* SpamapS installed originally from Xubuntu00:00
SpamapSI actually like Unity + Compiz when they work. A lot. Just tired of having to restart them because of some weird issue.00:00
SpamapShas been much better the last 2 weeks00:00
micahgSpamapS: there's always unity-2d00:57
SpamapSmicahg: indeed, it does seem to survive longer than unity between weirdness/crashing00:58
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
photonhi. I am scanning through archive.ubuntulinux.org, and I noticed that some .deb files are fairly small (just a few kb), like http://archive.ubuntulinux.org/ubuntu/pool/multiverse/u/ubuntu-restricted-extras/. why is that? shouldn't they be megabytes given how much software is in them?01:46
=== rextsai is now known as chihchun
gusnanphoton, those are probably pseudo-packages, not containing much information themself, but just dependencies on other packages with the actual stuff.01:52
TheMusoThat package in partiular is a metapackage, which pulls in other packages.01:53
photonI looked into the .deb archive and in /DEBIAN/control, there are only recommendations, not dependencies, and even those are far less than what apt-get install ubuntu-restricted-extras would install. so I'm confused01:53
photonmaybe I'm looking int he wrong file.01:54
TheMusoYes, recommends are installed by default, so the recommended packages get installed.01:55
TheMusoIt means that you can remove one or more of the recommended packages, and the rest of the packages don't get removed with them.01:55
photonGot it.01:56
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
pittiGood morning04:05
pittidoko_: vala FTBFS> looking04:06
didrocksgood morning05:38
=== tilgovi_ is now known as tilgovi
SpamapS@pilot out07:30
=== udevbot changed the topic of #ubuntu-devel to: Beta 1 released | Archive: Feature/UI Freeze | Development of Ubuntu (not support, not app development) | build failures -> http://bit.ly/or6CHJ | #ubuntu for support and general discussion for hardy -> oneiric | #ubuntu-app-devel for application development on Ubuntu | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots:
brodercould i get a core dev to do me a favor and accept the series nomination on bug #848687?07:32
ubottuLaunchpad bug 848687 in linux (Ubuntu) "DRI applications hang after DPMS standby on i965" [Undecided,New] https://launchpad.net/bugs/84868707:32
dholbachgood morning07:39
SpamapSbroder: done07:39
dholbachpitti, it seems like we need to bring libgdamm4.0 back if we want glom buildable at all07:41
pittidholbach: ah, glom isn't in Debian, so they could remove it07:41
pittiI guess someone can reupload it07:42
broderSpamapS: thanks :)07:42
dholbachpitti, ok - I'll have a look - seems like glom does not build - maybe a more recent version will fix it07:42
micahgdholbach: upstream commented as such, but one of the build depends was removed07:44
pittipresumably that very from above, libgdamm07:44
dholbachI think upstream has a PPA - I'll have a look how much of their work we can reuse07:44
dholbachI'll let you know07:44
micahgsorry, missed that somehow...07:44
dholbachpitti, maybe glom should be part of the please-update-these-desktop-packages overview, so we don't let old versions rot in the archive again :)07:45
pittinah, it's in universe07:45
pittia lot of universe packages are like that unfortunately :/07:45
micahgdholbach: shows up here where it belongs: http://qa.ubuntuwire.com/uehs/no_updated.html, unfortunately, we need more people looking here :)07:46
dholbachok :)07:46
* micahg actually thinks that page is out of date...07:51
* micahg moves this to -motu07:51
=== vila is now known as babune
=== babune is now known as vila
=== dholbach_ is now known as dholbach
=== doko_ is now known as doko
LaneyI have been trying to unblock glom in Debian for ages :(08:44
Laneycurrent problem is jsogo refusing to maintain goocanvas properly08:44
dholbachLaney, what's the issue?08:46
Laneythe current releases need a newer goocanvasmm which needs a newer goocanvas08:52
dholbachah ok, for now, I guess I just get 1.18.3 into Ubuntu which is the latest stable and builds08:54
dholbachfor P we can try something new08:54
dholbachis it hard to get a new goocanvas into Debian?08:54
Laneykind of is if the maintainer doesn't want to do it08:54
Laneywould the 1.18 release work?08:55
dholbachyes, it looks good08:56
dholbachI'm just sorting out a few other small issues while I'm at it08:56
dholbachwe just need to get gdamm4.0 added back to the archive08:57
dholbachbut I uploaded it already08:57
dholbachcan somebody get libgdamm4.0 out of NEW please? it was removed a bit too quickly (glom needs it)09:35
=== zyga is now known as zyga-afk
pittidholbach: done10:12
dholbachthanks pitti10:12
ogra_dholbach, do you still need an armel testbuild ?10:13
dholbachogra_, no it's all sorted10:13
dholbachthanks ogra_10:13
* doko can't decide if the archive looks like a garbage dump or a mortuary ...10:39
pittilooks like a big city; shiny center, but some suburbs are really chabby10:40
didrockspitti: I like this metaphor :)10:41
ajmitchpitti: does this mean that a bulldozer needs to be driven through some suburbs?10:43
pittiajmitch: in some cases a bulldozer, in some cases a citizen's initiative with some mortar and paint :)10:44
pittihildon or gnome 2 panel applets were definitively in the bulldozer class, but most packages have relatively easy fixes10:45
ajmitchit's sad to see the applets go :)10:45
dholbachcan somebody bulldoze libgdamm4.0 out of binary new? :-P11:06
dokofun ...11:07
dokoChecking if libc on this machine contains:11:07
dokogrep: library_contents: No such file or directory11:07
doko  vsprintf: No, I don't think11:07
dokogrep: library_contents: No such file or directory11:07
dokogrep: library_contents: No such file or directory11:07
doko  vfprintf: No, I don't think11:07
dokogrep: library_contents: No such file or directory11:07
dokogrep: library_contents: No such file or directory11:07
doko  strtoul: No, I don't think11:07
dokogrep: library_contents: No such file or directory11:07
doko  strtol: No, I don't think11:07
dokogrep: library_contents: No such file or directory11:07
doko  memcpy: No, I don't think11:07
cjwatsondoko: I'm guessing multiarch fallout11:08
dokosure, I just did "like" the grep approach11:08
cjwatsondholbach: one moment11:09
dholbachcjwatson, thanks muchly11:09
=== zyga-afk is now known as zyga
dokocjwatson, ahh, already done11:11
cjwatsonah, ok, cool11:12
dholbachfantastic, thanks11:13
=== ryanakca_ is now known as ryanakca
dholbachnice, we're down from 661 to 560 FTBFS bugs today11:44
nigelbmicahg: Congrats \o/11:48
=== MacSlow is now known as MacSlow|lunch
ScottKSweetshark: What are the odds of the LO FTBFS on i386 getting fixed today?  FYI, it's breaking amd64 live CD/DVD builds at the moment.12:05
cjwatsonmvo: I could use help with bug 848907; I don't see why it should have broken recently12:06
ubottuLaunchpad bug 848907 in apt (Ubuntu) "apt crashes on initial installation of oneiric" [Undecided,New] https://launchpad.net/bugs/84890712:06
SweetsharkScottK: Im currently building with a fix. but LO has quite some buildtime ...12:09
dokosiretart, ping12:11
SweetsharkScottK: It just went past the critical corner, so its looking good.12:11
mvocjwatson: is that reproducable?12:12
cjwatsonmvo: I reproduced it after seeing a user report of the same thing, which he's experienced both yesterday and today12:13
cjwatsonmvo: haven't yet figured out how to reproduce it from the command line12:13
cjwatsonregfree crash suggests memory corruption of some kind, but getting valgrind into d-i is ... painful12:14
mvocjwatson: right, but its reproducable by installing kubuntu in a vm with the altnerative CD?12:14
cjwatsonthat's all I did, yes12:15
mvothanks, I will rsync me updated images and try it out then12:16
cjwatsonI would expect Ubuntu to do the same thing; we haven't got to the flavour-specific parts of installation at this point12:16
cjwatsonI just made a 10G disk image and did a use-entire-disk install12:16
cjwatsonen_GB locale, otherwise all defaults12:16
cjwatsonwill try it under strace now12:17
mvocjwatson: great, the rsync is running. you did the install inside kvm I assume? is there a chance to get a slightly better readable backtrace then from the screenshot ?12:17
cjwatsonkvm yes; maybe, will try hacking about at base-installer12:17
mvocjwatson: I can poke the code in the meantime, there were some changes in the cdrom autodetect code recently (~4 weeks ago) that might be the issue12:17
cjwatsonmvo: I think this is recent though, I didn't see it when I was preparing the partman-crypto fix this guy was testing12:19
cjwatsonmvo: I wonder if it's something to do with the recent eglibc upload (and hope it's not)12:19
cjwatsonI think the backtrace must be being printked or something; I guess I'll just leave it on a sacrificial console12:21
dokoLaney, did the comment should go to bug 831235?12:27
ubottuLaunchpad bug 831235 in gdcm (Ubuntu Oneiric) "gdcm version 2.0.17-3.1 failed to build in oneiric" [High,Confirmed] https://launchpad.net/bugs/83123512:27
SweetsharkScottK: the fix is uploaded to chinstrap and waiting for sponsoring.12:28
Laneydoko: as in, the Debian version builds fine12:28
dokoLaney, yes transient build failure, you shouldn't upgrade gcj during a test rebuild ;)12:29
Laneythe current one works?12:29
cjwatsonmvo: reproduced, but the backtrace goes all over tty1 even if the current console is tty3 :-(12:29
Laneythere were a lot of probems that we helped the maintainer with in Debian12:30
* cjwatson extracts the strace12:30
dokostill running12:30
Laneyarch/indep issues12:30
ejatis this bugs 847591 still in progress ? or already fix but not release12:31
ubottuLaunchpad bug 847591 in packagekit (Ubuntu Oneiric) "Error org.freedesktop.DBus.Error.Spawn.ChildExited: Launch helper exited with unknown return code 127" [High,In progress] https://launchpad.net/bugs/84759112:31
ScottKSweetshark: Thanks.12:34
=== zyga is now known as zyga-afk
cjwatsonmvo: http://people.canonical.com/~cjwatson/tmp/base-installer.trace.xz - probably includes lots of junk, I'm about to start having a look12:35
cjwatsonthat's attaching strace to bootstrap-base.postinst after run-debootstrap has been forked12:36
cjwatsonI suspect it isn't helping that locale-gen removes /usr/lib/locale/C.UTF-8 which is owned by the libc-bin package12:40
* cjwatson goes to fix that12:40
SweetsharkScottK: uploaded (thanks pitti)12:49
Laneydoko: you are right, it is transient. I can't get on https://buildd.debian.org/status/logs.php?pkg=gdcm&ver=2.0.17-3.1 to remind myself what the problem was.12:54
Laneyget logs from*12:54
dokoLaney, I had given it back: https://launchpad.net/ubuntu/+archive/test-rebuild-20110816/+build/269375712:55
LaneyI just knew 3.1 failed and -4 did not, but, well it's not reproducible so "oh well".12:55
Laneyoh, the red text gives you the log12:56
Laneythat does not look like a link.12:56
=== zyga-afk is now known as zyga
=== MacSlow|lunch is now known as MacSlow
cjwatsonmvo: aha, reinstall libc-bin in /target after the crash and then apt segfaults again13:37
cjwatsonmvo: better screenshot attached13:38
cjwatsonso almost certainly due to:13:39
cjwatson  * debian/patches/localedata/locale-C.diff: Don't include ISO1465113:39
cjwatson    collation rules in C.UTF-8 locale.13:39
cjwatsonbut I don't know if this represents an eglibc bug or a wrong assumption in apt13:39
* cjwatson installs gdb and valgrind to try to dig further13:41
stgrabermicahg: can you confirm you are now on the DMB mailing-list?13:42
mvonice cjwatson! that is really much more helpful13:44
cjwatsonmvo: ... and now valgrind output attached13:49
mvocjwatson: thanks, I wonder if the compilation error is new or maybe triggering the bug13:52
achiangpitti: good morning, thanks for sponsoring some of my patches. i see that you created bzr branches... should i have done that on my own to save time?13:53
pittiachiang: I did?13:54
cjwatsonmvo: it doesn't happen in LC_ALL=C13:54
achiangpitti: well, someone created bzr branches. :) see LP: #770862 as an example13:54
cjwatsonnow, I wonder if I can recreate it in an unstable chroot13:55
mvocjwatson: *ick* it doesn't?13:55
cjwatsonmvo: indeed; that's why I was suspecting that localedata change13:56
cjwatsonif I can recreate it in unstable, I'll ask debian-glibc about it13:56
cjwatsonthere's no rationale for the change13:56
pittiachiang: presumably the auto-importer13:57
achiangpitti: ah, interesting. ok, thanks, i won't worry about it then13:57
dholbachdoes the publisher run every now and then nowadays? seems like libgdamm4.0 still did not make it through binary new13:57
cjwatsonit's crashing13:58
dholbachah ok :/13:58
cjwatsonI'll escalate13:58
=== zyga is now known as zyga-afk
dholbachthanks a lot cjwatson13:58
mvocjwatson: thanks, what channel? I would like to follow this - fwiw I think there is  a potential double free (or delte in this case) in the code that it worth fixing14:00
cjwatsonmvo: the mailing list14:01
hallynWTF?  alternate installer from sunday doesn't have a kernel it can install?14:02
jdstrandpitti: hi! it just occurred to me that we don't have an equivalent of gdm-guest-session for lightdm14:03
ScottKhallyn: Look at the newer one.14:04
pittijdstrand: no, it's built into lightdm14:04
jdstrandpitti: oh? are there docs on that?14:04
cjwatsonhallyn: see my discussion with mvo immediately above your question14:04
hallynScottK: ok, thanks14:04
=== chuck_ is now known as zul
pittijdstrand: docs for what? it's in the indicator and also in the greeter14:05
cjwatsonScottK: actually that wasn't really it14:05
cjwatsonScottK: that broke something slightly different ...14:05
jdstrandpitti: sorry, I was not at all clear. I meant the apparmor profile14:06
cjwatsonScottK: the weekend problem was wrong udebs on the CD, but this is apt falling over when trying to install the kernel on the target system14:06
cjwatsonor actually when trying to detect what to do14:07
mvocjwatson: I have some code that may fix the issue, it certainly fixes the double free but I'm not sure if that is enough. it seems to be only triggered (the double free) if there is a error in the regexp compilation14:07
pittijdstrand: hm, it ought to, but now that you mention it I can't see it14:07
jdstrandpitti: shall I file a bug?14:08
pittijdstrand: please14:08
pittijdstrand: I discussed the four bugs you mentioned on Friday with Robert this morning, FYI14:08
pittijdstrand: one is fixed, one committed, I just sent an MP for the third14:08
jdstrandpitti: I saw. thanks! I have updated our lists accordingly14:09
hallyncjwatson: (sorry, i'm not sure which you're pointing to, but it sounds covered, so great :)14:10
m4n1shprovided with say -lzeitgeist-1.0 can I find out to which this shared library it will link to?14:11
m4n1shmeans the actual file path14:11
m4n1shor even the name would be enough14:11
jdstrandpitti: fyi, bug #84902714:13
ubottuLaunchpad bug 849027 in lightdm (Ubuntu Oneiric) "lightdm does not provide an equivalent to the gdm guest session AppArmor profile" [Undecided,New] https://launchpad.net/bugs/84902714:13
pittijdstrand: thanks, bumped accordingly14:17
jdstrandcool, thanks14:18
mvocjwatson: if I upload a new version to my PPA will you be able to easy test if my fix is sufficient (assuming the double delete is actually the real problem here)14:24
=== kentb is now known as kentb-afk
cjwatsonmvo: I think so14:26
dokonjpatel, re bug 817691, do you really mean glibc, not glib2.0?14:26
ubottuLaunchpad bug 817691 in glibc (Ubuntu) "[Oneric] unity-panel-service crashed with SIGSEGV in getenv()" [Critical,Confirmed] https://launchpad.net/bugs/81769114:26
bdmurraysuperm1: Would you please look at my comments in bug 816199 and reconsider the fix?14:31
ubottuLaunchpad bug 816199 in dkms (Ubuntu) "dkms_packages.py crashed with ValueError in _apt_pkg(): package does not exist" [High,In progress] https://launchpad.net/bugs/81619914:31
slangasekScottK: honeyd> ack; would you like me to take this one?14:32
ScottKslangasek: Yes.  Please.14:32
slangasekok, will do14:32
mvocjwatson: I uploaded it to https://launchpad.net/~mvo/+archive/ppa/+packages?field.name_filter=&field.status_filter=published&field.series_filter=oneiric14:34
mvocjwatson: its building currently14:34
njpateldoko, yeah, going by what we found on Google and the other bugs in firefox etc with the same trace14:34
dokonjpatel, I'd rather assume some kind of memory corruption elsewhere. really doesn't look like eglibc's fault14:35
njpateldoko, it might well be, maybe gettext?14:37
dokonjpatel, then why assign it to eglibc, both ubuntu and upstream?14:38
njpateldoko, because the crash is in getenv and we needed some opinions from people familiar with that? I can guess it's in gettext but it doesn't make it so...14:40
njpateldoko, if you have a better place let me know and I'll change it asap14:41
njpateldoko, did you look at the linked bug list for the other projects?14:41
dokonjpatel, there is this one new report in unity, the other list are up to years old14:43
njpatelurgh, really? what the hell was I reading then :/14:43
* njpatel looks14:43
dokonjpatel, and the fact that every of these apps is using gtk, is not very assuring14:44
njpatelthe top 8 reports are from 2011, most of them after june/july14:45
njpatel doko it does seem gtk is involved in most of them14:46
dokonjpatel, e.g. bug 278095 is about corruption of the env14:47
ubottuLaunchpad bug 278095 in at-spi "MASTER crash in getenv() ... spi_atk_bridge_exit_func()" [High,New] https://launchpad.net/bugs/27809514:47
njpateldoko, yep, saw that, so unity-panel-service, gnome-keyring from O and evince and gnome-session from N14:49
njpateldoko, thanks, I'll move it back to unity for now and ask upstream gtk if they have any ideas14:50
dokopitti, didrocks, mterry, Sweetshark: any idea of the relevance of bug 725250?14:54
ubottuLaunchpad bug 725250 in lo-menubar (Ubuntu) "[MIR] lo-menubar" [Undecided,Confirmed] https://launchpad.net/bugs/72525014:54
slangasekcan someone tell me why the notification area seems to have completely disappeared from my panel due to an update sometime in the past couple of weeks?14:55
=== kentb-afk is now known as kentb
pittidoko: no, it's too broken for that14:56
hallynScottK: fwiw, today's alternate installer is no better :(14:56
mterrydoko, yeah, I don't think it's prime time yet14:56
pittidoko: bug updated14:57
cjwatsonhallyn: yeah, on it :-)14:58
tjaaltonirc admins available? need the bot back on #ubuntu-x15:00
tjaaltonbugbot that is15:00
bdmurrayslangasek: bug 553745 has some mention of it occuring again in Oneiric - subsequently I've removed the bugpattern for it so we can get another crash report in15:01
ubottuLaunchpad bug 553745 in plymouth (Ubuntu Maverick) "plymouthd crashed with SIGSEGV in ply_event_loop_process_pending_events()" [High,Triaged] https://launchpad.net/bugs/55374515:01
superm1sure bdmurray15:01
superm1someone with buildd knowledge, is it actually possible to call apt-get source  to fetch a source package from ftpmaster.internal on a buildd during a build?15:01
slangasekbdmurray: right, was talking to Jason about that - thanks, will be happy when we can get more info15:02
hallyncjwatson: great, thanks :)15:07
bdmurraypitti: Could you review / merge https://code.launchpad.net/~brian-murray/apport/ubiquity-dupe-sig-improvements/+merge/75050 ?15:10
pittibdmurray: merged, thanks15:12
pittiRiddell: "Give david.wonderly access to upstream docs" is the only remaining WI for desktop-o-kubuntu-documentation-review; is that blocked, or just needs to happen?15:17
Riddellpitti: I've no idea, I'm not working on Kubuntu15:17
Riddellpitti: but David is DarkwingDuck on #kubuntu-devel15:18
pittiRiddell: right, but that sounded like an ACL issue15:18
pittilike adding him to a team or so15:18
pittiRiddell: I'm happy to postpone that one if you want, just reviewing leftover WIs15:18
RiddellI have no opinion, you should ask the kubuntu team15:19
mvocjwatson: any luck with the ppa version?15:34
chrisccoulsondoko, hmmm, bug 831256 is a bit weird15:35
ubottuLaunchpad bug 831256 in dehydra (Ubuntu Oneiric) "dehydra version 0.9.hg20110609-1ubuntu1 failed to build in oneiric" [High,Confirmed] https://launchpad.net/bugs/83125615:35
chrisccoulsoni don't see how that fails :/15:35
cjwatsonmvo: not yet, I was chasing things down in eglibc: see thread at http://lists.debian.org/debian-glibc/2011/09/msg00036.html15:41
mvocjwatson: aha, nice!15:48
cjwatsonbut it's true that segfaulting is a bad response to regcomp failing15:52
cjwatsonjust a moment, testing15:52
cjwatsonmvo: right.  that does indeed fix the segfault, thanks, and valgrind is now happy.  of course we still need to fix the eglibc bug since now the regexes are incorrectly not recognised, and I suspect this will break other things15:54
cjwatsonhopefully aurel32 will reply quickly15:54
=== beuno is now known as beuno-lunch
ahasenackhi, could someone please nominate https://bugs.launchpad.net/landscape-client/+bug/813477 for maverick and natty?15:54
ubottuLaunchpad bug 813477 in Landscape Client "Update landscape-client to" [High,Fix committed]15:54
cjwatsonahasenack: done15:55
LaneyI would if the ... oh, never mind15:55
mvocjwatson: yeah, much agreed, thanks a bunch for testing, I will upload the fix next15:56
ahasenackcjwatson: thanks15:58
=== paulliu is now known as paulliu_gd
=== zyga-afk is now known as zyga
tjaaltonbarry: ping re libfsobasics, libfsotransport; I believe these could be synced?16:08
tjaaltonseems that libfsobasics is blacklisted currently, is there a way for me to see why?16:14
slangasektjaalton: the blacklist is unfortunately not synced to anything public; let me look for ou16:14
tjaaltonslangasek: thanks16:14
tjaaltoni know it ftbfs16:15
slangasektjaalton: in fact, I don't see it in the blacklist - how did you determine that it's blacklisted?16:15
tjaaltonslangasek: I ran syncpackage16:15
tjaaltonLaney: ooh16:15
LaneyLaunchpad autoblacklists when the Ubuntu version deviates AFAIK16:15
slangasekright, so you have to override16:15
dokochrisccoulson, given it back, and it did build. thanks16:15
cjwatsonslangasek: http://people.canonical.com/~ubuntu-archive/sync-blacklist.txt FYI16:16
slangasekcjwatson: oh, ok16:16
slangasekcjwatson: where's the cronjob that syncs that?  I have trouble figuring out how things get to that directory16:16
cjwatsonslangasek: cocoplum:~lp_archive/dak/cron.sync16:16
cjwatsonbut yeah, in this case it's more confusing UI from syncpackage16:17
slangasekgot it16:17
tjaaltonso, are the ffe-rules relaxed when fixing ftbfs bugs? :)16:18
slangasekexceptions may be more likely to be granted, but there's still a need for due diligence16:18
infinity(Unless you mean you've JUST fixed the FTBFS bug, which isn't a feature, and not subject to FF)16:19
infinityBut yeah, "I've added 7 new features, and accidentally fixed a build failure at the same time" doesn't fly. :P16:19
tjaaltonin these cases the packaging has been fixed too, but all our changes could be dropped as well16:19
tjaaltoni'll file bugs then16:19
bambeeHi, plymouth is not shown at all on startup, except if I remove "vt.handoff=7" manually (also with CTRL+ALT+F1 during the boot process, it works). it works fine on reboot/shutdown. Any ideas ?16:20
bambeethe system does not switch back to vt 1... it's strange :\16:20
slangasekhow do you mean, "does not switch back to vt1"?16:21
SpamapSslangasek: so, the "wait for interfaces" change has ruffled some feathers. It seems that d-i always leaves behind a permanent 'auto ethX' after installation.. which means a lot of people are upgrading and then seeing a 2 minute delay16:26
bambeeslangasek: I just wonder why I don't see anything (nouveau and kms work)16:28
bambeeif I type "CTRL+ALT+F1" during the boot process, I see plymouth, otherwise I see nothing16:29
=== deryck is now known as deryck[lunch]
slangasekSpamapS: wait for interfaces> yep, caught that the other day in the mail trail.  Do you think it's feasible to clean that up on upgrade?  Or maybe we should exclude interfaces that are configured as merely auto/dhcp in favor of checking for NM running?16:37
slangasekbambee: you see plymouth during the boot process only when typing Ctrl+Alt+F1, *and* vt.handoff=7 is being passed?16:37
cjwatsonI still think NM is a red herring; if NM was installed, then 'auto' lines would've been removed at the end of installation16:38
ScottKslangasek: If it's relevant, bambee is on Kubuntu and it's not very good about transitions being invisible.16:38
slangasekScottK: well, so far we're discussing flavor-neutral bits of the boot16:39
SpamapScjwatson: ok so there's already some intelligence there in NM's postinst?16:39
ScottKJust mentioning it because he may be seeing stuff that's normally not visible in Ubuntu.16:39
cjwatsonSpamapS: I gave specifics in the bug16:39
SpamapSoh I haven't caught up yet16:39
* SpamapS reads16:40
bambeeslangasek: exactly16:40
slangasekcjwatson: do you have any suggestions on where to look next with bambee's handoff issue?  Sounds like something falling down in the junction between grub, kernel and plymouth16:40
bambeean interesting info: it works just fine with {kubuntu,ubuntu}-11.10-desktop-amd64 (the livecd), and it does not work on my system16:42
cjwatsonmaybe plymouth:debug=file:/run/plymouth.debug (assuming oneiric)16:42
slangasekbambee: could you try booting with that option, plus vt.handoff=7?16:43
cjwatsonplymouth is supposed to display on vt7, not vt116:43
bambeeslangasek: sure16:43
SpamapScjwatson: I agree 100% with the position that a change in behavior at this change would be more harmful than good. I wonder if we can notify users in plymouth that the system is waiting for network configuration so, if nothing else, we know how to tell them to fix it when they report a slow boot.16:43
cjwatson(because otherwise the smooth transition from grub to plymouth wouldn't work)16:43
cjwatsonSpamapS: that sounds straightforward and a good idea, yes16:43
slangasekbambee: two major differences between the live CD and an installed system are the boot loader used, and the fact that plymouth is always started in the initramfs for a liveCD16:43
SpamapSthat should be easy to do in failsafe.conf16:44
slangasekbambee: you could eliminate one possible source of this noise by installing the cryptsetup package on your installed system, which will cause plymouth to be started from initramfs on the installed system16:44
bambeeslangasek: trying with both (option+package), bbl16:46
bambeeslangasek: well, nothing is generated in /run   (/run/plymouth.debug does not exist)16:52
bambee/proc/cmdline: BOOT_IMAGE=/vmlinuz-3.0.0-11-generic root=UUID=5b2e8c3c-eacf-494b-bb21-2363d0d235f2 ro plymouth:debug=file:/run/plymouth.debug quiet splash vt.handoff=716:52
slangasekbambee: blast; that's somewhat consistent with my own experiences trying to get output from plymouth in oneiric, but I was hoping cjwatson had the magic recipe (/run vs. /var/log)16:53
slangasekbambee: so I'll take some time to hack on that today and see if I can figure out what's going on with logging.  In the meantime, can you try installing cryptsetup?16:54
slangasek(and rebooting, of course)16:54
bambeeslangasek: sure, trying16:54
cjwatsonmaybe try /run/initramfs/plymouth.debug instead16:55
cjwatsonhistorically I used /dev/.initramfs/plymouth.debug, actually, but nowadays somewhere under /run ought to work16:55
cjwatsonI haven't debugged plymouth since the /run transition16:55
slangasekI thought the log only got written to disk when /etc/init/plymouth-log.conf fired16:55
SpamapScjwatson: one last question.. does d-i call ifblacklist_migrate.sh ? I only see it called in n-m's postinst configure if we're coming from before 0.6.5-0ubuntu12 ..16:56
bambeecjwatson: trying too16:57
* SpamapS can probably c/o d-i and confirm that if you don't know off the top of your head ;)16:57
dobeyhrmm, no patch pilot atm?16:57
cjwatsonSpamapS: yes, it does, that was why I mentioned it16:57
cjwatsonSpamapS: netcfg16:57
cjwatson./finish-install.d/55netcfg-network-manager:11:         sh /usr/lib/NetworkManager/ifblacklist_migrate.sh16:57
SpamapScjwatson: ahh ok, thanks. :)16:58
cjwatsonslangasek: true16:58
cjwatsonslangasek: but that should happen as soon as the rootfs is writable16:59
cjwatsonslangasek: actually, I don't think that applies to the debug log16:59
cjwatsonplymouth-log is about /var/log/boot.log16:59
slangasekcjwatson: ok - I thought it did, I'll have a look at the code a bit later to confirm.  Anyway, it doesn't seem to work for me at all, possibly because the job races plymouth-stop16:59
cjwatsonwhich interestingly appears to date from 2011-07-13 on mmy system, hmm17:00
slangasekmaybe we should make 'plymouth quit' do a last-ditch attempt to write the logs17:00
dobeyshould maybe ubuntu-sponsors be subscribed to all merge proposals into lp:~ubuntu-branches owned branches?17:01
SpamapSdobey: effectively sponsors are, because they show up in the sponsorship queue17:01
ahasenackSpamapS: hi, finally finished the other ubuntu releases for #813477, can you upload the missing Maverick and Natty packages?17:02
dobeySpamapS: do they get spammed by LP for them though? or do they just show up on the queue page? :)17:02
SpamapSdobey: the queue. I don't think I'd want to read all of ubuntu-sponsors' bugmail ;)17:03
SpamapSahasenack: ACK, will take a look later today, in the middle of a few things.17:04
ahasenackSpamapS: ok, thanks17:04
bambeeslangasek: even with cryptsetup installed, it does not work17:04
dobeySpamapS: well this isn't bug mail, it's code review mail. :)17:05
slangasekbambee: thanks, that eliminates one source of difference.  The remaining major difference is the bootloader itself, which again points to the grub+kernel vt handoff17:05
bambeecjwatson: plymouth produces no debug outputs here :\ (even with /var/log)17:05
dobeySpamapS: hopefully it's not as bad as bugs mail17:05
SpamapSdobey: the queue is, IMO, a much better way to approach it, as things are sorted by their age.. so right away you know which thing has been sitting in the queue the longest17:06
SpamapSemail is an interrupt driven system and favors the squeaky wheels17:06
bambeeslangasek: oh! I still have a 2.6.38 kernel here (just in case)17:07
bambeeI could test with the 2.6.38....17:07
cjwatsonbambee: "even with /var/log"> doesn't make sense17:07
dholbachcan somebody please get the glom packages out of binary new?17:08
slangasekbambee: sure, that may also be a useful data point - even in the event that it works with the older kernel, this doesn't necessarily point to a kernel bug, but it at least gives us a place to look (i.e., git-bisect)17:09
dobeySpamapS: i take it you mean https://launchpad.net/ubuntu/oneiric/+queue ?17:11
cjwatsondobey: http://reports.qa.ubuntu.com/reports/sponsoring/17:12
dobeycjwatson: ah, ok17:13
bambeeslangasek: it does not work with linux-image-2.6.3817:17
bambeeI don't think it's kernel related... (it worked fine on natty with linux-image-2.6.38)17:18
=== beuno-lunch is now known as beuno
bambeecjwatson: http://paste.ubuntu.com/688477/17:18
cjwatsondinner trumps plymouth debugging :-)17:19
=== deryck[lunch] is now known as deryck
bambeebtw, "[ply-utils.c]                               ply_open_module:Could not load module "/lib/plymouth/renderers/x11.so": /lib/plymouth/renderers/x11.so: cannot open shared object file: No such file or directory"17:41
Laneybts --mbox show17:43
Laneyarg, sorry17:44
=== eyes_ is now known as EyesIsServer
cjwatsonbambee: the X11 plugin isn't intended to be used there18:13
cjwatsonbambee: so that part is a red herring18:14
bambeeoh ok18:14
cjwatsonbambee: hmm.  there really isn't anything obvious (to me) there.  it looks as though plymouth's internal logic is doing all the right things, but it's just talking to nouveau wrongly :-(18:14
bambeeso , is it a problem with nouveau ? don't compute it worked just fine on natty with linux-image-2.6.38... and the kernel is still the same...18:16
cjwatsonI wouldn't expect so; presumably it's dealing with X just fine18:16
cjwatsonplymouth has a variety of horrible code to talk to each of the different framebuffer implementations in different ways, since at the time we first put that all together they all had different requirements18:17
cjwatsonit seems quite plausible that nouveau's requirements have changed and our plymouth package hasn't kept up18:17
dobeycan i bug someone to PLEASE sponsor a critical fix asap? https://code.launchpad.net/~dobey/ubuntu/oneiric/couchdb/fix-780972/+merge/7523818:26
cyphermoxdobey: looking at it now18:31
dobeycyphermox: thanks18:31
slangasekhallyn: do you know what hvm is?  (Bug #849224)18:39
ubottuLaunchpad bug 849224 in qemu-linaro (Ubuntu) "hvm domU doesn't start (qemu-keymaps can't load)" [Undecided,Incomplete] https://launchpad.net/bugs/84922418:39
hallynhardware assisted virtualization?18:39
hallynas in kvm18:40
Laneydoko: do you want to give back fgarch or are you happy the vr removal will have fixed bug #749138?18:40
hallyn(or xen)18:40
ubottuLaunchpad bug 749138 in fgarch (Ubuntu Oneiric) "fgarch version 2110.80-1 failed to build on i386" [High,Confirmed] https://launchpad.net/bugs/74913818:40
Laney(can't reproduce it)18:40
hallynslangasek: interesting.  we might need to talk to lool about that.  /usr/share/qemu is usually provided by qemu-common.18:41
hallynpresumably qemu-linaro shoudl be patched to look under/usr/share/qemu-linaro?18:41
slangasekhallyn: qemu-linaro *already* looks under qemu-linaro.  I don't know what the bug submitter is talking about, because "hvm domU" doesn't tell me what architecture this is...18:41
hallynno, it doesn't :)18:42
hallynslangasek: I also have no idea what v0.1.1-2ubuntu2 refers to18:45
dokoLaney, given back18:46
dokocould you close the report?18:46
slangasekhallyn: ok, I'll not worry over it any more then, we'll wait for the response from the submitter - thanks :)18:47
hallynnp :)18:47
DktrKranzLaney: python-httplib2 uploaded in sid18:49
loolslangasek: hvm: I think this is by opposition with pv; in the latter case, paravirtual, the kernel is aware that it's virtualized, while it's not the case in the hvm case (e.g. to run windows); I think this is xen terminology but might apply to kvm18:58
cyphermoxdobey: done19:09
dobeycyphermox: thanks much!19:10
slangaseklool: ok - so in any case we have no idea what the submitter is doing19:12
bdmurraypitti: is there any chance bugpatterns.xml was out of date on people.canonical.com for a while?19:20
=== chrisccoulson_ is now known as chrisccoulson
bdrungthe ppa builder seems to have a problem on maverick: https://launchpadlibrarian.net/79806678/buildlog.txt.gz19:55
bdmurrayslangasek: is there a multiarch tag in use?20:01
slangasekbdmurray: yes20:02
bdmurrayand that tag is multiarch?20:02
bdmurrayslangasek: and that tag is multiarch?20:03
slangasekbdmurray: yes - sorry, I thought that was the first question :)20:04
* micahg added the multiarch tag to the wiki tags page20:04
bdmurraymicahg: maybe it should be official20:05
micahgbdmurray: I thought I made it official already20:06
micahgindeed, it autocompletes20:06
bdmurraywell neat20:06
loolslangasek: Sorry hadn't looked at the actual bug; seems like someone trying to run a piece of xen by hand which expects the keymaps in usr/share/qemu; let's wait for his reply indeed!20:11
=== tkamppeter_ is now known as tkamppeter
SpamapShmm, how does one make strings in upstart jobs translatable?21:14
JanCSpamapS: what do you mean by "strings in upstart jobs"?21:40
SpamapSJanC: I want to display a message to users when the system is waiting on something for a long time.21:53
JanCthat doesn't sound like an upstart issue, but something service-specific?21:55
SpamapSJanC: no, its one of upstart's jobs21:55
SpamapSbut even if it were, say, apache's.. how would one make it translatable?21:56
SpamapSthe gettext cli program seems a likely candidate21:58
JanCin general, daemon/service-specific messages can only be translated "inside" the daemon, but for anything outside that gettext is an obvious tool of course22:00
broderSpamapS: wouldn't that be more plymouth's job?22:11
SpamapS$PLYMOUTH message --text="Waiting for network configuration..."22:13
SpamapSbroder: not that I can see22:13
broderSpamapS: how does, uh, mountall handle this?22:14
SpamapSthe burden lies on the origin of the literal string22:14
SpamapSbroder: _()22:14
* SpamapS assumes22:14
broderi feel like i've seen a way to access gettext from sh...22:14
slangasekSpamapS: by "make them translatable", do you mean "get them somewhere that they will be translated", or "get the right translated string into the upstart job"?22:15
SpamapSslangasek: both. :)22:16
slangasekthe gettext commandline program should do for the latter22:16
slangasekplymouth message --text=$(gettext -d upstart-jobs "Waiting for network configuration...")22:17
slangasekor gettext -d upstart, if you can work out how to get it into upstart/po/upstart.pot in a semi-automatic fashion22:17
SpamapSThere's a lot of pieces of the boot that are english only.. so I'll leave it in the "lets solve this soon" bin, and not do it now..22:18
SpamapSslangasek: the descriptions of all the upstart jobs would also be good to translate.. since they're used by the plymouth bridge22:19
slangasekSpamapS: hmm, ick :)22:20
SpamapSslangasek: seems like that would be much easier to do in a "semi automatic fashion" with dh_installinit22:20
slangaseknot particularly22:20
slangasekdh_installinit runs when building binary packages22:21
SpamapSslangasek: but then we have to have upstart know to feed it through gettext at the right moment. :-P22:21
SpamapSslangasek: it could scan the init script it is installing, and produce the .pot, no?22:21
slangasekthe problem you need to solve is extracting the strings so that they can be fed to something that interfaces with launchpad translations, which looks at *source* packages22:21
slangasekif LP Translations pulls .pot from binary packages at build time, I'm not aware of it22:22
SpamapSslangasek: err, not installing, but rather, the init script it is setting up to install22:22
SpamapSI see your point..22:22
SpamapSwrong stage of package building22:22
slangasekScottK: the honeyd configure.in is making me angry22:25
SpamapSseriously, I think ESR may be right.. autoconf/automake need replacing. :-P22:25
SpamapSIf for no other reason than to knock all the bad habits out of peoples heads22:25
slangasekI challenge you to show me a build system that doesn't encourage *worse* habits :)22:26
slangasekanyway, honeyd fix in hand now22:31
slangasekwell, except for the subsequent unrelated build failure22:31
charlie-tcaTheMuso: bug 836798 seems to be still happening22:38
ubottuLaunchpad bug 836798 in at-spi2-core (Ubuntu Oneiric) "natty to oneiric upgrade failed: Could not perform immediate configuration on 'python-pyatspi2'" [Critical,Triaged] https://launchpad.net/bugs/83679822:38
TheMusocharlie-tca: Yes I know, I am subscribed to it, I will get to it today.22:39
charlie-tcaJust want to make sure I didn't mess you up with it22:39
slangasekScottK: hah; the other build failures are caused by API changes in libevent, which I see is precisely what puts honeyd on the NBS list22:49
ScottKslangasek: If you want a break for something fun, have a glance at Debian 637509 and then remove it from Ubuntu.22:50
ubottuDebian bug 637509 in ftp.debian.org "RM: dtc -- RoQA; consistently buggy and non-policy compliant" [Normal,Open] http://bugs.debian.org/63750922:50
slangasekthe latter is Debian bug #63276522:50
ubottuDebian bug 632765 in honeyd "FTBFS with libevent 2.0 in experimental" [Normal,Open] http://bugs.debian.org/63276522:50
slangasekheh, dtc22:51
slangasekyeah, we can do that22:51
slangasekoh good, honeyd is by the same upstream author as libevent22:53
slangasekand is not ported to the new API22:54
ScottKOnly 13 RC bugs against dtc in Debian.22:54
SpamapSslangasek: /win 2122:56
* slangasek runs to the window to look!22:56
SpamapSshh don't tell anybody22:56
SpamapSeveryone will want one22:57
ScottKWas it http://www.backcountry.com/sog-knives-double-headed-axe-w-nylon-sheath ?23:04
slangasekScottK: bug #845481 is a treat23:11
ubottuLaunchpad bug 845481 in dtc (Ubuntu) "URGENT: Please sync DTC 0.34.1 from SID" [Undecided,New] https://launchpad.net/bugs/84548123:11
ScottKslangasek: I think you'll like https://bugs.launchpad.net/ubuntu/+source/dtc/+bug/845481/comments/123:12
SeqHello. My boot times in oneiric have skyrocketed to about 2.5 minutes. Can anybody help me see why? Here is the bootchart: http://i.imgur.com/XS36E.png23:15
SpamapSSeq: bug 84778223:16
ubottuLaunchpad bug 847782 in upstart (Ubuntu Oneiric) "installer writes a permanent ethernet entry in interfaces file" [High,In progress] https://launchpad.net/bugs/84778223:16
SpamapSSeq: I am about to propose a fix, that hopefully slangasek will take a look at for me. ;)23:17
SpamapSSeq: the short description is .. if you have interfaces in /etc/network/interfaces that aren't actually able to be configured, you should likely remove the line 'auto ethX' from that file so that the system won't wait for them or try to bring them up at boot.23:18
SpamapSslangasek: https://code.launchpad.net/~clint-fewbar/ubuntu/oneiric/upstart/add-plymouth-messages/+merge/75278 .. would you mind taking a peek?23:20
SeqThanks. I've removed the 'auto eth0' line, although there was no actual configuration indicating it was dhcp.23:21
SpamapSSeq: so it had no configuration at all?23:21
Seqonly an 'auto eth0' line, it did not have anything such as 'iface eth0 inet dhcp'23:22
SpamapSSeq: do you know why it might have been there?23:22
SpamapSSeq: do you have a physical eth0 ?23:22
Seqdid the timeout value in /etc/failsafe.conf change recently? I mostly reboot when on wifi only, and I've been running oneiric for a few weeks now, rebooting every few days (mostly for kernel updates)23:23
SeqSpamapS: yes.23:23
SpamapSSeq: the timeout did raise from 30 -> 120 seconds about a week ago23:23
SeqThat could be it. I probably haven't rebooted in a week23:24
SpamapSSeq: the proposed fix is just to display to the user that its waiting for network configuration..23:24
ScottKThat's not much of a fix.23:24
SpamapSSeq: but I suppose we should also make sure that the interfaces we're waiting for actually have configurations!23:24
SpamapSScottK: your snark cuts me deep. ;)23:25
SeqThat would be a good start. I ditched the 'auto' line so hopefully it gets skipped over. But when I disabled 'quiet' and 'splash', I didn't get any indication either, just that cupsd had started.23:25
SpamapSSeq: right, the MP I just linked to for slangasek has the messages bits added23:26
SpamapSits not in oneiric yet23:26
ScottKIt's not really snark.  I've got a system that boots wicked fast and explaining to me why it's no longer fast once I upgrade it isn't going to make me into a happy user.23:26
SpamapSScottK: can't make everybody happy. :)23:27
SpamapSMost users should *not* have these lines23:27
SpamapSBut I think its worthwhile to also make it not wait on any interfaces that will clearly *never* come up.23:27
ScottKOTOH, compared to a UEC user upgrading to oneiric, only booting a bit slow is a detail.23:27
ScottKSpamapS: Agreed.23:28
SeqI've actually got an issue I'm planning on narrowing down tonight that causes my network to not work. Maybe it is related to bringing up an unconfigured interface..23:28
SpamapSSeq: I would expect that ifup would still bring it "up", just not give it an IP23:28
SpamapSSeq: which is actually what I'm testing right now23:29
SeqIf I'm on wifi and connect my dock (which has ethernet plugged in), each device is configured and working, but my routes get a little messed up. I need to disable both in NM, then pick either one to start again23:29
SpamapSSeq: yeah maybe the auto line was borking your configs23:29
slangasekScottK: bug #84954423:29
ubottuLaunchpad bug 849544 in dtc (Ubuntu) "remove dtc from oneiric and blacklist: multiple security and policy bugs" [High,Fix released] https://launchpad.net/bugs/84954423:30
ScottKslangasek: Thanks.23:30
SeqSpamapS: Unfortunately I'm on that machine right now. I'm about to dock, so I may disappear for a second23:30
SpamapSso, as I suspected, if you have an 'auto ethX' line with no config, it comes up instantly23:34
SpamapSif the interface *exists*23:34
SpamapSone of the whole points of this is we're waiting for interfaces that may take 2 minutes to be detected..23:35
SpamapSSeq__: wb ;) so I tested auto eth1 with and without an eth1 attached to the system.. and without it.. you get the delay, but with it, you don't. This is desired behavior, as one thing we're waitign for is hardware detection that takes a long time.23:39
=== Seq__ is now known as Seq
=== Seq is now known as seq
seqSo you said I shouldn't get the delay if it has 'auth eth0' while disconnected?23:41
seqbtw, I just tested again. If I have both network interfaces active, I lose speedy connectivity. Bringing up a web page takes minutes. If I have just one interface connected (either one) web pages load instantly23:42
seqthis is not necessarily related to the boot issue23:43
SpamapSseq: are they both on the same network?23:47
seqSpamapS: `route -n` > http://pastebin.com/3J11yM9u23:48
SpamapSseq: that is a little weird.. not sure why you'd see the slowdown.. but can't you just turn wlan off when you plugin to ethernet?23:49
seqIf I recall, on previous releases the wifi route had a metric of 1, whereas the wired one was 0. They are both 0 here23:49
SpamapSseq: interesting indeed23:49
seqSpamapS: I could, but I'd prefer it just work magically like it used to.23:49
SpamapSseq: yeah it should be magical like that if they're essentially on the same network23:50
seqI'll try manually removing that route over wifi and adding it back with a metric of 123:50
seqSpamapS: Also, disconnecting wifi would sever existing connections, which I don't necessarily want to happen23:51
SpamapSseq: as for your boot slowdown .. the issue isn't if its "connected" or "disconnected" .. but whether the hardware is present at all... if its not present, we wait.. in case it shows up (not entirely unlikely w/ servers ;)23:51
seqSpamapS: eth0 is always present, it just usually isn't connected23:51
seqI still had to wait 120 seconds23:51
SpamapSseq: if it was present, with no configuration, the system should have booted. I just tested that and we don't wait, because ifupdown brings it "up"23:52
SpamapSI wonder if there are situations where it doesn't show up until later. Hm.23:52
seqIt didn't do that here. Is there something I can poke to get additional output?23:53
seqSpamapS: Never been a problem with it showing up before. It's just a standard e1000 on a standard intel chipset in a thinkpad23:53
SpamapSseq: if you want to pull the failsafe.conf from that branch and put it in /etc/init/failsafe.conf, that would at least tell you its waiting on network configuration..23:53
seqAccording to dmesg, it gets detected pretty quickly: [    2.213280] e1000e 0000:00:19.0: eth0: (PCI Express:2.5GT/s:Width x1) 5c:ff:35:00:6b:0523:54
SpamapSseq: it would actually be really interesting to see ls -l /run/network23:55
cjwatsonwhy is there an auto line with no iface line anyway?  d-i won't do that23:56
seqIs there something better than pastebin to post this to?23:56
SpamapScjwatson: I'm pretty sure we're going to dig up all the hacks and voodoo people have applied to their interfaces file with this change. ;)23:56
cjwatsonthe only places it writes auto, it writes iface straight afterward23:57
SpamapSseq: I like paste.ubuntu.com (easily accessed from cli with pastebinit)23:57
seqThis is a fresh install on oneiric, and I do not touched the interfaces file on machines that I use NM on23:57
cjwatsonseq: which installer?23:57
seqalternate, I use dmcrypt+lvm23:57
cjwatsonI've audited the code and it never writes auto without writesing iface next23:57
seqIs there a log file I can check to see the install date?23:58
cjwatsonso I'd like to see the exact original contents of /etc/network/interfaces plus /var/log/installer/syslog23:58
SpamapScjwatson: perhaps more importantly, if there's only an 'auto ethX' line and it is detected, /etc/network/network-interface.conf brings it up.. so the system won't wait for that.23:58
cjwatsongive me the latter and it has the install date in it plus lots more23:58
slangasekSpamapS: sent some comments on the merge request23:58

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!