11:23] fabbione: done [11:24] elmo: thanks === fabbione starts a build orgy === jbailey [~jbailey@CPE00501836c657-CM014260028338.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain [12:44] fabbione: ping? [12:45] jbailey: pong [12:45] fabbione: Tell me good things about glibc on sparc? =) [12:45] it's building :) [12:45] it's half way trough [12:46] Can you peek into the source directory? It would be interesting to see if it's finished the first pass, and if so, was it succesful or did it error? [12:46] jbailey: i am saving the entire buildlog [12:46] so you can look at it when it's done [12:47] actually.. you can check yourself too :) [12:47] you can still ssh to the sparcbuildd :) [12:47] it's building in chroots/manual/home/sparcbuildd/ [12:48] BAh. Cat just threw up for the 6th or 7th time this morning. =( [12:49] *sighs* === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain [12:51] fabbione: First pass looks good as sparcv9 minimum and nptl [12:52] jbailey: cool [12:52] It's way less than halfway done, though, I think. === doko [~doko___@dsl-082-082-201-214.arcor-ip.net] has joined #ubuntu-toolchain [12:52] yeah it's around half way [12:52] the previous log was 43MB [12:52] this is around 20 [12:52] but if you disabled stuff around, i expect the log to be smaller [12:52] Right, I disabled one full pass. [12:53] I don't see any reason why the v9b build would fail from here. sparc64 didn't change yet. [12:54] so no NTPL? [12:54] Not for sparc64, yes for the others. [12:54] ok [12:54] It needs binutils 2.16.1 and some extra patches. [12:55] ah ok [12:55] So early next week maybe... [12:55] yeah no rush dude :) [12:57] Yeah, but it would be nice to have kickass sparc support in a mainstream distro. =) [01:01] than we should get more buildd's :) [01:01] we are in catch 22 right now [01:02] but i am getting a nice buildd from maswan [01:03] i hope to get it up and running soon [01:03] fabbione: I can probably fire up this u5 to help along, too, after I move. [01:03] jbailey: with the u5 it would be better to test the installer and so on [01:03] we really need somebody to test :) [01:03] and i can do it when i am not building [01:05] 'k =) [01:05] I almost brought an E250 home with me from my last place of employment, but the thing is too loud, I never would've run it. [01:05] when are you going to finish your relocation btw? [01:05] yeah i know the e250 :) [01:05] nice toy [01:05] I do my actual move on July 4th. [01:06] It's to a new city, so I'm not doing one of these 'spread out over 4 days' types of moves. =) [01:06] lucky you [01:06] it took me ages to change city [01:06] moving stuff myself [01:07] <-- has to move also [01:07] fabbione: We're loading everything into one big truck, driving for 7 hours, and unloading. [01:08] jbailey: that's the best.. really [01:08] one shot and that's it [01:08] Yup. [01:08] And we've hired people in the remote city to handle getting it from the truck into the appt. [01:09] We'll be way too tired after all that to figure out how to get furniture up two tight flights of stairs. [01:09] wise :) [01:09] I've gained so many things over the years. I can only hope that wisdom is one of them. ;) [01:10] s/hope/pray... beg..../ [01:10] ehhehe [01:38] fabbione: Neat, v9b compiled fine. [01:43] jbailey: cool [01:45] I wonder how much difference there really is between v9 and v9b [01:46] no clue :) [01:46] that's your playground :) [01:46] Fabio, my friend... [01:46] *you're* the sparc port guy. =) [01:46] jbailey: you are the glibc guys :P [01:47] I've asked someone. I'm curious if just the hwcap was different, so v9b systems weren't picking up the v9 optimisations. [01:47] It would be nice if we could drop that pass completely, too. [01:48] And then my last big dream for sparc: Kill the gcc wrapper. =) [01:50] ahhaha === Monkness [bananas@hamilton-maki.res.uct.ac.za] has joined #ubuntu-toolchain === Monkness [bananas@hamilton-maki.res.uct.ac.za] has left #ubuntu-toolchain [] [02:23] jbailey, [03:57] chmj: Here === bhome is now known as BenM [04:12] I forgot what I wanted to say [04:14] I've started to give comments after the word ping now so that I stop forgetting. =) [04:14] good idea === lamont__ [~lamont@15.238.6.137] has joined #ubuntu-toolchain === jbailey [~jbailey@CPE00501836c657-CM014260028338.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain [04:50] Bah, strip/tls fuckage. === jbailey hunts. [05:34] lamont / lamont__ / infinity: Is it possible to tell what versions of the base packages were in the buildd chroot at package build time? [05:35] I'm looking for binutils in this case. [05:35] build-essential Depends are in the log file [05:36] Checking correctness of source dependencies... [05:36] Toolchain package versions: libc6.1-dev_2.3.2.ds1-20ubuntu11 linux-kernel-headers_2.5.999-test7-bk-17 gcc-3.3_1:3.3.5-8ubuntu2 g++-3.3_1:3.3.5-8ubuntu2 binutils_2.15-5ubuntu1 libstdc++5_1:3.3.5-8ubuntu2 libstdc++5-3.3-dev_1:3.3.5-8ubuntu2 [05:36] for example [05:37] Poifect, thanks. [05:37] (I _THINK_ it pulls the list of packages-to-dump-versions-for from build-essential.... [05:37] jbailey: and you asking is pretty much the reason it's there. [05:38] well, for correct values of 'you'... [05:38] It may very well have been me at some point in the past. =) [05:39] Hmm, no ia64 success or fail message for http://people.ubuntu.com/~lamont/buildLogs/g/glibc/2.3.5-1ubuntu6/\ === lamont__ goes looking === jbailey whines about there being no hppa or sparc ones there either, but at least that's expected... === jbailey hides. [05:40] buildLogs/g/glibc/2.3.5-1ubuntu6/glibc_2.3.5-1ubuntu6_20050614-1141-ia64-successful.gz [05:40] uh.... [05:41] _AMD_64, otoh... [05:42] Umm. Right. === jbailey smokes another one. [05:42] amd64 is only 29 hours into the build... do you suppose it hung? [05:43] GCONV_PATH=/build/buildd/glibc-2.3.5/build-tree/amd64-libc/iconvdata LC_ALL=C /build/buildd/glibc-2.3.5/build-tree/amd64-libc/elf/ld-linux-x86-64.so.2 --library-path /build/buildd/glibc-2.3.5/build-tree/amd64-libc:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/math:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/elf:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/dlfcn:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/nss:/build/buildd/glibc-2.3 [05:43] .5/build-tree/amd64-libc/nis:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/rt:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/resolv:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/crypt:/build/buildd/glibc-2.3.5/build-tree/amd64-libc/nptl /build/buildd/glibc-2.3.5/build-tree/amd64-libc/inet/test-ifaddrs > /build/buildd/glibc-2.3.5/build-tree/amd64-libc/inet/test-ifaddrs.out [05:43] OK... so can we just kill test-ifaddrs???? [05:43] Bah, 29 hours is normal! [05:43] infinity: it's not an m68k box [05:43] glibc_2.3.5-1ubuntu6_20050614-1139 00:20:50 (12 entries, sigma 00:04:02) [05:43] (It's not normal on m68k, either... glibc isn't that big) [05:43] infinity: you don't build for m68k ... [05:44] lamont__: Right, or possibly just fix the kernel. [05:44] lamont__: But yeah, I think I'll prune test-ifaddrs [05:45] (Also, sbuild doesn't pull the list from build-essential, since that wouldn't work at all) [05:45] You would just get the gcc/g++ metapackages, not the current gcc-X.X... [05:46] And no binutils. [05:46] On the flipside, you'd get dpkg-dev, which I've been considering adding to the list to log ever since keybuk started seriously screwing with us. :/ [05:46] foreach $name (@main::toolchain_pkgs) { [05:46] Bah, busybox-cvs fails because of old binutils. (Me cries at the thought of wondering why it worked before....) [05:46] infinity: *please* [05:47] @toolchain_regex = ( 'binutils$', 'gcc-[\d.] +$', 'g\+\+-[\d.] +$', 'libstdc\+\+', 'libc[\d.] +-dev$', 'linux-kernel-headers$' ); [05:47] Yeahp. That's the one. [05:48] I'll add dpkg-dev now === jbailey does a snoopy dance! [05:48] then pester elmo about breezy-test, and push a config this week sometime [05:48] Thanks. I need to branch your arch repo and start merging in meuro's changes. [05:48] yes. [05:48] neuro, too. [05:48] upstream neuro is there, so it's really a 3-way merge [05:49] jbailey: want any more packages while I'm there? [05:50] I don't think so, nothing else has been getting churn that usually sits in the chroots that I can think of. [05:50] note that we do a dist-upgrade of all the chroots starting shortly(heh) after 0215 DCT [05:52] and that's still pending on yellow, because of glibc. [05:53] glibc_2.3.5-1ubuntu6_20050614-1139 02:33:52 (13 entries, sigma 07:59:40) === lamont__ giggles [05:53] GO test-ifaddrs [05:53] Joy [05:53] infinity: pushing a mirror to chinstrap now [05:54] Yeah, I'll prune it next time around. [05:54] I'll also bump the build-dep to 2.16 to avoid the strip error. [05:54] and fix hppa's missing directory. :-) [05:56] mirror fresh === infinity ponders heading to bed soon and getting back on a "correct" sleep schedule... [05:56] lamont__ : Huzzah. [05:57] hppa finally passed sparc for %installed [05:57] which reminds me... [05:58] lamont__: Wow. [05:59] jbailey: actually, I've been watching #installed the last couple of days... [06:00] hppa:Total 3342 Installed. [06:00] hppa:Total 2842 Needs-Build. [06:00] hppa:Total 6184 [06:00] sparc:Total 3312 Installed. [06:00] sparc:Total 2905 Needs-Build. [06:00] sparc:Total 6217 [06:00] Nice! [06:00] Are thos eboth on track for breezy then? [06:00] that is, it's a relative-suckage thing [06:01] I really need to make some time to see about arts (and 4? others) ICE. [06:01] which kinda blocks a large chunk of main [06:01] Bah, just not-for-us them... [06:01] ;) === lamont__ goes christmas shopping... back in 20-40 minutes === infinity checks the date. [06:09] You have to admit... [06:09] A Christmas tree is probably really cheap right now. [06:10] Much more unexpected when you chop down the one in the neighbour's front yard in July... [06:10] Erm.. June, even. [06:10] While yelling 'Merry fucking christmas!' [06:13] fabbione: Around? [06:18] fabbione: It appears to build fine. the make check sequence didn't run since sparc != sparcv9. I'll poke that in the next upload. [07:37] jbailey: it builds, but it doesn't create the .debs [07:37] jbailey: so it's FTBFS [07:38] the overall is still exactly where i did show you [07:38] lamont__: eh.. hopefully numerbs will change soon.. i am getting another buildd :) [07:39] fabbione: ditto. :-) [07:39] although I may have to wait for monday before I can scrounge good hardware [07:39] i will get it sometime during this week [07:39] so i can finally dedicate one machine to main and one to universe :) [07:45] fabbione: Is that the usual split? [07:45] jbailey: ? [07:46] jbailey: you mean for building? [07:46] Yes, sorry. =) [07:47] jbailey: yeah [07:48] cp: cannot stat `./debian/tmp-libc/usr/lib/librpcsvc.a': No such file or directory [07:48] wha? [07:49] jbailey: funny eh? [07:49] Oh! [07:49] That'll be fixed in the next build too. =) [07:49] It doesn't build any of the sunrpc stuff when cross compiling. [07:50] uh ? cross compiling? [07:50] i don't want to know === fabbione covers his ears with the hands and start yelling TRALLA LA' TRALLA LA' [07:50] I set the libc_configure_target=sparcv9-linux [07:51] _build was still set to sparc-linux [07:51] different processors == cross compiling. [07:51] Even if its on the same glibc. [07:51] err gcc. [07:51] TRALLLAAAAAAAAAAAAA LAAAAA [07:51] TRALLLAAAAAAAAAAAAA LAAAAA' [07:51] TRALLLAAAAAAAAAAAAA LAAAAA' [07:51] eheh ok [07:51] stick the source somewhere [07:51] so i can do a test build [07:52] It'll be pulsed out at :33 (Unliekly to make :03), mind just fetching it then? [07:52] jbailey: oh you already upload? [07:52] than it's ok [07:52] i understood you wanted me to test before [07:52] No, I'm still running a test. [07:53] I'd just like you to start it before you go to sleep. [07:53] ok [07:53] That way when I get up in the morning, I'll have testsuite results. [07:55] hm ok [07:55] i guess i will have to come later to start it [07:55] Okay, lemme send you the source then. [07:55] Sheesh [07:55] ;) [07:55] jbailey: that's fine.. i can wait [07:56] I don't want your wife to hurt me for making you come back. [07:56] it's just the mirror round-robin adds at least one hour [07:56] Can I safely drop the .diff.gz and .dsc in ~ on the sparc buildd? [07:56] yup [07:57] YAR BLOODY STUPID KERNEL BUG [07:57] lamont__: Well, it doesn't seem to be test-ifaddrs that's the problem. [07:57] lamont__: Removing it, I happened to get the failure on the test immediately after it. [07:57] And lucky me on ppc64, it seems to prevent me from running sudo [07:58] fabbione: Is there any way to catch state information at this point, or should I just hit the power button like I usually do? [07:58] jbailey: so nuke the test _before_ test-ifaddrs :-) [07:58] lamont__: Ayup =) [07:59] jbailey: ctrl+sysrq+s ? [07:59] That might explain why noone's been able to hunt this down, though. [07:59] bah i can't never remember the sequences [07:59] Which one is sysrq on a mac keyboard? [07:59] MEHHH [07:59] no idea [07:59] i think it's the same where it says: "HIT HERE TO POWER OFF THE MACHINE" [08:00] i never had a ppc [08:00] so i dunno [08:00] but i will be glad to host a few here at home :) [08:01] fabbione: So, yeah, best to come back. This package isn't baked yet. [08:01] How long until you go to bed? [08:01] 2/3 hours [08:01] but i can still build it if you put the sources on vultus5 [08:01] Okay good, so I have time for another full run with the testsuite to make sure there isn't some place after this that I need to remove the test too. [08:02] ok [08:02] Yeah, I'll do that. The .diff and .dsc safe to drop in ~ ? [08:02] yup [08:12] fabbione: I've uploaded it. [08:13] fabbione: err, scp'd it rather. [08:13] jbailey: rocking [08:13] yeah [08:13] fabbione: With any luck this will be what I upload, even. =) [08:13] It's off of a 'debuild', not debuild -S [08:14] i will build it.. if you add or change i will just rebuild [08:14] otherwise we can test/upload that set of packages [08:14] Thanks. [08:14] Yup [08:14] i have no issues building glibc.. [08:15] it's relatively fast :) [08:15] What's the usual time that it takes? [08:16] 7 hours more or less [08:16] but it might take longer [08:16] Nice, so it might be finished before I go to bed. [08:17] i doubt.. it takes 7 hours when it builds alone [08:17] it's running in parallel now [08:17] but it has been ccached with the previous run [08:17] so it's a tricky build :) [08:17] anyway it's already building [08:20] Cool [08:20] I'm off to get food, back in a bit. [08:20] sure later [08:32] *Sigh* [08:32] Killing the previous test doesn't do it either. [08:32] Then test-ifaddrs kills it. [08:32] I've never hit this bug so many times in a row [08:35] i am off :) === jbailey disables that patch and builds this on a power4 kernel instead === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-toolchain === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain === jbailey [~jbailey@CPE00501836c657-CM014260028338.cpe.net.cable.rogers.com] has joined #ubuntu-toolchain