=== doko_ [n=doko@dslb-088-073-071-016.pools.arcor-ip.net] has joined #ubuntu-toolchain === cjwatson [n=cjwatson@82-69-40-219.dsl.in-addr.zen.co.uk] has joined #ubuntu-toolchain [12:23] Listing ubuntu/feisty (DONE) 0/0 [12:23] y'all are slackers [12:23] :-) === mdz [n=mdz@george.kkhotels.co.uk] has joined #ubuntu-toolchain [12:45] cjwatson: is edgy-updates open? [12:45] or -proposed ... [03:00] So, what's the status of toolchain mangling, kids? [03:00] Anyone have uploads ready for me? [03:29] yes, for edgy-updates; sparc has problems === jb-home [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-toolchain [03:48] doko, fabbione: Whenever you guys wake up. I'm going to patch glibc such that it build-dep's on linux-libc-dev 2.6.19. Some defines went away, so I'll patch glibc to have them directly (and submit the patch upstream) [03:51] jb-home: So, am I going to do a manual bootstrap to get glibc in place, then do the rest of the toolchain (and kernel), then we'll upload glibc again with the correct build-dep? [03:51] infinity: Eh, you're awake earlier than expected. =) [03:51] It's almost noon... [03:51] glibc doesn't build without it. [03:52] So I'd start with the kernel upload. [03:52] I suggested that to Ben, and he doesn't want people usng the kernel, so wouldn't update linux-meta. [03:52] Oh, yeah, starting with the kernel seems sane, since it should build fine regardless. [03:52] Right. [03:52] Do we have a kernel upload for me? [03:52] Best to poke BenC. He has one that he was using to give us linux-libc-dev .debs from. === BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-toolchain === BenC was summoned [03:53] Woo. [03:53] BenC: How quickly can I get that linux-source-2.6.19 in edgy? [03:53] feisty, even. [03:53] That's going to take some brain mangling. [03:53] feisty, feisty, feisty. [03:53] I wonder how many changelogs I'll mess up in the first month... [03:54] I can upload tonight [03:54] Where "tonight" = ? [03:54] withing a few hours [03:54] Cool. Should do. [03:54] want me to get on it? [03:54] Please. :) [03:56] there's no lrm, and not touching linux-meta, and don't suggest to anyone to install it :) [03:56] infinity: It's only bad because you have the rights to approve uploads to edgy, don't you? =) [03:56] I don't have all the third-party drivers in place yet [03:58] jb-home: I can disable double-128 for sparc for now, but we shouldn't open the archive before we can fix that [03:59] doko_: Right, it seems like it's just a patch missing from the backport, yes? [03:59] jb-home: tell me which one ... [03:59] =) [04:00] Considering I had trouble *reading* the bloody assertion that was tripping... [04:00] I'm currently trying to work around the C++ regressions, (fixes of accepts-invalid-code :-/ ) [04:01] yeah, and davem doesn't know either [04:03] Does it happen with pristine CVS? [04:04] jb-home: s/edgy/feisty/ in the above? [04:04] jb-home: 4.1 doesn't support double-128, 4.2 does [04:04] Anyhow, if all the above babble is telling me we're not ready for the bootstrap, I can hold off. [04:04] I have other stuff to do right now anyway. [04:04] interim solution: make 4.2 the default on sparc and let fabbione fix the fallout bugs =) [04:05] infinity: No. As in, it's bad if you type edgy in the changelog, because you have launchpad god rights. You can accidentally make it go through. =) [04:05] BenC: But please, do get us the kernel upload soonish, so when jeff and doko are ready, we're good to go with linux-libc-dev. [04:05] infinity: re: hppa... [04:05] infinity: on its way, at 30k/sec, so give it some time [04:05] jb-home: I'd have to put edgy back in "development" before I could do that, which would be a lot of effort to work around a typo. [04:05] infinity: Are there feisty chroots there based off of dapper? [04:06] jb-home: There are no chroots for any arch, but yes, there will be an hppa/feisty chroot that's essentially a dapper chroot. [04:06] jb-home: just ronne/i386 [04:06] (No chroots intentionally, so uploads don't trigger builds accidentally) [04:06] infinity: Cool. So with any luck, it'll give us the kernel headers there. [04:06] Assuming linux-2.6.19 can build on dapper, yeah. [04:07] If not, this'll be a fun ride. [04:07] Oh, which it certainly can't. Forgot about that. [04:07] Although I can tell you which pieces need to be built. =) [04:07] If the only thing we need to pull in is linux-libc-dev, and one of you has already bootstrapped that on hppa, I'm all for cheating. [04:08] I have all the basic stuff to do it, just need Ben's source package. [04:08] Okay, cool. [04:08] I'm less concerned about hppa today anyway. [04:08] But if we can squeeze it in here and there and make it happen, cool. [04:55] infinity, jb-home: 5 more minutes to upload completion [04:56] BenC: \o/ [04:56] Thanks, dude. [04:56] Were Fabio's "The headers are so busted that the kernel's scripts/ directory can't even be compiled against the new linuc-libc-dev" concerns addressed? :) [04:57] yep, retested today with latest git [04:57] Awesome. Thanks. [05:02] done [05:02] morning guys [05:03] hey fabbione [05:03] so what's the situation now? [05:03] new kernel is up? [05:03] Hola Fabbione! [05:04] jb-home: glibc still doesn't build on ppc [05:07] fabbione: see the reply from davem, that's our major sparc problem [05:07] doko: is that long-double thing REALLY required ? [05:07] doko: i am going trough my emails right now... [05:09] jb-home: did you commit that patch? [05:09] fabbione: yes, see PR28701 [05:09] fabbione: No, I went to eat dinner. =) [05:10] doko: url? [05:10] jb-home: ok [05:10] jb-home: i am going to try to build glibc on ppc32 [05:10] fabbione: Starting by build now. [05:10] but that will take some time [05:10] fabbione: http://gcc.gnu.org/PR... [05:10] The patch commits, anyway. [05:10] I think this thing usually takes about 20 minutes to build. [05:11] doko: oh, is that the shortcut? I can never remember it. =) [05:12] infinity: got the NEW for l-s-2.6.19 [05:13] jb-home: we also still need to port the ubuntu locale patches to 2.5. in theory they could wait for the second upload.. [05:14] jb-home: what was Alzheimers first name? [05:14] doko: Is "I don't remember" a good answer? =) [05:14] fabbione: They didn't apply cleanly? I thought I had seen them in the patch set. [05:15] jb-home: i didn't even look at debian/patches YET [05:15] you know.. been doing only a merge for 8 hours or so [05:15] jb-home: yeah, ... [05:16] ppc32 fired up [05:22] doko: when is gcc-4.2 due to? [05:22] just branched [05:22] no release date [05:23] what does that mean in gcc world usually? [05:23] 3 months? 20 months? [05:30] maybe 3, maybe 5 months [05:30] wow... [05:31] jb-home: building glibc on ppc32 basically froze my entire machines except the ssh sessione where it's building [05:31] fabbione: Fucked kernel? [05:31] jb-home: possibly.. [05:31] Linux daltanius 2.6.19-1-powerpc #1 Mon Oct 23 05:18:52 CEST 2006 ppc [05:31] it's rc2 === fabbione reboots in .17 [05:34] infinity: feisty doesn't exists on the mirrors yet? [05:38] doko: i dunno what to do with gcc... i have no idea what to look for and so on.. [05:38] It won't until something is published in it, I susect. [05:38] should we go for 4.2 branch for all arches? [05:38] Woo, the queue tool still defaults to edgy. [05:38] \o/ [05:38] will new gcc build hppa kernels yet? [05:39] BenC: yes. 4.1 should do [05:39] we don't know yet if they will boot, but to start with, it's enough it builds :) [05:39] Although, hrm, I'd expect something to be published by now... [05:39] Maybe the publisher hasn't been run since the import. [05:39] Feh. [05:39] No Team Soyuz around to confirm which step we're at. [05:40] I may have to wait until malcc wakes up before I start this dance. [05:40] infinity: don't sweat it... glibc and gcc are not ready yet [05:40] Kay. I'll get the kernel building as soon as I have malcc's okay, at least. [05:40] infinity: yeps... [05:40] hppa will have to wait, of course, cause it'll need manual love, but the other 5 should work, right? [05:40] not yet === ..[topic/#ubuntu-toolchain:fabbione] : STATUS: glibc: ppc busted | gcc: sparc busted [05:41] BenC: On which arches will the kernel actually build? :) [05:41] should build on all [05:41] infinity: all of them.. i did test build on all 6 a couple of days ago [05:41] Even hppa/dapper? [05:41] hppa will need edgy kernel-package and kernel-wedge otherwise it should be good [05:41] (I suspect that will require love) [05:41] Ahh, that's it? Cool. [05:41] yes [05:42] I can manage that. [05:42] the resulting kernel will NOT boot [05:42] but it will build [05:42] Booting is not critcal, getting linux-libc-dev is. [05:42] BenC: i assume you did pull also my changes to debian/d-i/ for ia64 and hppa, right? [05:42] i didn't notice the commit logs after that depatch-repatch [05:43] fabbione: yep [05:43] that should do [05:43] if nothing drastic changed in the meantime [05:51] Looks like I got the patch to build with the 2.6.19 headers right enough this time. [05:54] "right enough" [05:55] infinity: As in, the build didn't fail. [05:55] Not that I've reduced the patch to the set that I would expect to get past drepper. [06:00] doko: should we go for 4.2 branch for all arches? [06:01] Eh, no. [06:01] Having the merge be done with 4.2 in it's current state would almost certainly mean a rebuild after something crazy was discovered. [06:01] They only branched what, a week ago? maybe two? [06:01] fun :) [06:02] If we were sane and kept the results of the rebuild at the end for publishing, I'd say go for it. [06:02] But I'd also track glibc-2.6 actively through the release. [06:03] jb-home: that's MV spec (glibc-2.6 tracking).. we could start packaging it immediatly together with gcc-4.2 and use them offbw for some crazy rebuild.. we would have feisty+1 toolchain ready at the same moment it opens [06:06] fabbione: By spec for the last two release, we're supposed to have had places to upload toolchains and the ability to do test rebuilds. [06:06] Hard to get enthusiastic for that a third go 'round. [06:07] jb-home: well let's put this way ... with my addition to the toolchain slackers, at least i have enough CPU power at home to do main once in a while for all arches [06:07] fabbione: Ah, are you officially added to the crew now? =) [06:08] jb-home: didn't you add me to the team? [06:08] Yeah. You just hadn't acknowledged it before. =) [06:08] yeah yeah... whatever you say.. i am committing to glibc for fun :) [06:14] fabbione: Do you remember on which test she zombie'd out? [06:14] ppc ? [06:15] Yeah. [06:15] no, and i just started a build on davis disabling the test suite [06:15] you are late by 30 secs or so [06:15] Late how? [06:15] No build logs? [06:16] late with me doing ./debian/rules clean && rm ../*.build [06:16] Ah. [06:16] I'm just curious if I've passed that point already or not. [06:16] don't worry.. i can rebuild running the test suite and claim that i thought i was using my machine [06:16] hostnames are very similar :) [06:17] davis versus datbloodppcbox? [06:17] well get to the debs .. there might be an error in sysdeps/powerpc.mk anyway for the headers install in ppc64 [06:17] bloody. even. [06:17] davis versus daltanius [06:17] you know.. tab completion shit [06:17] i blame bash and its maintainer ;) [06:18] it means that in case i will never get root on ANY of the DC machines, but i don't think that will ever happen anyway [06:18] Eh, why don't I have any swap on my box... [06:18] Oh I see the problem you were describing. [06:18] So it doesn't actually stop the build at all. [06:19] are you getting Zombies? [06:19] Zl to be exact [06:19] 23241 jbailey 21 0 0 0 0 Z 95 0.0 3:17.46 ld.so.1 [06:19] 23103 jbailey 16 0 0 0 0 Z 13 0.0 0:40.09 ld.so.1 [06:19] 23159 jbailey 16 0 0 0 0 Z 13 0.0 0:32.74 ld.so.1 [06:19] 23196 jbailey 16 0 0 0 0 Z 13 0.0 0:28.47 ld.so.1 [06:19] 23120 jbailey 16 0 0 0 0 Z 13 0.0 0:36.73 ld.so.1 [06:19] 23139 jbailey 16 0 0 0 0 Z 13 0.0 0:33.48 ld.so.1 [06:19] 23176 jbailey 16 0 0 0 0 Z 13 0.0 0:29.87 ld.so.1 [06:19] 23216 jbailey 21 0 0 0 0 Z 12 0.0 0:29.16 ld.so.1 [06:19] yeps [06:19] i am still not getting them on ppc32 kernel [06:20] well at a certain point the testsuite will hang [06:20] it won't take long from there [06:20] Swap: 19531232k total, 0k used, 19531232k free, 1398728k cached [06:20] Much better. [06:20] * tim (n=tim@carl-sgc-sg-1.inter-touch.net) has joined #canonical [06:20] who is he? [06:20] ops.. ECHAN [06:21] fabbione: You're one of thoses detectives that work for HP, aren't you? [06:21] ehhehe [06:22] So hmm. Are those zombied because of a bug in glibc, the kernel, or upstart? [06:23] those are spawned by glibc build so i can't think of anything upstart related. [06:24] i am trying to exclude the kernel, building on ppc32 [06:24] They're parented to 1 [06:24] So wouldn't that mean they might be hanging around for not having SIG..CHLD? acknowledged by upstart? [06:24] Especially since they appear to be spinning? [06:24] Or using up CPU time somehow? [06:24] hmm [06:25] no i exclude upstart.. davis is running dapper [06:25] and i could reproduce it there [06:25] and it's running 2.6.15.x [06:25] compared to .17 on your box perhaps? [06:26] Yeah, I'm current edgy from 2 days ago or so. [06:27] right.. so am i [06:27] kill -9 won't take it out, so it's not a userspace problem. [06:28] as i told you yesterday you need a hard reboot of the box [06:29] brb [06:29] Bedtime soon for me. [06:30] I'll try asking Steve Munroe from IBM tomorrow. [06:32] fabbione: It made it through the nptl tests, which is where I would've expected permanent problems if nothing else. [06:33] jb-home: since it doesn't take too long to build there, could you please get it to the deb before you head to bed? [06:33] fabbione: It's working it's way through now. [06:33] ok [06:34] i am running 2 builds here: full on ppc32 and without test suite on ppc64 [06:34] But with a zombie taking up 98% of one CPU, and a collectiong of others fihting for the second one, it's not going quickly. [06:34] so what i would suggest is: [06:34] In the meantime I'll commit the build fix. [06:34] if ppc32 builds fine without zombies and ppc64 goes trough, we look at the tests results from ppc32 and upload temporary disabling the test suite [06:34] (assuming the results are good enough) [06:35] Errr... [06:35] I'd *really* rather talk to Steve first. [06:35] well that can be parallelized :) [06:35] it's not going to happen before monday anyway [06:35] What part of "first" can be parallelized? =) [06:36] our test build can parallelize with you talking with Steve :) [06:36] Lovely, I've commited that. Builds will now demand newer linux-libc-devs than edgy has. [06:37] And it's 00h36, time for sleep. [06:37] See y'all. [06:37] night jeff [07:06] jb-home: it hangs on ppc32 too [07:06] F U C K [07:07] davis did build without testsuite but clearly that's NOT good [07:58] jb-home, infinity: mailed all the ppc hackers for glibc issue.. let see what happens [09:50] BenC: latest git is FTBFS on ppc [09:51] oh nevermind [09:51] it's Olaf patch [09:58] jb-home: portforwarded the patches from glibc-2.4 [09:58] so we should be good with that [10:01] doko: edgy-{proposed,updates} are open as of edgy release; follow StableReleaseUpdates as usual [10:02] (open and functional, even) [10:04] infinity: publisher should've been run since i-f-p happened ... [10:04] cjwatson: Should've, as in "you think it has", or as in "why hasn't it"? [10:04] cjwatson: I've been waiting for malcc to show up to give me a status report. [10:05] $ ls ubuntu/dists/feisty/ [10:05] Release Release.gpg main multiverse restricted universe [10:05] can't see what else would've put that there. :-) [10:05] Yeah, fair point. :) [10:05] (I thought someone said that wasn't there earlier..) [10:05] Well, then, I guess we can NEW the kernel, and I can work on getting it building. [10:05] at this point I'm happy to turn the cron job back on [10:05] Also, the queue tool still defaults to edgy. Fun. [10:06] now done [10:06] yeah, I noticed that [10:06] I wonder if that's because feisty is EXPERIMENTAL not DEVELOPMENT [10:06] Could be. Why is it, anyway? [10:06] I'll happily change it. [10:07] I thought EXPERIMENTAL was meant to only exist if there was *also* a DEVELOPMENT release. [10:07] (And, afaik, we've never actually used it for anything) [10:07] right [10:07] i suggest you feisty in frozen [10:07] till we bootstrap the toolchain [10:07] I'm down with that idea. [10:07] that will work for queue too [10:07] infinity: it wasn't on the mirror a couple of hours ago [10:08] There, it's frozen now. [10:08] queue (and everything in launchpad that does Distribution.currentrelease) tries FROZEN, DEVELOPMENT, CURRENT in that order [10:08] and if there isn't one of those then it just picks the first one [10:08] cjwatson: Err, did you already NEW the kernel? [10:08] Oh, feh. The changes list is empty. [10:08] infinity: no === infinity fixes. [10:09] does feisty-changes exist yet? [10:09] it didn't last night [10:09] Yes. [10:09] wait until I subscribe? :-) [10:09] not that I'm a sad completist or anything [10:09] Be quick. :) [10:09] Okay, NOW it's frozen (and has a changes list) [10:11] cjwatson: Subscribed yet, huh, huh? :) === Keybuk [n=scott@quest.netsplit.com] has joined #ubuntu-toolchain [10:11] BenC: you seem to have gone back to native packaging rather than .diff.gz ...? [10:12] cjwatson: yes it's an -rc.. [10:12] no point to go .diff.gz [10:12] i guess he will switch with .19 final [10:12] infinity: yes. :-) [10:12] fabbione: ah [10:12] cjwatson: this upload is only to bootstrap [10:12] not for use [10:13] I know [10:14] Well, she's accepted. [10:14] Now I get to play. === infinity does a publisher run, real quick-like. [10:16] doesn't that need a toolchain? :p [10:16] Oh, I remember what I wanted to ask malcc (he says, 2 minutes after the publisher starts)... [10:17] cjwatson: Are we sure the dsync changes got made, so we don't keep altering edgy? [10:17] Keybuk: linux-libc-dev is the first bit of the chain that we want. [10:18] ahh [10:18] of course [10:18] I'd forgotten about that === infinity runs to 7-11 for nutrience while the publisher does... Stuff. [10:18] infinity: yes [10:18] malcc commented out link-dups; we have a better solution in testing [10:18] cjwatson: Oh, good. I didn't want to have to test how well the publisher deals with a SIGINT. [10:19] hah [10:19] we made sure of this before creating feisty. :) [10:21] Keybuk: carlos wants to start up feisty translations before we do the first big auto-sync, if possible [10:21] which requires downing launchpad for a few hours [10:21] cjwatson: we don't have glibc yet.. so that shouldn't be an issue [10:22] cjwatson: that's ok [10:22] jb-home and I think we will need a new kernel on ppc buildd to let glibc build [10:23] we also need to sync edgy -> feisty before I'll do the unstable -> feisty one ;P [10:23] and that needs a toolchain [10:23] edgy -> feisty? [10:23] isn't that done automatically when they "clone" edgy into feisty? [10:24] apparently not [10:24] it appears to have been done now though [10:24] it wasn't when I looked yesterday [10:24] qh ok [10:27] yeah, that was done early on [10:41] cjwatson: I'm down with the initial translation run being done anytime after this publisher run is done, since my kernel builds will likely be out-of-band anyway. [10:42] And, the publisher just finished. [11:19] *sigh* === Keybuk so can't remember how mom works [11:19] She just sits at home, does laundry, and watches Oprah while dad works, doesn't she? [11:19] At least, that sounds like my mom. [11:19] this is the first release cycle where she hasn't had a PMT attack, so I literally haven't touched her for sixth months! [11:41] we should let Hobbsee loose on the archive, if only to free up some disk space on casey :) [12:42] Keybuk: i knew about PMS attack.. PMT????? [12:43] fabbione: another word for PMS, I guess s/Stress/Tension/ [12:44] syndrome ... yeah gotcha [12:44] dict has a lot of interesting other definitions tho :) [01:41] infinity: are you the one accepting stuff? [01:42] infinity: note that debootstrap has some C programs in it [01:42] cjwatson: I am. [01:42] cjwatson: Don't mind so much, I'd just like to see it built and useable. [01:43] cjwatson: It'll get reuploaded at least once, I'm sure. :) [01:43] fair enough :) [01:43] fabbione: The build hang is *tee* hanging. [01:43] Whacked. [01:43] (Not accepting anything else, mind you) [01:44] infinity: just uploaded debhelper; you might want that [01:44] I'm skipping dpkg because (a) it's bloody hard to merge and iwj can do it and (b) it doesn't seem necessary [01:45] mind you, debhelper doesn't look like a huge deal either [01:45] fabbione: Oh, I guess that makes sense since the children haven't exitted. === cjwatson_ [n=cjwatson@82-69-40-219.dsl.in-addr.zen.co.uk] has joined #ubuntu-toolchain === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-toolchain === jb-home [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has left #ubuntu-toolchain ["Twisting] [02:20] fabbione: Still around? /proc/##/wchan says that the process is sitting in do_exit [02:21] Is there any way to tell where it is in that? [02:25] fabbione: gcc-4.1 built on 2.4 at davis:gcc/4.1/ [02:30] jbailey: twisting like a baby in a slow flame? you sick puppy [02:30] lamont: Eh? [02:30] Oh. It's a quote from a Cult song called "edie" [02:30] 27-10-2006 06:17:36 -!- jb-home!n=jbailey@modemcable139.249-203-24.mc.videotron.ca has left #ubuntu-toolchain ["Twisting like a flame in a slow dance, baby..."] [02:30] maybe I transposed a few words [02:31] Perhaps. =) [02:31] Although it's an incredibly depressing song. One of the ones I really enjoyed as a teenager. [02:34] jbailey: i am here now. i did mail all of upstream. [02:34] doko: it does?!?!?? [02:34] doko: glibc-2.5? please doublecheck [02:37] fabbione: I'm currently building in feisty-libc [02:37] doko: ok... [02:37] these Niagara failures are scary. i need to talk to david === lamont wonders if there's any signficance to neither of his pet architectures being in the topic [02:38] lamont: ia64 is ok already as i told you [02:39] fabbione: right. must wake up. [02:39] lamont: hppa we need to do some extra bootstrapping love but getting ready for it [02:39] and hppa has the signals, um, patch. plus bootstrapping === ..[topic/#ubuntu-toolchain:lamont] : STATUS: glibc: ppc busted | gcc: sparc busted | hppa: bootstrapping needed [02:39] doko: looks good glibc.. dunno what to say === ..[topic/#ubuntu-toolchain:fabbione] : STATUS: glibc: ppc busted | hppa: bootstrapping needed [02:40] so we only need ppc glibc [02:40] rock [02:40] jbailey: anyway i did build a set of debs for ppc [02:40] without running the test suite [02:40] fabbione: Any thoughts on my debugging questiosn? [02:41] no i have no idea. i did spoke with benh and he also believes it's a kenrel bug [02:41] and he said that they were going to look at it very quickly [02:41] Cool. === jbailey_ [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-toolchain [03:12]