/srv/irclogs.ubuntu.com/2009/01/28/#ubuntu-devel.txt

NecrosanJust an FYI, after building sunffb on jaunty, xorg is crashier than ever.00:02
NecrosanWill not run.00:02
=== CarlFK1 is now known as CarlFK
NecrosanNot sure if it's sunffb or xorg itself.00:02
slangasekwhere's the best place to assign bugs in langpacks?  (bug #286065)01:03
ubottuLaunchpad bug 286065 in apt "Swedish translation typo: opeation -> operation" [Undecided,New] https://launchpad.net/bugs/28606501:03
slangasekif I assign it to the actual langpack that's at fault, will it get looked at?01:03
cjwatsonI generally subscribe the relevant translation team and hope they'll pay attention01:04
cjwatsonI'm not convinced bugs against individual language packs get looked at remotely promptly01:04
cjwatsonthough they probably ought to01:04
slangasekoh - n/m, this bug was present in the intrepid source and is fixed in the jaunty source, so I guess not a langpack bug in this case01:05
slangasekbut noted for the future, thanks01:05
cjwatsonwow, how did partman-crypto ever work01:07
cjwatsonit's attempting to find UUIDs by looking at the encrypted block device01:08
* slangasek squints01:08
cjwatson"<shuffle> <shuffle> I guess this looks a bit like a UUID, I'll use that"01:08
cjwatsonthis bug appears to have existed since gutsy ...01:09
cjwatsonbug 32173201:10
ubottuLaunchpad bug 321732 in partman-crypto "Alternate and Server CD installer fails to create encrypted swap parition" [Undecided,Incomplete] https://launchpad.net/bugs/32173201:10
Necrosancj01:11
Necrosanill pastie now, had some car problems.01:11
Necrosanhttp://pastebin.com/m7c3de6401:11
cjwatsonthough even then I'm confused - this code seems to be supplying a UUID where the encrypted block device was desired. Surely that doesn't work?01:12
cjwatsonNecrosan: OK, you described the problem incorrectly earlier :-)01:12
Necrosancjwatson: Eh? how'd you interpret it? =P01:13
cjwatsonNecrosan: aptitude hasn't installed linux-image-sparc64-smp, so there's no bug of the form "installs package without dependency satisfied" here01:13
cjwatsonNecrosan: instead, what it's done is attempted to install as many of the dependencies of that package as it can01:13
NecrosanOh, a miscommunication.01:13
slangasekI'm wondering if that wasn't an unused code path before now; I'm pretty sure I went through partman-crypto to set up my crypted LVM in earlier iterations01:13
Necrosancjwatson: Alright. Good to know.01:13
cjwatsonwhich is perfectly legitimate although (to me) surprising01:13
cjwatsoni.e. the result is still a sound dependency graph01:13
Necrosancjwatson: an FYI, x11 is all messed up on sparc64..01:14
cjwatsonI'm glad my initial diagnosis was wrong01:14
NecrosanBad...01:14
cjwatsonNecrosan: noted but I'm not the person to do anything about it :-)01:14
NecrosanC'mon, take one for the team ;)01:14
cjwatsonI'm directly involved with neither X nor sparc01:14
NecrosanIs there anyone actively working on the 9.04 build?01:15
NecrosanIE, someone to fix the problems before release?01:15
cjwatsonyou've already spoken to the main X developers in Ubuntu01:15
NecrosanAh..01:15
cjwatsonbryce and tjaalton01:15
NecrosanDo either of them  own sparc machines? or do they cross compile?01:16
bryceno01:16
Necrosanwhich one, bryce?01:16
cjwatsonneither is particularly required, since we have sparc build daemons in our datacentre for building binary packages01:16
slangasekcompiling is done by the central buildds01:16
cjwatsonUbuntu developers do not themselves build the binaries you install on your system01:16
NecrosanI see. I'm a little unfamiliar with all ofthat.01:16
cjwatsonwe upload source packages which are built centrally01:16
cjwatsontherefore as long as the source doesn't break, a port can in theory be kept going with very little manual attention01:17
bryceI always build packages against at least amd64 before uploading, sometimes i386 as well, but not really anything beyond that01:17
NecrosanPretty slick01:17
bryces/build/test build/01:17
Necrosanbryce: The stuff builds fine.. just segfaults.01:17
slangasekbryce: presumably not the sunfb package, though. ;)01:17
bryceif I were to modify it and upload it, yeah I'd definitely try building it first ;-)01:18
NecrosanI had to manually go grab libdrm though, as the package is broken horribly on sparc64..01:18
cjwatsonslangasek: it seems to give me some kind of UUID when used in encrypted LVM mode, although I have no idea what it represents01:22
slangasekhmm, I got nothin'01:23
cjwatsonmaybe the LUKS header01:24
cjwatsonah yes01:24
cjwatsonso it'll work as long as you're using the passphrase key type01:27
slangasekasac: hrm; libnss3-1d is using dpkg-symbols, but all of the symbols have been revved to version 3.12.2~rc1?01:42
slangasekasac: that's quite a bit of overhead for the same result as a .shlibs file :)01:43
asacslangasek: thats a one time thing ... because we flipped SONAME01:45
slangasekoh?01:45
asactthe win of this move was that we get back to upstream binary compatibility in both directions01:45
slangasekfrom what I can see, libnss3-1d in hardy already provided /usr/lib/libnss3.so, so there really isn't a need to bump the minver so high in practice01:47
asacslangasek: yes, versions for a bunch of symbols were even below the addition of that symlink01:48
asacso i hit the reset button.01:48
slangasekI don't think I understand01:48
asachmm ... right. sorry misread. i didnt research carefully enough. we probably added the symlink during hardy dev cycle then.01:49
slangasekappears so01:50
asacbut is not really a big issue from what i can tell01:50
slangasekanyway, minor issue - I was really looking at bug #272314 and ended up down a rabbit hole :)01:50
ubottuLaunchpad bug 272314 in update-manager "update-manager disabled multiverse on 8.04 -> 8.10 upgrade (Was: NSPlugin Viewer ERROR: libnss3.so)" [Undecided,Invalid] https://launchpad.net/bugs/27231401:50
* asac looks with tired eyes ;)01:50
slangasekdon't worry about it :)01:51
asacthats not related ;) ...01:51
asacbut the min version would still be good to have i think01:52
slangasekheh, it's related in that I was looking for a shlibs file to parse mentally. :-)01:52
TheMusoasac: Re bug 318985, have you had a chance to try my suggestions?02:07
ubottuLaunchpad bug 318985 in linux "line-in doesnt work for my HDA NVIDIA device" [Undecided,Confirmed] https://launchpad.net/bugs/31898502:07
=== nhandler_ is now known as nhandler
RAOFAnyone feel like applying a 3.7 MiB debdiff to f-spot for me? :/03:30
RAOFBuildsystem changes suck.03:31
* ScottK steps away.03:31
TheMusoRAOF: What build system changes?03:33
RAOFTheMuso: Well, there's autoreconf-ing to make it support passing in CSC, for the mono 2.0 transition.  Then, I need to patch it again to make it pick up Mono.Cairo (libmono-cairo2.0-cil doesn't ship mono-cairo.pc).  Then, there were buildsystem changes (accidentally) rolled up in the .diff.gz for -0ubuntu5, which I rolled into 98_autoreconf.03:35
ajmitchminor things, only 3.7MB03:35
RAOFIf it wasn't CDBS I'd have just switched it to running autoreconf during build.03:36
RAOFActually, that still would've been a 3MB debdiff, but at least the next buildsystem change we needed to make wouldn't be so big.03:39
TheMusoRight.03:49
* TheMuso is not sure he wants to take that on atm.03:49
RAOFEh.  Someone will, eventually.  Someone will want both f-spot and banshee to be simultaneously installable.03:53
ScottKThat takes me off the hook for sure.03:58
RAOFNow that some part of it actually works, KDE4 appears to be kinda neat, yes.04:02
ScottKBased on some of the ranting I've seen here recently I get the impression that KDE decided on purpose to take a hit and redo a bunch of stuff in order to get to a better future and is now starting to come out of it while this is happening to Gnome bit by bit and it's always something new.04:04
ScottKOf course I may just be listening to channel trolls too much.04:04
ScottKSince I don't know much about Gnome it's hard for me to tell.04:05
RAOFScottK: Yeah, that's about right.  GNOME's breaking stuff just about slowly enough that not too much stuff is broken at any one time, it seems.04:16
calcslangasek: ping05:08
slangasekcalc: yo05:08
calcslangasek: what did we decide about the help support wrt search/cd05:08
calcslangasek: dropping help for english off the cd?05:09
calciirc there was the issue of how the user would get it installed05:09
calcversus help being broken without lucene05:09
* calc checks what happens when he removes lucene to see how bad the issue is05:09
calcwell it appears that 'Find' in Help just does nothing at all if its not there05:11
calcno error, no nothing05:11
slangasekcalc: right, it's not resolved how we would get the help package in later; I think embedding a copy of lucene might be simplest?05:12
calchmm yes, i'll see what that entails, brb05:12
slangasekthat solves the problem of not being able to recommend: lucene without pulling in the whole java stack; leaving only the "java is missing so things don't work" problem common to the rest of OOo05:12
calcslangasek: yea that looks doable, i'll just have to verify it still works after a build05:13
calci don't think it will work for jaunty+1 but i can deal with that later, heh05:13
* slangasek nods05:13
calcok, well off to start the build and then to bed :)05:15
* calc wonders why his laptop goes into automatic sleep mode when he unplugs the power cord05:43
calcseems like some sort of ubuntu bug to me05:44
Necrosanor a feature05:44
calcNecrosan: hah05:44
Necrosan=P05:44
RAOFcalc: What does gnome-power-manager think your battery life is like?05:44
calcwell i am running Gnome so perhaps they do consider it a feature05:44
slangasekpresumably it's built-in troll detection. :)05:45
calcRAOF: 99%05:46
calcRAOF: i had to pull the power then bring it back from sleep again to see05:46
RAOFSo, I guess it's not getting freaked out about low power, then.05:47
calcapparently its just braindead05:47
TheMusocalc: something to get debugged at the sprint next week.05:47
calcTheMuso: yea :)05:47
calchopefully its not a heisenbug, i've seen it happen several times already though05:47
slangasekrun gnome-power-manager --no-daemon --verbose, capture the output when it does it, and dump it to a bug?05:47
calcslangasek: would that actually work since it shoves it immediately into sleep mode?05:48
slangasekit'll still be running when you resume05:48
ajmitchmaybe it thinks your battery is at 1% or so05:48
calcok05:48
calcheh heisenbug05:49
calci killed g-p-m then ran it from command line and pulled the power and it was fine05:49
slangasekheh05:49
* ScottK is currently chasing a bug that goes away when you install debug packages.05:49
calci'll have to look at it again later when i reboot05:50
calcScottK: those are fun05:50
calcScottK: OOo has one of those with accessibility05:50
calcScottK: forwarded it to upstream they saw it happened but when building it in debug mode its fixed :\05:50
ScottKRight, well OOo is just chock full of insanity.05:51
dholbachgood morning06:05
* slangasek waves06:07
calcslangasek: hey are you able to sync packages from debian? :)06:10
slangasekcalc: yes06:10
calcslangasek: can you sync suitesparse and lp-solve from experimental for me? :)06:10
slangasekwhat are the chances I'll regret it if I do? :)06:11
calcslangasek: hopefully none :) its the new version OOo 3.0.1 depends on06:11
calcslangasek: i will be working on breaking up suitesparse later after it syncs (probably at the sprint)06:12
calccalc needs lp-solve but only libcolamd from suitesparse06:12
slangasekcalc: ah; will you continue to leave the lp-solve dep out of OOo for alpha-4, then?06:12
slangasek(since alpha-4 comes out during the sprint)06:12
calci can get it broken out early in the week, the issue is calc actually really links to it, so anyone trying to use the solver has a good chance of it blowing up on them currently06:13
calcactually i can try to get it done before i leave for the sprint (maybe tomorrow)06:13
slangasekwgrant: I've gone and subscribed you to bug #303587, since this bug seems to be fixed by merging from Debian unstable and you appear to have TIL :)06:14
ubottuLaunchpad bug 303587 in maxima "Error in CONDITIONS::CLCS-UNIVERSAL-ERROR-HANDLER [or a callee]: Caught fatal error [memory may be damaged] " [Undecided,Confirmed] https://launchpad.net/bugs/30358706:14
calcslangasek: does that sound ok?06:14
slangasekcalc: that seems to imply OOo uploads next week during the sprint so probably not06:15
slangasekanyway, suitesparse and lp-solve synced06:15
calcok06:15
calchmm i'll break suitesparse up before uploading tomorrow then06:15
calcso it will only need one OOo upload (assuming OOo doesn't show up any other issues) ;-)06:15
slangasekthat would do nicely06:16
calci'll try to get it done tomorrow morning06:16
pittiGood morning06:25
=== ara_ is now known as ara
siretartmdeslaur: okay, I've now merged the xine package, but I still need to have a look at the tons of dpatches lool added to it. I currently think that all of them have been merged upstream (at least the package builds fine in jaunty), but I need to double check it07:08
=== tkamppeter_ is now known as tkamppeter
loolsiretart: Cool, thanks07:42
siretartlool: what's the status about the broken translation dpatch?07:46
siretartlool: I've applied it inline now, but I remember you were talking with darren about it?07:46
loolsiretart: So basically Darren didn't want to protect strings in xine-lib in this way07:46
loolsiretart: It's basically assuming that all translations are perfect and you can trust all translators or you review each translation file in full when you merge it07:47
loolsiretart: So feel free to remove the dpatch for these two strings after the next langpack update in jaunty07:47
siretartI don't find that too unreasonable, TBH.07:47
siretartok, I've applied it inline for now, I'll add a note in debian/changelog about that07:48
loolI think it's the case of many upstreams so not worth fighting07:48
siretartok07:48
loolConcerning the patches to port to new APIs, these should be all upstream I guess07:49
loolFinally there's the imagemagick build-deps mess07:49
loolI think I changed them in a way which will work in both jaunty and backports, but in the mean time I've readded the imagemagick compatibility packages07:49
loolI didn't recheck, but I don't think these should have any bad influence07:50
siretartI've dropped them for now, because the package builds just fine in jaunty07:51
siretartTBH, the hole merge was terribly painful07:51
siretartcould you perhaps double check my merge before I upload?07:51
loolsiretart: Hmm ok07:51
siretart(or upload yourself if you are OK? - need to run to work now)07:51
siretarthg clone http://hg.debian.org/hg/xine-lib/pkg/xine-lib-deb-ubuntu07:51
siretartthanks07:52
siretartlool: btw, I've committed your debian/rules cleanup to the debian packaging branchj07:52
loolsiretart: Thanks07:57
* lool shivers on mercurial07:58
LaserJocklool: not a fan?07:58
loolI'm not yet confortable with it07:59
loolSo it's a pain each time I need to use it08:00
=== Jon is now known as Jonbo
=== mcasadevall is now known as NCommander
=== thekorn_ is now known as thekorn
tkamppeterpitti, I tried "bzr get lp:~ubuntu-core-dev/cups/intrepid/", commit, "bzr push lp:~till-kamppeter/cups/intrepid" to create a BZR repo of the newest CUPS SRU for you, but the bzr push falls into an infinite loop.08:54
pittitkamppeter: oh? how does that look?08:54
pittitkamppeter: maybe name the branch "intrepid-till"?08:55
pitti(it should work with "intrepid", too, but maybe not)08:55
tkamppeterpitti, when I look on Launchpad, an empty lp:~till-kamppeter/cups/intrepid is created, telling that there was never pushed anything into it yet.08:55
pittitkamppeter: okay, then that should work; maybe you just didn't wait long enough or so? (should take some 3 minutes)08:56
pittitkamppeter: the branch is visible as soon as it gets created, but the push needs to finish until lp can browse it08:56
tkamppeterI have deleted the branch now and I am retrying.08:57
pittitkamppeter: recursive loop> did you get some output, or did it just get stuck?08:58
pitti(it usually looks like the latter)08:58
tkamppeterpitti, I have also another problem: I have uploaded GS 8.64RC1 yesterday and RC2 today. In both cases they build only on i386 and x86_64, on powerpc, ia64, lpia, ... they do not build.08:59
pittitkamppeter: hm, and it's a non trivial build failure?09:00
tkamppeterThe error message is the following:09:00
tkamppeterThe following packages have unmet dependencies:09:00
tkamppeter  freeglut3-dev: Depends: xlibmesa-gl-dev but it is not going to be installed or09:00
tkamppeter                          mesag-dev or09:00
tkamppeter                          libgl-dev09:00
tkamppeter                 Depends: xlibmesa-glu-dev or09:00
tkamppeter                          libglu-dev09:00
pittioh, then that needs to be fixed, and the build can be attempted again09:01
tkamppeterI uploaded a development snapshot of GS 8.64 last week and it installed without problems.09:01
pittitkamppeter: (don't upload a new version just for a rebuild, that can be done in the web ui)09:01
tkamppeterpitti, I know, todays upload was RC2, it was not an attempt to force rebuilding.09:01
pittitkamppeter: right, I know; just mentioning09:02
tkamppeterI have posted about this problem here yesterday, too, but no one showed any reaction.09:02
apwTheMuso, don't suppose you are about still?09:04
pittitkamppeter: right, it's because mesa failed to build on all those09:05
pittitkamppeter: so don't worry about it for now (mesa should be fixed at least for lpia, though)09:05
pitti^ tjaalton09:05
pittibryce: ^ too09:05
brycepitti: thanks09:06
tkamppeterpitti, now my bzr push is transferring, so the repo should arrive before Christmas.09:12
pitti*chuckle*09:12
tjaaltonpitti: yes, lpia has switched to 2.6.28 so it should now build on it09:14
pittitjaalton: oh, shall I give-back?09:14
tjaaltonpitti: a retry should do09:14
tjaaltonif that's the same :)09:15
tkamppeterpitti, now the bzr push finished, the progress bar was at around 49 %, disappeared, and it said "Created new branch.".09:15
pittiyes, sorry09:15
pittitkamppeter: ok, thanks; please just commit to that one then, and prod me to pull from it once you have updates to be uploaded09:15
tkamppeterpitti, I have already given back lpia.09:15
pittitkamppeter: won't work, mesa needs to build first09:15
pittitjaalton: kicked09:15
MirvArneGoetje/mvo: noticed that desktop team would (still) not have time for language-selector. anyway, I'd hope someone could review and import my bzr branch to fix the bug #31122809:16
ubottuLaunchpad bug 311228 in language-selector "[jaunty] Extra Translations and Writing Aids not installed for the default language" [Undecided,New] https://launchpad.net/bugs/31122809:16
tkamppeterI have already committed before pushing, so the changes should already be inside.09:16
Mirvit's the last bug people encounter in the flow "install non-English Ubuntu from a CD without a network connection and boot to it (and get language support installed)"09:16
tjaaltonpitti: thanks09:17
mvoMirv: I have a look09:17
mvoMirv: patch looks fine, thanks a lot09:22
Mirvmvo: great that it looks proper, I tried a few more ugly approaches before finding that one09:24
mvoMirv: just to confirm, the old way was that it would have prompted the user about missing translation packs and after the next re-open of the language-selector it would have noticed the missing support packages? this now makes it one go (which is much better)09:25
tjaaltonpitti: still failed.. I need to relax the dependency on linux-libc-dev for lpia, and maybe other archs too09:29
Mirvmvo: (just tested it again). no, it never suggested support packages. ie. after suggesting the basic support to be installed, no matter how many times l-s is opened, it will not offer support packages to be installed for the default languages (ie. there is no cross mark for the default language, only the "half mark" (color, no check)09:34
pittisuperm1: great work wrt. moving hotkeys out of hotkey-setup! I'm committing everything upstream now09:43
siretartlool: regarding hg, if you are happy with my merge, just upload it to the ubuntu archive, I'll import the upload to the hg branch09:59
siretartlool: if you have changes and want them committed to the branch, just tell me the url, I'll fetch & push them. otherwise just upload, I'll import them10:01
loolsiretart: I'm uploading as we speak10:01
loolsiretart: I resisted doing a trivial change: moving from XS-Vcs-* to Vcs-*10:02
loolsiretart: I built the source with debuild -i'\.hg|\.hgtags' -S -sa; hope that's fine10:02
siretartlool: yes, that's fine. I'm building it with hg-buildpackage -S, but haven't noticed a difference yet10:03
siretartjust make sure that you grab the orig.tar.gz from debian10:03
siretartah, I'll commit that change in debian, so we'll get that on the next merge10:04
siretartpushed10:05
tkamppeterpitti, the CUPS SRU is successfully committed. I have checked on Launchpad. I have also proposed to merge it into your branch. So you have probably some click button to merge it in now.10:05
pittibryce, tjaalton: hm, (for bug 311895), it seems that Xorg.0.log doesn't report pipe underruns any more? I still get them, though10:05
ubottuLaunchpad bug 311895 in xserver-xorg-video-intel "[i945] spontaneous black screen (major pipe-A underrun)" [High,Triaged] https://launchpad.net/bugs/31189510:05
bryceheya pitti10:07
tjaaltonpitti: that's with the current driver?10:07
pittitjaalton: jaunty du jour, yes10:07
bryceI think eric may have suppressed some error/warning messages10:07
pittiokay; it was useful for counting, but I still see the flickering, so I can still test it10:08
pittiI'm currently trying the "FramebufferCompression" "off" suggestion10:08
ogracjwatson, http://www.nslu2-linux.org/wiki/Debian/TroubleShooting ... acording to martin michlmayr the -r and -k options dont work for slugimage or do i read that wrong ? (you use them both in latest d-i)10:33
=== asac_ is now known as asac
cjwatsonogra: *I* don't use anything10:48
cjwatsonogra: that code is straight from Debian, just with numbers tweaked10:48
ograwell, the d-i buid does10:48
ogra*build10:48
cjwatsonogra: I am not going to get involved in debugging this; somebody needs to work out what's wrong and send a patch10:48
cjwatsonI have already spent far too much time trying to get d-i going on armel10:49
ograi'm not sure i'm reading it right though ... "Don't be fooled by the documentation -- the -r and -k options don't work for Debian-on-slug, because of the APEX bootloader "10:49
ograok, thats fine (you said yesterday to leave you alone with it ... )10:50
=== davmor2 is now known as davmor2-away
pittiRiddell: any idea what to do with the KDE 4.1.4 in intrepid-proposed? only two packges have a bug mentioned in the changelog, so we have no written feedback for them whatsoever11:19
Riddellpitti: ScottK was keeping track of them, I believe there is a bug open11:23
Riddellalthough can't seem to find it11:25
Riddellmm, no, can't find it, we'll need to wait for ScottK-desktop to wake up11:27
dholbachseb128: you were quicker than I was on pochu's vinagre upload :)12:11
dholbachmid-air collision :)12:11
Kalozhey. libasound2_1.0.18-1ubuntu3_i386.deb breaks audio on intel hda for me12:25
Kalozreverting to libasound2_1.0.18-1ubuntu2_i386.deb and restarting pulseaudio fixes it12:25
seb128dholbach: sorry about the race upload ;-)12:26
dholbachno worries12:26
dholbachI'll go and do something else now :)12:26
seb128dholbach: I'm trying to be responsive on desktop requests when I'm not too busy ;-)12:26
pochudholbach: that explains the reject mail - diff.gz already in the archive ;-)12:31
dholbachpochu: yeah12:40
=== LucidFox_ is now known as LucidFox
fabbioneUploading to ubuntu (via ftp to upload.ubuntu.com):13:24
fabbioneConnection failed, aborting. Check your network (111, 'Connection refused')13:24
fabbioneknown issue?13:24
pittifabbione: yes, cocoplum is down ATM13:24
fabbionepitti: ok thanks..13:24
fabbionepitti: do you have an ETA?13:24
pittifabbione: might take a while, thanks to my stupidity we need to recover some files13:25
fabbioneit's nothing urgent anyway.. just wanna flush the queue :)13:25
pittifabbione: an hour or two perhaps13:25
fabbionepitti: no worries man.. i doubt it was your stupidity.. rather your fingers didn't obbey the master :P13:25
pittieither way :)13:26
loolinfinity: Hmm I think we need manual intervention to fix the lpia and probably other chroots to install a newer linux-libc-dev13:26
loolinfinity: This affects e.g. mesa which has a libdrm-dev bdep which has a versionned dep on linux-libc-dev; mesa failing to build affects a bunch of packages13:27
DexterFhi13:32
DexterF8.10 in vmware workstation 6.5.0, linux host: no X driver from vmware tools, won't compile. options?13:32
=== DktrKranz is now known as DktrKranz_nobrow
=== DktrKranz_nobrow is now known as NoBrowser
=== evand_ is now known as evand
=== davmor2-away is now known as davmor2
superm1pitti, great thanks14:12
=== NoBrowser is now known as DktrKranz
* ogra pokes pitti in the ribs 14:21
pittiogra: eek14:25
ograheh14:25
ograpitti, i have some touchscreen issues with the latest hal14:25
ograapparently the "touchkit touch" on the Q1 forcefully gets capabilities input.touch*pad* set so hal uses synaptics (which doesnt work)14:26
ograit should either be evdev or evtouch, where does the synaptics suddently come from ?14:27
ogra*suddenly14:27
DexterF8.10 in vmware workstation 6.5.0, linux host: no X driver from vmware tools, won't compile. options?14:28
ScottKpitti: On 4.1.4 we have two issues we're investigating.  It's working quite well for me and others, but I don't want to say verification is done until we get the issues sorted.14:33
ScottKSo I think we'll get them verified, but it may take several days.14:33
ScottKRiddell: ^^14:33
ograpitti, aha, removing /usr/share/hal/fdi/policy/20thirdparty/11-x11-synaptics.fdi fixes it14:33
tjaaltonogra: lshal should show the capability of input.tablet14:39
ogratablet14:39
ograthat would be wrong as well14:39
tjaaltonwhy?14:39
ograsince touchscreen != tablet14:39
tjaaltonwell that's what it's called in hal anyway14:40
ogratouchscreen us definately closer to touchpad14:41
ograbut it needs to use the right driver14:41
tjaaltonnow it matches 10-x11-input.fdi14:41
tjaaltonif you remove the synaptics one14:41
ograno, it matches evdev14:41
ograerr14:41
ograevtouch14:41
ogralet me remove that as well and see if evdev recognizes the taps14:42
ogragrrr, if only ctrl alt BS would work now to kill the calibration X server14:42
tjaaltonwell, I don't know what evtouch does to it, but it shouldn't list input.touchpad as a capability14:42
ograwell, touchpad isnt to wrong here14:43
ograjust synaptics is14:43
cody-somervilleslangasek, Ugh... the new kernel update prevents Xubuntu users from logging in??!14:43
cody-somerville:S14:43
ScottKbryce: I was wondering what your thoughts on a SRU for Bug #254468 are?  I can guarantee you lots of testers, starting with me.14:43
ubottuLaunchpad bug 254468 in xorg-server "MASTER: momentary video garbage upon drawing new objects (particularly in KDE)" [High,Confirmed] https://launchpad.net/bugs/25446814:43
ScottKcody-somerville: Should cut down on bug reports then.14:43
cody-somerville: P14:44
tjaaltonogra: but because it lists input.touchpad (because of some evtouch trickery?) it'll use synaptics14:44
cody-somervillecharlie-tca, Can you fill me in on what you know?14:44
charlie-tcaI ran the updates to 7.10 yesterday, about 83 of them installed. When I restarted, I got the GDM login screen.14:45
charlie-tcaIt went into a continuous loop using automatic login, just trying to login14:45
tjaaltonScottK: probably not going to happen14:45
ogratjaalton, evtouch doesnt touch any capability settings, it matches only vendor and device names14:45
ScottKtjaalton: Why not?  It's a pretty major bug.14:46
tjaaltonogra: then the kernel gets it wrong14:46
ogratjaalton, removing it *and* the synaptics fdi doesnt make it use evdev though14:46
tjaaltonScottK: I think the performance regression would be greater14:46
ograthere is no x11 driver in lshal at all now14:46
charlie-tcaI logged in through tty, changed to not auto login, and every time I tried to login it just keeps asking for a name and password, never leaving the gdm screen14:46
ScottKtjaalton: Apparently we already decided to take that hit for Jaunty.14:46
charlie-tcaWhen I start the -15 kernel, it works14:46
tjaaltonScottK: yes14:46
tjaaltonwell, for now anyway14:46
ScottKtjaalton: So it's been decided the performance hit is acceptable.14:47
tjaaltonScottK: I can't see it with current intel stack. maybe radeon is safe too since we are using EXA now14:47
charlie-tcaSomething in the 2-6.22-15-generic kernel is blocking me from logging in through the GUI14:47
ScottKtjaalton: Can't see the performance hit or can't see the screen corruption?14:48
tjaaltonScottK: neither14:48
charlie-tcaNo, 2.6.22-16-generic kernel is blocking me14:48
tjaaltonScottK: it was dropped during feisty, but got added back almost immediately14:48
tjaaltonI think the legacy nvidia* would be hit the most14:49
tjaaltonmaybe the problem is more evident in KDE.. it never bothered me much14:51
ScottKtjaalton: It really has an attrocious affect on KDE.  In addition to looking horrible, it causes hangs.14:51
AmaranthWait, you dropped the "make X go fast but break KDE" patch?14:51
AmaranthPut it back, put it back14:51
ScottKyep14:51
tjaaltonAmaranth: why?14:51
AmaranthI wondered why I could watch windows redraw again14:51
tjaaltonAmaranth: on intel? it's not that14:52
AmaranthKDE is only broken because they're not doing something correctly I thought anyway14:52
ogratjaalton, creating an fdi file to make it use evdev doesnt make evdev recognize any input events14:52
ScottKtjaalton: And if we got to upstream and say "We've patched X to make it suck on KDE, please help us work around it" I don't think we'll get a lot of support.14:52
ScottKAmaranth: Tell us what so we can get it fixed.14:52
AmaranthScottK: Did fedora drop the patch as well then?14:53
ScottKCurrently KDE is fine with an unmangled X.14:53
tjaaltonScottK: fedora still has it14:53
ScottKAmaranth: I have no idea.14:53
AmaranthscottK: the upstream bug report says what to do14:53
ScottKGet upstream to accept the patch and then I'll accept it's a KDE problem.14:53
AmaranthscottK: Upstream wrote the patch14:54
AmaranthWell, ajax did anyway14:54
ScottKWhat's the upstream bug report?14:54
AmaranthThere is something wrong with the spec I think that causes the problem14:55
AmaranthscottK: I dunno, I'd have to find the launchpad one first14:55
* Amaranth woke up about 30 minutes ago, brain not firing right yet14:55
tjaaltonAmaranth: so, are you on intel or what?-)14:55
Amaranthtjaalton: Yep14:55
tjaaltonAmaranth: ok, bug 32081314:56
ubottuLaunchpad bug 320813 in mesa "[GM45] with EXA compiz animations cause temporary freezes" [High,Confirmed] https://launchpad.net/bugs/32081314:56
tjaaltoncaused by vblank14:56
Amaranthtjaalton: X3100, vsync disabled in compiz, using UXA14:56
tjaaltonAmaranth: ah :)14:57
tjaaltonI don't have problems with UXA, other than X crashes on resume14:57
tjaaltonoh and the alpha corruption14:57
Amaranthtjaalton: Well for me gutsy was the last one to work right14:57
AmaranthMost people wish we'd go back to hardy graphics, that was the worst one for me14:58
tjaaltonintel is in flux..14:58
AmaranthUntil recently performance in jaunty was pretty good, now it's pretty slow on the desktop yet 3D is nice and speedy14:58
AmaranthSo it's either UXA is really slow or put it back, put it back (the patch) :P14:58
tjaaltonAmaranth: please verify that it would help ;)14:59
ogratjaalton, so any suggestions ?14:59
Amaranthbleh, the bug report in KDE is on my OtherOS partition14:59
AmaranthWhat is the launchpad bug report for KDE being broken?14:59
ScottKAmaranth and tjaalton: We currently have a KDE SRU pending for Intrepid, so if there's a bug that says what needs to change in KDE to work around this patch, please point me at it and I'll see if I can get someone to look into it.15:00
AmaranthThe one where people are screaming for blood to get the patch dropped :P15:00
ogratjaalton, i see it being configured properly in Xorg.0.log as mouse device now ... but apparently no events are getting through at all15:00
ogra(by forcing evdev15:00
ogra)15:00
tjaaltonogra: sorry, not really.. file a bug with the log and output from evtest15:00
tjaaltonogra: might be a kernel bug then15:00
AmaranthscottK: There is a KDE bug report where ajax tells them what the patch does, why, and what to do to KDE to work with it15:00
tjaaltonhttp://bugs.kde.org/show_bug.cgi?id=17046215:00
ogratjaalton, well, they get through with using the evtouch driver15:00
AmaranthscottK: But I have no idea how to find it15:01
ubottuKDE bug 170462 in compositing "Video garbage when drawing new object (Comment #54)" [Normal,Resolved: downstream]15:01
ScottKOK15:01
ogratjaalton, so i'm pretty sure its in the X layer15:01
AmaranthScottK: You up for adding a major chunk of plumbing to Qt? I hear they accept community contributions now :/15:03
tjaaltonogra: then try evtest to be sure15:03
ScottKNope.15:03
ogratjaalton, hrm, why is that in the joystick ackage15:03
tjaaltonogra: no idea..15:03
AmaranthQt needs to support _NET_WM_SYNC_REQUEST on initial drawing, not just resizing15:03
ograweird15:03
AmaranthUntil they KDE is broken on lots of X servers when you use a compositor15:04
Amaranths/they/then/15:04
ScottKAs the bug says, this patch violates the current design, which is why it isn't accepted upstream.15:05
ogratjaalton, hmm, funny, i see events in evtest15:05
ograand even with the proper coordinates15:06
ScottKNow that I've read that, I'm more convinced than ever dropping the patch as a bad hack is the right answer.15:06
AmaranthScottK: It breaks what applications _expect_ the server to do, I don't think it violates the spec15:07
AmaranthScottK: I dunno if SuSE carries this patch but Fedora still does, what do they do to Qt/KDE to not have this problem?15:08
ScottKLast I heard they had the problem.15:08
ogratjaalton, so am i right to assume the kernel side is fine if it passes evtest ?15:08
tjaaltonogra: probably. file a bug against evdev with all the findings15:09
ograwill do ...15:09
* ogra wonders if the joystick driver would work ... *g*15:09
ScottKAmaranth: In the bug report, I find, "<ajax> trying to do what the protocol says is intensely painfully slow since it's a readback from the card and those aren't fast".  That sounds to me like "we don't like the spec so we'll just ignore it."15:10
ScottKSo the right way is to change the spec.15:10
ScottKNot just hack around it and not care what users get screwed in the meantime.15:11
ScottKBut I guess I'm done.15:11
bbshi -- i'm a programmer -- this is the first time i've used ubuntu on my own system coming from gentoo type system15:11
bbswhere can i find the vim / emacs/ gedit syntax files15:11
bbsso i can program with highlighting15:11
AmaranthScottK: I'd still rather not destroy the performance of everything else15:12
ScottKI understand KDE isn't a priority.15:12
AmaranthScottK: There is a clear thing Qt can do here that will fix the problem _and_ make Qt look better on unpatched servers too15:12
hyperairwhere can i get access to ia64 debs?15:12
AmaranthScottK: That is pretty much what I am saying, yes.15:12
jpdshyperair: ports.ubuntu.com15:13
hyperairjpds: ah thanks. i'll take a look15:13
Tm_TAmaranth: without knowing better, doing something that break Qt-related stuff and just point them is not helpful IMO15:13
jpdsbbs: /etc/vim/vimrc15:13
Tm_TAmaranth: point/point at15:14
AmaranthTm_T: Ok then, let's just permanently disable compositing for everyone that way they don't run into the problem the patch is trying to fix.15:14
bbsjpds: everything is installed by default?15:14
bbsthe syntax files15:14
Tm_TAmaranth: that's not the solution either (:15:15
jpdsbbs: You might want to intall vim-full for that.15:15
ScottKAmaranth: Fix the design first.15:15
AmaranthTm_T: Or just disable compositing for kwin since that seems to be where the problems are15:15
AmaranthScottK: We can rewrite the core X protocol now?15:15
* Amaranth goes to do something else15:15
ScottKAmaranth: That or carry an out of tree hack forever15:15
AmaranthWhich one sounds more likely?15:16
Tm_TAmaranth: anyway, for this particular case, I don't comment on details until I know what is this all about15:16
ScottKWhich upstream will never support and we're continually screwed.15:16
Tm_TAmaranth: but, just breaking and waiting others to fix is not nice, it can be done more collaborative way too15:16
AmaranthScottK: One of the main upstream developers wrote the patch and maintains it in Fedora15:16
ScottKAnd KDE is still borked there too.15:16
ScottKJust because he's an upstream dev doesn't make it less of a hack on the protocol.  He even says as much in that bug.15:17
Tm_TScottK: this is the buffer thing?15:17
ScottKYeah.15:17
Tm_T...no point breaking stuff because of that, until it's properly fixed IMO15:18
AmaranthScottK: It does mean as long as he cares about the performance in Fedora the patch will be maintained by someone else15:18
ScottKIt's not the maintenance of that patch that worries me.15:18
tjaaltonTm_T: we've had the patch for 2,5 years.. now who's breaking what again?-)15:19
ScottKIt's the knock on effects and getting upstreams to accept change to work around it.15:19
ScottKWe didn't have KDE4 for that long.15:19
Tm_Ttjaalton: exactly, its broken at begin with (:15:19
ScottKObviously we aren't going to solve this.15:20
tjaaltontrue15:20
AmaranthI consider Qt broken15:20
Tm_TAmaranth: I consider too15:20
AmaranthSo then fix Qt or wait for someone to fix Qt15:20
Tm_Ttjaalton: anyway, IMO this is the kind of issue that must be taken cautiously, it's not helping anyone to just break things for some people to help performance elsewhere, unless it really is greater benegit15:20
Tm_Ttjaalton: this in short term, in long term it must be fixed everywhere properly15:21
Mirv(tjaalton: nice to know X is crashing for you, too, on resume)15:21
AmaranthTm_T: At this point dropping the patch would be a large regression in performance15:21
JontheEchidnathen compiz should be fixed!15:21
Tm_TAmaranth: how large?15:21
tjaaltonTm_T: no-one has been breaking anything, it's been there for a long time15:21
Tm_Ttjaalton: I know15:21
JontheEchidnakwin works just fine without the patch, compiz must be broken15:22
Tm_TJontheEchidna: heh15:22
AmaranthJontheEchidna: The bug was reported against kwin's compositor?15:22
JontheEchidnaI'm saying, kwin has no slowdown after removing the patch. If compiz does it must be broken15:22
tjaaltonJontheEchidna: maybe not with your driver15:23
Tm_Ttjaalton: Amaranth: JontheEchidna: but anyway, really, thinking compiz or kwin is more important than other is not a solution, we don't like to divide community, do we?15:24
AmaranthTm_T: I'm absolutely only thinking about GNOME/Compiz users15:24
Tm_TAmaranth: I know that15:24
AmaranthTm_T: The rest of them either don't have the problem or have always had the problem15:24
Tm_TAmaranth: I'm not, because I have to answer to every user, not only those15:25
AmaranthTm_T: If you change the patch we have a regression15:25
Tm_TAmaranth: and Qt breaking isn't regression then?15:25
AmaranthI think we've had the patch longer then we've had Qt415:25
Tm_Tperhaps15:25
Tm_Tperhaps not15:25
* Tm_T remembers doing Qt4 stuff for over 3 years now15:26
AmaranthI know we've had it longer than KDE 415:26
Tm_TAmaranth: I dealt with KDE4 3 years ago first time15:26
Amaranthtjaalton would be able to tell you for sure but I thought we added that patch for dapper15:26
Tm_Twell -2 month but anyway15:26
AmaranthTm_T: When was it in the archive?15:26
Tm_TAmaranth: I have no idea, I do a lot with upstream15:27
tjaaltonAmaranth: nope, for edgy (Mon,  7 Aug 2006)15:27
AmaranthTm_T: My point is the patch is older then our having KDE4 or possibly even Qt4 in the archive so from an ubuntu point of view KDE4 has always been broken and changing the patch is a regression for others15:27
Tm_TAmaranth: could be15:28
Tm_TAmaranth: that doesn't mean we as a whole, not you alone, should ignore qt15:28
JontheEchidnaQt4 has been in the archives since 200515:29
AmaranthSo fix Qt, doubt it'd take more then a week15:29
Tm_TAmaranth: anyway, are you willing to wait until qt4 side is fixed?15:29
JontheEchidnahttps://launchpad.net/ubuntu/+source/qt4-x1115:29
Tm_TJontheEchidna: thanks15:29
AmaranthIt'd be a nice test for community involvement in Qt Software15:29
RiddellI've filed a task with Qt upstream15:29
tjaaltonTm_T: the patch is disabled in jaunty (for now anyway)15:29
Tm_TRiddell: thanks sir, bug number?15:29
AmaranthFixing this would make KDE look better on unpatched X as well so it's a win-win15:30
RiddellTm_T: no number until it's public15:30
Tm_TRiddell: roger15:30
Tm_TAmaranth: true15:30
Tm_TAmaranth: that's what I meant by collaborating15:30
Tm_TAmaranth: one side soloing isn't helpful (:15:31
Tm_Ttjaalton: I see, thanks for info, so, are we now evaluating situation or what?15:31
* Amaranth tries to free up HD space to build X with the patch again15:31
Tm_TRiddell: have any our friend trolls been poken about it?15:31
tjaaltonTm_T: yes.. nvidia updated their legacy drivers to support the new xserver video ABI, so once tseliot has updated the packages we should know how they are affected15:32
Tm_Ttjaalton: appreciate all info in the future too, son (:15:33
tseliotI'll do it soon15:33
RiddellTm_T: 15:29 < Riddell> I've filed a task with Qt upstream15:33
tjaaltontseliot: now! :)15:33
Tm_TRiddell: ah, with! thanks15:33
Tm_Ttjaalton: these issues are messy, I notice15:34
tseliottjaalton: I'll go back in time and since "now" has just passed15:34
tseliots/and//15:35
tkamppeterpitti, I have added also the fix for bug 309314 to the CUPS SRU. Can you merge it into the Intrepid CUPS repo? Thanks.15:38
ubottuLaunchpad bug 309314 in cups "Documents silently fail to print" [Undecided,Fix committed] https://launchpad.net/bugs/30931415:38
=== ssweeny_ is now known as ssweeny
pittitkamppeter: did you see my merge request reply? did you pull before?16:02
ScottKpitti and tkamppeter: Can we arrange it so all core-dev don't have to get emailed about those?16:03
pittiScottK: I don't know how... I get annoyed by those, too (getting merge requests for firefox etc.), but ATM there's nothing in LP to change that :(16:04
ScottKI think there have been some recent changes.16:04
* ScottK looks....16:04
pittitkamppeter: ah, apparently you did; looks fine, pulled16:05
ScottKpitti: https://code.launchpad.net/~ubuntu-core-dev/cups/intrepid/+subscription/ubuntu-core-dev - Unsubscribing the team will, I think, do it.16:06
pittiScottK: I set it to "no email"16:07
ScottKpitti: Excellent.16:07
tkamppeterpitti, thanks for taking that one in.16:08
pittiScottK: thanks, didn't know about that16:08
ScottKpitti: It's a recent development.  A few weeks ago and went and bitched about it in #launchpad and they pointed me at the feature.16:09
=== dholbach_ is now known as dholbach
slangasekcharlie-tca: was '7.10' a typo?16:37
dholbachslangasek: so... sponsoring :)16:43
slangasekdholbach: it was actually less a comment on the sponsoring queue per se, than a repeat of my kvetch from UDS about there being packages in main with authoritative repos that core-dev don't have commit rights to16:44
dholbachslangasek: ah, I see - sounds like something we should discuss on the mailing list16:45
dholbachslangasek: I don't think that "make ubuntu-core-dev join a bunch of other teams" will help16:46
cjwatsonthe main reason I haven't just added ubuntu-core-dev to the ubuntu-installer team is that I think it would generate a lot of mail16:46
dholbach*nod*16:46
cjwatsonin principle, I think it would be correct enough, actually16:46
Keybukthe way this was always envisioned as working was that you had a lp bzr branch that ubuntu-core-dev can commit to, that matches what's in the archive16:46
cjwatsonperhaps somebody could write up a primer on LP mail handling as it affects team memberships16:46
Keybukand you have a second team branch16:46
cjwatsonI actually do that for ubiquity16:46
Keybukwhich the entire mozilla team can commit to16:46
slangasekdholbach: I don't think there are "a bunch" of other teams; I know of 416:46
dholbachin an archivereorg + bzr-for-everything world it will be different, I don't know what the best way to do it now16:46
Keybukand when you do releases, you merge from the team branch into the core-dev branch16:47
slangasekmozilla, OOo, network-manager, and (apparently) installer16:47
cjwatsonbut I often forget to push to the core-dev branch (I certainly don't want to merge if I don't have to)16:47
dholbachslangasek: artwork stuff too16:47
cjwatson(I'd rather revision numbers stayed the same)16:47
dholbachdocs too16:47
slangasekdholbach: oh, those don't count because I don't ever commit to them. ;)16:47
slangasekthough, I think g-p-m is also in this state, now that I think of it16:47
cjwatsondesktop16:48
Keybukwhat broke when we set the core-dev mailing address, can anyone remember?16:48
cjwatson~ubuntu-desktop/vte/ubuntu was one I ran into recently16:48
slangasekdholbach: anyway - I agree it should go via mailing list; I'm not volunteering to drive this discussion just at the moment.  Was just kvetching. :)16:48
cjwatsonKeybuk: I think the "matches what's in the archive" constraint is weird, because it doesn't hold if a core-dev not in the team commits something but doesn't think it needs to be uploaded right away16:49
dholbachslangasek: kvetching... is that something like czechnology?16:49
slangasekif czechnology is Yiddish, then yes16:49
Keybukcjwatson: right16:49
dholbachslangasek: no, not really :)16:49
cjwatsonKeybuk: so the effect of that policy is that the branch is up-to-date for anything core-devs commit, but not for anything that the team most qualified to work on the package is concerned, which seems precisely backwards16:49
Keybukcjwatson: I agree with you there16:49
cjwatsonKeybuk: which is why I just push ubiquity to the ubuntu-core-dev branch when I feel like it ;-)16:50
cjwatsonreally, though, I'd rather that the branch be "ubuntu-core-dev + ubuntu-installer"16:50
seb128slangasek: hello, so the g-s-d changes fix your issue, good ;-)16:50
Keybukbranch permissions?16:50
cjwatsononce they exist ...16:50
KeybukI don't think they are even planned16:50
cjwatsonif ubuntu-core-dev can have a mailing address set, I can just make it a member of ubuntu-installer and not care16:50
cjwatsonKeybuk: they are, as part of the distro branch namespace work16:51
Keybukcjwatson: we tried setting a mailing address for it16:51
Keybukit broke some things16:51
james_wcjwatson: you would like the branch permissions to be different to the upload permissions?16:51
cjwatsonKeybuk: i.e. eventually it'll just be ubuntu/+trunk/ubiquity or whatever, and anyone who can upload ubiquity can commit to it16:51
slangasekseb128: yep - like magic!16:51
Keybukoh, interesting16:51
Keybukthat would certainly fix the problem16:51
cjwatsonjames_w: perhaps, but that isn't required for this; we could just say that some extra people can upload ubiquity16:52
gesermvo: about bug 243550: I've just tested your patch, but it doesn't work (for me). gdebi still lists all recommends if I enable recommends again in my host system (I've checked, it's still disabled in the pbuilder)16:52
ubottuLaunchpad bug 243550 in python-apt "gdebi --root doesn't use the apt config from the chroot but the one from the "host"" [Medium,Triaged] https://launchpad.net/bugs/24355016:52
cjwatsonwe originally moved it to ubuntu-installer because evand wasn't in core-dev but needed to work on ubiquity, and I was fed up merging his stuff on autopilot16:52
james_wcjwatson: I think that makes more sense at this point.16:52
mvogeser: yeah, thanks for testing it16:52
mvogeser: sorry that its not quite there yet, I'm investigating16:52
cjwatsonI *would* like extended branch permissions at some point, but we could get by without16:53
gesermvo: does it matter that I'm still on intrepid?16:53
james_w(p.s. lp namespace and branch permissions based on upload permissions are well on the way)16:53
dholbachslangasek: here's where I first stumbled over the word: http://jimmac.musichall.cz/log/?cat=3116:53
mvogeser: the patch I did initially had some bad side effects,  I need to redo it I think16:54
gesershould I add my test results to the bug?16:54
mvogeser: feel free, but I hope to get a new version up soonish :)16:55
=== azeem_ is now known as azeem
charlie-tcaslangasek: Sorry, No, 7.10 was not a typo.16:59
slangasekcharlie-tca: ah.  so cody-somerville was hitting the panic button about a bug in gutsy then?17:00
charlie-tcaI'm running a fresh install now, to see if it happens again17:00
charlie-tcaGutsy is 7.10. We do still support it, right?17:00
slangasekthere's security support for it.  You're not going to get much support for other bugfixes.17:01
charlie-tcaSo if the kernel won't boot, what?17:01
charlie-tcaWe just tell the user "too bad, it is not a security issue"?17:02
slangasekcharlie-tca: direct them to the forums for a workaround that will let them upgrade to a version of Ubuntu whose kernel *is* actively maintained?17:02
charlie-tcaBut the update to the new kernel was just this month!17:03
slangasekoh?17:03
charlie-tcaThe kernel was upgraded from 2.6.22-15 to -16 and that is what broke17:03
slangaseksorry, I thought you were saying that you had just recently upgraded your system to 7.10 from an older release17:03
slangasekif it's a regression *in* a security update, yes, that should be tracked down and fixed17:04
charlie-tcaThe -15 kernel will start, the -16 kernel will loop the GUI login screen17:04
charlie-tcaOnce it upgrades the kernel, you can not log in at the GUI17:04
slangasekcharlie-tca: is there a bug report open for this issue?17:10
charlie-tcaNo, not that I can find. I can file it, though17:11
slangasekcharlie-tca: please do, and give me the bug number so I can escalate17:12
charlie-tcaOkay, I will.17:12
slangasekalso, can you capture an strace of what gdm is doing when it's supposed to instead be logging you in?17:13
charlie-tcaI can try, but so far none of the logs are showing anything.17:13
slangasekI wouldn't expect them to; hence, 'strace'17:14
charlie-tcaI'll do that17:14
=== warp10_ is now known as warp10
pittiRiddell: FYI, bug 297152 is the boost transition thing, and immediate uploads just for this aren't really required (but thanks for your kdepim upload)17:31
ubottuLaunchpad bug 297152 in linux-ports "boost -> boost1.35 transition / demote gcc-4.1 and gcc-4.2 to universe" [High,In progress] https://launchpad.net/bugs/29715217:31
Riddellpitti: yep, found that out17:32
keescharlie-tca: that sounds like video driver version mismatch.  did all the other kernel-related packages get updated too?17:38
charlie-tcayes, to the best of my knowledge17:39
Kanohi, could syslinux upgraded to 3.72 to use hybrid iso mode of isolinux?17:42
Kanoor newer... like in experimental17:43
Kano3.7317:43
keescharlie-tca: when you've got a bug #, let me know, and I can poke around at it.17:45
Kanothen you could call isohybrid on the created iso image17:53
sorenKano: What would that do?17:53
Kanowell you can cp the iso image 1:1 onto hd or usb stick17:53
Kanoand it is bootable17:53
Kanolike that moblin iso, which is created that way17:54
sorenI'm almost certain we can already do that.17:54
Kanoyou have got a pretty bad installer for usb stick17:54
sorenI've done some testing of extboot in kvm that suggests that that is the case, anyway.17:54
sorenKano: Perhaps I'm not fully understanding what you're saying.17:55
cjwatsonsimply changing the way the image is written to the USB stick would not change the other problems with running from a USB stick in any way, and those are more urgent.17:55
cjwatsongenerally speaking upgrading syslinux is a complete pain to do because we have to upgrade the gfxboot patch as well17:55
Kanosoren: cp image.iso /dev/stick17:55
Kanois what you do, then you can boot it17:55
cjwatsonso I try to avoid it whenever possible. If you want it faster, send a fully-tested patch that preserves all of the things we use17:56
sorenKano: I'm 99% sure that works now.17:56
Kanocjwatson: why dont you use the standard routines?17:56
cjwatsonKano: because people screamed at us until we did gfxboot17:56
cjwatsonand it's a pain to change now17:56
Kanosyslinux has nice gfx support17:56
cjwatsonsoren: no, you have to use usb-creator17:56
sorenKano: But I appear to be wrong.17:56
cjwatsonKano: it's different, and would require substantial porting. I will not be harangued by you about it, I'm afraid17:57
cjwatsonand it didn't have any of that support when we first added gfxboot17:57
Kanovesamenu17:57
cjwatsonwould require substantial porting work17:57
cjwatsonread my lips17:57
cjwatsonit's not just as simple as saying that the facilities exist17:58
cjwatsonI would get shouted at if I made it look less pretty17:58
sorencjwatson: Hm... That is interesting, actually. I had a libvirt bug once where Fedora CD's failed to boot. The conclusion was that due to some misunderstandings between libvirt and kvm, we were booting the CD's as though they were hard drives, and for some reason, Ubuntu and Debian CD's worked fine, but Fedora ones didn't.17:58
cjwatsonwhich I think is pretty shallow, but that's not my problem17:58
cjwatsonsoren: they might use some different bootloader on their CDs, e.g. grub17:58
calci have a question i am modifying suitesparse to break a library out and it uses cdbs, how do i make it link automatically to the other package (being produced at the same time) or should it know how to do that already and something possibly broken?17:59
cjwatsonright now, we depend very extensively on the customisations that have been applied to gfxboot17:59
Kanowhich are not working with pxelinux btw.17:59
calciirc dh_shlibdeps (or something like that) is supposed to do it, but i thought cdbs does this automatically?18:00
sorencjwatson: I think they also used syslinux, but ICBW. Anyhow, this sheds new light on the issue.18:00
cjwatsonKano: yes, I am aware that gfxboot doesn't currently work with pxelinux; there is a bug about it18:01
cjwatsonnobody has sent a patch yet AFAIK and I haven't had time to dust off my assembler and figure it out18:01
Kanocould you put that into a module?18:02
cjwatsonput what into a module?18:02
Kanoyour gfxboot hack18:02
cjwatsonit's not my gfxboot hack, it's SuSE's18:02
cjwatsonask them, I'm not interested in that level of work on it18:02
Kanoi know that gfxboot is from suse18:03
cjwatsonnor do I think I would do a good job on it18:03
TheMuso8/c18:03
* ogra really wonders how "cp image.iso /dev/stick" would work though ... 18:03
* calc thinks he found the issue18:03
ograare you sure you didnt mean dd ?18:04
Kanoogra: you can use dd, cat or write your own app, all would work18:04
cjwatsonour gfxboot customisations are in gfxboot-theme-ubuntu, which is written in gfxboot's own interpreted language. They would not make a lot of sense as a standalone syslinux module18:04
ograbut not cp ... thats what confused me18:04
Kanocp too18:04
Kanono diff18:04
cjwatsoncp image.iso /dev/stick => replace the /dev/stick device node with a regular file18:05
Kanojust dont forget sysnc18:05
Kanosync18:05
cjwatsonhmm, that's interesting, cp foo devicenode does actually appear to write to devicenode18:06
ograhmm, i always thought you had to pipe a binary stream to it ...18:06
calcyep got it working18:06
ograwhich is what cat and dd would do18:06
cjwatsonI take it back18:06
ion_cjwatson: echo foo >foo >bar; strace cp foo bar: cp just open()s bar for writing and write()s to it.18:09
cjwatsonyes I already did18:09
slangasekcalc: should already do it automatically; is it not working?18:09
slangasekcalc: oh, scrolling down, n/m18:10
slangasek:)18:10
Kanobtw. when you use: default vesamenu.c3218:11
Kanothen you see at least the logo and have got a menu18:12
Kanoof course you have to put that module on the iso18:13
calcslangasek: i found the problem it uses a .symbols file i forgot to update18:14
calcslangasek: so it was claiming the symbols were in the wrong file18:14
slangasekaha18:14
slangasek:)18:14
* calc notes he isn't sure how good a feature that is ;-)18:14
slangasekdpkg-gensymbols has an option to make it complain if there's a mismatch18:15
slangasekhmm, except the levels are wrong, warning if libraries have disappeared should be a lower level than warning if symbols have been added18:16
slangasekdoh18:16
slangasekcharlie-tca: have a bug number yet for us? :)18:35
=== ivoks_ is now known as ivoks
DexterFdoes ubuntu 8.04 support the GeForce 8x00 chipsets out of the box or is there a backported kernel that does?18:44
directhexDexterF, are you talking plug-in gpus, or motherboard chipsets?18:48
DexterFdirecthex: chipsets18:48
directhexDexterF, you be wantin' Envy.18:51
directhexenvyng-gtk package18:51
DexterFuh... no, *not* gpus - chipsets. sATA ports, usb and such.18:51
calcthe nvidia geforce 8300 chipset came out after 8.04 so probably not supported i would imagine18:52
DexterFthere goes that plan18:52
DexterFany backported kernels available?18:52
calcafaict it wasn't even reviewed by sites until ~ July 200818:53
calcDexterF: 8.10 might support it18:53
DexterFnot an option18:53
directhexsata ports should be AHCI these days shouldn't they?18:55
directhexUSB hasn't changed for about a decade18:55
calcthe newest kernel in hardy is 2.6.24-23.4618:55
DexterFall irrelevant if the chipset is unsupported18:55
calcDexterF: you could try booting a live cd and see if it works18:55
directhexi don't think you know exactly what "the chipset is supported" means though18:55
DexterFsystem is in planning stage right now18:56
DexterFdirecthex: in order to talk to the board components the chipset has to be supported by the kernel. it's quite obvious I'd say18:56
calcbut yea in general its not a good idea to buy a chipset newer than the version of linux you want to run on it ;-)18:56
DexterFdang I need a board with nvidia based onboard video...18:57
directhexyou'd be wrong, though. the chipset's individual components, as reportsed to given busses, is what matters18:57
DexterFthe old 7050 work fine I know but I need a *tad* more punch18:57
calcdirecthex: i don't know if its as problematic as it used to be but generally linux has had to add quirks, etc for new chipsets18:57
directhexso whilst sata_nv might be too old, as an example, EHCI is EHCI and usb2 will just work regardless of vintage18:57
calcDexterF: so why is it you have to run 8.04? :)18:57
directhexversion matching for mythtv at a guess18:58
calcDexterF: even if there was a backported kernel you would have to somehow get the machine bootstrapped probably on another machine before upgrading the kernel18:58
DexterFfriend's machine, 8.10 means kde4 which I can't support because I don't use it. kde4 quite frankly and imo sucks wet farts from dead pidgeons18:58
calcDexterF: you could try installing just the kernel from intrepid or jaunty i suppose18:58
directhexcalc, nah, too much changed18:59
DexterFhmm, manually backport, eh?18:59
calcdirecthex: oh? need new userland for them?18:59
directhexand building ubuntu kernels manually sucks18:59
* calc is running jaunty kernel on intrepid, but hasn't run hardy in a long time18:59
directhexcalc, lots of things changed around, so you'd lose the entire support structure for, say, nvidia graphics drivers18:59
calcdirecthex: ah ok18:59
DexterFi see where this is going18:59
* calc avoids nvidia due to their stance on floss ;-)19:00
DexterFif that fglrx drivers wasnt that crappy I'd go for a 780G board...19:00
directhexin the end, ubuntu has a particularly bad mechanism for supporting new hardware in old LTS releases19:00
calcwell thats not completely true i run nvidia card with the nv driver because ati at the time had no fanless dual dvi cards i could find19:01
calcDexterF: intel ftw :)19:01
DexterFcalc: $$$-issue. not applicable. that is.. wait. hm. the X3100 series packs a punch in terms of integrated video, right?19:02
directhexdepends. what do you want it to be able to do?19:03
directhexgames are right out19:03
calcintel cpu can be really cheap eg the c2q 830019:03
calcand are much faster than amd19:03
calcbut yea no gaming on them, probably not too good gaming on any integrated chipset really19:03
calcsome can run games only 5 years old instead of 10 ;-)19:04
calcDexterF: the one good point about amd is that there are open drivers in active development with specs so you are less likely to get bitten by ugly nvidia bugs19:04
charlie-tcaslangasek: no, not yet. I'm trying to verify it on another system first.19:05
calclike not being able to see text on menus in applications19:05
slangasekcharlie-tca: well, even if it's only reproducible on a single system, regressions in security updates are a critical issue19:05
ograpfft, menus are overrated anyway19:05
directhexi wonder whether intrepid will install on my new pc, once i get the broken motherboard issue solved19:05
directhexi was surprised enough that it worked on this laptop19:05
calcDexterF: but no fglrx is not the open driver :)19:06
calci'm waiting to buy a new computer until ~ dec/jan at the end of the year19:06
DexterFcalc: I fell for ATi PR chatter for five f?!*ing years, I hardly have adequate words for how through I am with their video cards19:06
slangasekso is libxcb upgrading ok for folks now on jaunty, as opposed to being held back in update-manager?19:06
slangasek(I already forced the upgrade on my system, so don't have a convenient test case anymore)19:06
calcDexterF: phoronix has articles about the work, it has taken a few years to get them running (aiui)19:06
calcDexterF: and nvidia still has no real open source drivers at all after claiming they would help floss over 10 years ago now, so ymmv ;-)19:07
directhexDexterF, i benchmarked it about 5 years ago. ut2004 performance, £100 geforce ti4200 versus £300 ati 9800 xt... guess which was 3x faster than t'other :)19:07
superm1calc, was it ever discerned if that was a driver bug, or did it only happen with compiz running (and is possibly a compiz implementation bug?)19:07
superm1i remember hearing about it a long time ago, but never followed it19:07
calci was in a discussion with nvidia (iirc on irc) back around 199819:07
calcsuperm1: aiui it only happens on nvidia so it is almost certainly their bug19:08
DexterFI've been monitoring the open drivers dervelopment lately and from what I saw/heard full featured stable drivers are *way* in the future, 2010 perhaps when everything current is more outdated than socket A boards and don't get me started on moronix' fanboi drivel19:08
calcsuperm1: but i don't know the current status19:08
directhexDexterF, phoronix mean well, they just don't really understand how to benchmark things properly19:08
DexterFthey are fanbois19:08
superm1calc, well it only happens with them doesn't mean it's "their" bug, it still could just be that their driver emphasizes a bug with a lower layer.  i know that was happening with gnome screensaver a release or two ago19:09
DexterFeach time I glimpsed in it was "NEW!! ATi free drivers now *even better!!"19:09
superm1in this case turned out to be a compiz bug19:09
DexterFyeah, like: it doesn't crash now when changing resolution or such19:09
directhexDexterF, you don't think that counts as "even better"? :)19:09
calcDexterF: well if you do go with nvidia do realize your hardware is going to become unsupported in the not too distant future, just look at the list of drivers for nvidia in ubuntu, we have to maintain 5 sets of drivers because nvidia dropped support for older cards at various times19:09
DexterFI switched to nv finally and while I'm aware of the perf issues I can't describe how happy I am with that evil closed src driver19:10
calcDexterF: we have 71, 96, 173, 177, 18019:10
DexterFspeaking of which: I _don't_ habe 180 in 8.04. any way to get it? I'd like the perf extra plus...19:10
calcback when i used nvidia it would always crash when i tried to switch to console19:11
directhexDexterF, sure. don't be 8.0419:11
DexterFkde4 still = teh suck19:11
directhexcalc, anyway, i have similarly bad glitching on intel19:11
RiddellDexterF: can you stop flaming please19:11
calcDexterF: xubuntu? :-)19:11
DexterFRiddell: k19:11
directhexcalc, if i allow usplash to use its default resolution, i can't switch to a console after x starts19:11
calcdirecthex: ah ok19:11
directhexcalc, and sadly usplash doesn't correctly work with widescreen19:12
calcdirecthex: luckily we'll be rid of usplash with 9.1019:12
ogradirecthex, until them patches are hapily accepted ;)19:12
ogra*then19:12
directhexogra, i had a look at the code, and ran screaming19:12
directhexogra, both at the code and at the idea of telling mjg59 that something sucks19:13
ogradirecthex, just call it a challenge ;)19:13
DexterFcalc: I'd like nv to port over the old stuff, too, yes, just pointing out everything works with those. 3D, Xv... with ATi couldn't get even decent Xv.19:13
directhexDexterF, ati sorta kinda got xv support in driver 8.5 or so19:14
calcDexterF: nv can't since nvidia refuses to give out any documentation at all19:14
calcDexterF: the ati drivers are better because amd released nearly full docs19:15
DexterFdirecthex: fglrx had Xv ever since, only problem was serious artifact issues over month they didn't solve. last time I checked was... mmh... 8.3 or 8.4.19:15
calcDexterF: they won't even release documents for the nv0119:15
DexterFcalc: yeah, sad story. thing is they dont really need to.19:15
directhexcalc, zomg trade secrets!!!19:15
calcDexterF: yea what with intel having more market share than them ;-)19:16
DexterF?19:16
calcDexterF: nvidia isn't in a market dominant position so probably should be releasing documents19:17
calctheir stock price is less than 1/4 what it was a year ago19:17
calcthey had product recalls that nearly killed them as well19:17
DexterFdidn't know about the stock thing. well, I guess their main interst is windows gamers. I came to think ATi made the open src move because they were desperate before the 48x0 series rolled in real cash19:18
calciirc intel has the vast majority in the market for video chipsets with ati and nvidia trailing fairly far behind19:19
directhexthe real cash is in OEM sales19:19
calcDexterF: not really before the amd buyout even finished they had already announced they were going to release the documents19:19
DexterFcalc: intel? Oo19:20
calcdirecthex: yea and that is where nvidia has been having trouble, esp with all the bad parts19:20
calcDexterF: in the past there have been graphs published showing the breakdown of the market between the vendors19:20
calcslangasek: i think the new suitesparse i uploaded might be stuck in new19:34
calcslangasek: i haven't received an email about it yet though, it has a new binary package19:34
slangasekdid you get confirmation of the source upload?19:34
slangasekyeah, I see the binaries in new; processing19:35
calcslangasek: not yet19:35
TheMusoKeybuk: Afaik you are working on initramfs-tools. Are you likely going to upload soon, or is it safe for me to do an upload today?19:35
calcslangasek: thanks :)19:35
slangasekcalc: oh, the binaries I see are from 3.2.0-1, which is the one we synced - if there's another upload, I don't see it19:35
calcyea a 3.2.0-1ubuntu1 upload19:35
KeybukTheMuso: I'm not doing any work on initramfs-tools right now19:35
calcslangasek: hmm maybe i should try uploading again19:36
TheMusoKeybuk: Ok thanks.19:36
slangasekcalc: doesn't show up on https://launchpad.net/ubuntu/+source/suitesparse, so if it's been more than 15 minutes, yeah probably19:36
calcslangasek: ok19:36
calcslangasek: should it show up there even if it is new? btw i just uploaded again19:37
slangasekcalc: yes, the source should show up there.19:38
calcok19:38
calconce i see it finished building i'll let you know for the new processing bit19:39
superm1tseliot, tedg should we be pulling in http://svn.gnome.org/viewvc/gnome-power-manager?view=revision&revision=3167 to fix the high CPU usage bug on gnome power manager? does jaunty have RR1.3 yet so we could do that?19:40
superm1(bug 307306 for reference)19:42
ubottuLaunchpad bug 307306 in gnome-power-manager "upgrade to 2:1.2.99.2-0ubuntu1 makes session utterly slow" [High,In progress] https://launchpad.net/bugs/30730619:42
=== hyperair_ is now known as hyperair
maxbsuperm1: I'm running with that patch locally - it works, and helps, but doesn't completely fix things (as per my later comments in the bug)19:47
superm1maxb, considering it's upstream, perhaps we should pull it in then to at least get the behavior improved until a final resolution is found.19:47
maxbIt would undoubtably be appreciated by people testing Jaunty on problem setups who haven't done local builds.19:48
superm1maxb, have you explored at all, do you have any idea why there are so many randr calls happening in the first place?19:49
AmaranthYou mean X will stop chewing 50% CPU for 3 minutes after gnome-power-manager starts soon? :)19:49
maxbUnfortunately my X-fu is insufficient19:50
Amaranthsuperm1: The call they are making used to return a cached result, now it probes19:50
maxbI'm hoping the log I posted will mean something to someone19:50
AmaranthAnd apparently GDK sends the "monitors-changed" signal when you change the brightness which is broken19:51
superm1Amaranth, so who is really to blame then and should be fixed? GPM to cache the result and/or make it less frequently?19:52
Amaranthsuperm1: There is another function they can use to get the cached result in randr 1.319:52
superm1Amaranth, is that the call that was added in the patch i posted above, or you are meaning a different one yet?19:52
Amaranthsuperm1: probably that one19:53
superm1Amaranth, "XRRGetScreenResourcesCurrent" ?19:53
Amaranthyep, that one19:53
Amaranthgnome-settings-daemon needed patched as well, iirc19:53
superm1okay well i'll get this into an upload to GPM then, it's upstream already, and way annoying.19:54
Amaranthoh, the gsd one was actually in gnome-desktop which we have already19:55
calcslangasek: i uploaded suitesparse again 30m ago and i still haven't received an email and its not showing up on LP, any ideas?20:07
slangasekcalc: see other channel where you've been highlighted. :)20:08
slangasekthere are some delays right now with package accepts20:09
tseliotsuperm1: my patch for g-p-m is ok but we still need other two patches for gnome-control-center. And yes we have randr 1.320:09
calcslangasek: thanks :)20:10
tseliotsuperm1: we're discussing with upstream on the patches for gnome-control-center here: http://bugzilla.gnome.org/show_bug.cgi?id=56816020:16
ubottuGnome bug 568160 in libgnome-desktop "Gnome Settings daemon causes high CPU usage with an expensive call" [Major,Resolved: fixed]20:16
mohbanaam i the only one who is experiencing a really slow file browser (the gnome one)20:26
mohbanayet the  kde file explorer - dolphin - is very fast20:26
mohbanait's very noticeable20:26
=== mthaddon changed the topic of #ubuntu-devel to: Launchpad is going down from 22:00 UTC until 23:00 UTC for a code update | Archive: open, MoM running, alpha-3: released | Ubuntu 8.10 released! | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/Helpi
thinkgnuwhich one is better for my system (dual core2  , 1GB ram) ?  kubuntu32 but or kubuntu64 bit ?20:41
Riddellthinkgnu: user questions in #kubuntu20:41
thinkgnu:D20:41
=== sabdfl1 is now known as sabdfl
* directhex smiles sweetly at tseliot and superm1, asks gently about nvidia 180.22 in intrepid21:00
tseliotdirecthex: I'll work on it tomorrow21:01
* directhex offers tseliot a muffin21:01
tseliotdirecthex: keep an eye on https://bugs.launchpad.net/bugs/32241621:01
ubottuLaunchpad bug 322416 in nvidia-graphics-drivers-180 "update request: nVidia ver 180.25, stable" [Medium,In progress]21:01
directhexcan't find 180.25 on nvidia.com :/21:02
directhexat any rate, wanna be sure my new pc works, when i eventually get a fixed motherboard ;)21:03
superm1tseliot, okay i pushed the GPM patch in, so that should be some improvement at least.  thanks for writing it :)21:05
tseliotsuperm1: great, thanks21:05
calcslangasek: can you process suitesparse now, or does it have to be built on all archs (for the new processing bit)21:17
calcits still working on armel/hppa21:18
charlie-tcaslangasek: bug 32249721:44
ubottuLaunchpad bug 322497 in linux-source-2.6.22 "gutsy 7.10 will not login after updating to kernel linux-source-2.6.22.16 *regression*" [Undecided,New] https://launchpad.net/bugs/32249721:44
charlie-tcacould not reproduce on two different systems with fresh installs21:44
ScottKcharlie-tca: I'm having similar symptoms on 8.10 right now.21:48
ScottKcharlie-tca: If you do a console login does df -k say there's no space left?21:48
charlie-tcaI'll go look21:48
charlie-tcaThis just started on 7.10 after the latest kernel update21:49
charlie-tcaScottK: no, mine shows lots of space left; over 40% unused21:52
ScottKcharlie-tca: And it shows block in the available column?21:52
charlie-tcayes21:52
ScottKOK.  Then I've got a different problem.21:53
charlie-tcayeah, quick glance shows it about right, too21:54
ScottKMine is basically no user level ability to write to disk because it thinks it's full.21:54
ScottKI can write as root and delete as user, but writes as user fail.21:55
charlie-tcaI see. This is different.21:55
ScottKIf anyone has thoughts on that, I'm all ears.21:56
slangasekcalc: if I make it before the LP maintenance it can be, sure21:58
calcslangasek: ok21:58
slangasekkees: ^^ seen bug #322497?21:59
ubottuLaunchpad bug 322497 in linux-source-2.6.22 "gutsy 7.10 will not login after updating to kernel linux-source-2.6.22.16 *regression*" [Undecided,New] https://launchpad.net/bugs/32249721:59
slangasekcharlie-tca: thanks, fixed up the bug state21:59
calc"Launchpad will be going offline for maintenance very very soon." lol21:59
keesslangasek: I'll check in a moment... doing kernel updates.  ;)21:59
slangaseksure21:59
charlie-tcaYou're welcome. I hope there is enough there to count22:00
keesslangasek: will you be around after LP is back to do some pocket-copies for me?  (kernel updates from -security to -updates)22:02
slangasekcalc: should libsuitesparse-dev have a dep on libcolamd-3.2.0, given that it ships the .so for it?22:03
slangasek(it currently depends indirectly)22:03
slangasekcalc: anyway, LP maintenance window hit before I got a chance to accept it; will try again in an hour22:04
calcslangasek: ok np22:28
calcslangasek: not sure wrt your question, i saw it was getting pulled in by libsuitesparse-3.2.0 already so just left it at that22:29
calcslangasek: apparently this will probably be fixed in debian soon as well, so we hopefully will be able to go back to syncing in the near future22:29
* slangasek nods22:30
cody-somervilleslangasek, Hows that Xubuntu killing bug going?22:37
slangasekcody-somerville: it's been filed and the security team is in the loop22:38
keesslangasek: you've meaning 322497?22:40
slangasekyes22:41
slangasekcharlie-tca: why are all of these vboxsf mounts happening at login time?22:49
slangasekcharlie-tca: and I bet you don't have a virtualbox-ose modules package to match your kernel, because vbox isn't in main so didn't get updated with the kernel security release22:50
slangasekcharlie-tca: also, please check for gdm output in /var/log/auth.log; the strace shows some of it being written, but it's truncated because I didn't ask you to use -s :)22:51
charlie-tcaThe virtualbox came direct, I ran /etc/init.d/vboxdrv setup to update it.22:51
slangasekok22:52
slangasek[pid  4995] 07:36:56.235593 writev(2, [{"gdm[4995]: PAM adding faulty mod"..., 71}, {"\n", 1}], 2) = 7222:52
charlie-tcaI don't know why all the mounts happen at login22:52
slangasekthat looks promising; pam_gnome_keyring.so is referenced but absent22:52
charlie-tcayou're right about /var/log/auth.log22:54
charlie-tcalet me get a copy out22:55
charlie-tcaslangasek: added the auth.log to the bug report23:00
slangasekthanks23:00
charlie-tcaThank YOU. I know you are very busy23:01
calchow often do packages get put in the archive (eg ACCEPTED -> DONE)?23:01
slangasekcharlie-tca: could you please confirm that 'grep pam_gnome_keyring /etc/pam.d/*' returns only 'optional' lines?23:01
charlie-tcahow?23:02
slangasekcharlie-tca: well, regressions in security updates are kind of a priority23:02
slangasekcharlie-tca: run the command 'grep pam_gnome_keyring /etc/pam.d*', and tell me if there are any lines that don't say "optional"23:02
=== mthaddon changed the topic of #ubuntu-devel to: Archive: open, MoM running, alpha-3: released | Ubuntu 8.10 released! | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-intrepid | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs | this week: https://wiki.ubuntu.com/UbuntuDeveloperWeek
charlie-tcaslangasek: no lines at all23:03
charlie-tcanothing comes up23:03
slangaseker, oops, wrong command23:03
slangasekcharlie-tca: grep pam_gnome_keyring /etc/pam.d/*23:03
charlie-tcaslangasek: two lines, both with 'optional'23:04
slangasekok23:04
slangasekso it doesn't appear to be an authentication failure23:04
slangasekcharlie-tca: I notice that the PIDs from the logfile and from the strace don't match; I guess they're from two different tests?  (it probably doesn't matter, just want to be sure)23:05
charlie-tcayes23:06
cjwatsoncalc: hourly23:06
charlie-tcaI been in there so many times I got confused23:06
slangasekcharlie-tca: the problem is in /etc/gdm/PostLogin/\:0 - could you send a copy of that file to the bug?23:11
charlie-tcaslangasek: sure, let me get it.23:12
slangasekthis is the script that's trying to do all of the vboxsf mounts; it's the only thing run between the 'session opened' and 'session closed' lines, and there are a lot of failures, with the script eventually exiting non-zero23:13
slangasekso this still points back to a vbox problem, ultimately23:13
charlie-tcaslangasek: Oh, no. Do I just pull that and try to login, then?23:14
charlie-tcaOh, sent the :0 file to the bug report23:15
slangasekcharlie-tca: you could try that, yes23:15
slangasekcharlie-tca: alternately, you could change the 'exit' at the end of the script to an 'exit 0'...23:16
calccjwatson: ok23:18
charlie-tcaslangasek: I renamed the file, it did not allow the primary user to log in, but did allow another one.23:18
charlie-tcaIs this all my fault, then?23:19
cjwatsoncalc: (this is aside from all the problems there've been specifically today of course)23:20
slangasekcharlie-tca: well, as far as I can tell the login failures were the result of your local PostLogin script; but the reason for that script to start failing may still point to a regression that should be addressed23:20
charlie-tcaI lied. wrong system23:20
calccjwatson: ok, i see it got marked as done, i guess it takes a while after that for it to show up on archive.u.c?23:20
slangasekyes, it's marked as 'done' at the very beginning of the publisher run23:21
cjwatsoncalc: yes - the bug is closed at the ACCEPTED->DONE transition, but that's actually at the ... what slangasek said23:21
calcslangasek: ah i see :)23:21
slangasekand the publisher run lasts ~1h23:21
charlie-tcaslangasek: I haven't changed that script in a very long time.23:21
calcwell sometime tonight i should have new OOo in after its deps get finished anyway ;-)23:21
slangasekcharlie-tca: sure.  I think your virtualbox-ose setup is broken following the kernel ABI change, which may be an Ubuntu issue23:22
slangasekcharlie-tca: did I understand you to say that you weren't using the virtualbox-ose packages from the archive...?23:22
charlie-tcaslangasek: Okay, it does allow login with out the script.23:23
* slangasek nods23:23
TheMusosuperm1: Re bug 32250, do you think that could have anything to do with the missing kernel hid quirk modules from the initramfs? I uploaded an initramfs-tools earlier to add the hid quirk modules.23:23
ubottuLaunchpad bug 32250 in gnome-power-manager "g-p-m critical power warning states "will power-off" which is misleading" [Medium,Fix released] https://launchpad.net/bugs/3225023:23
TheMusosorry sorry bug 32250623:23
ubottuLaunchpad bug 322506 in busybox "Caps lock doesn't work in initramfs environment" [Undecided,New] https://launchpad.net/bugs/32250623:23
TheMusosuperm1: ^^23:24
cjwatsoncaps lock has occasionally been a source of console-setup-related glitches23:24
slangasekcharlie-tca: it should also allow login if you put the script back, but edit the last line to say 'exit 0' instead of 'exit'; that can be a permanent change to make the script more robust, then we can focus on debugging your vbox problem23:24
cjwatsonis this consistent across keyboard layouts?23:24
cjwatsonsuperm1: attaching /etc/default/console-setup would probably be a good thing, as well as confirming that that file is also present in the initramfs23:25
charlie-tcaslangasek: I am running virtual peul. It is not virtualbox-ose23:25
charlie-tcavirtualbox peul23:25
charlie-tcaI installed it before virtualbox-ose was in the repositories, and the system has been upgraded to 8.0423:25
TheMusocjwatson: ah ok.23:27
Hobbseehrm.  Preseeding packages won't get their recommends, will it?23:27
slangasekcharlie-tca: ok.  what does 'lsmod | grep vbox' give you?23:30
cjwatsonbug 28194023:30
ubottuLaunchpad bug 281940 in tasksel "Debian-installer can't be (and isn't) told not to install recommends by default." [Wishlist,Confirmed] https://launchpad.net/bugs/28194023:30
cjwatsonerr, no, maybe not that one23:30
charlie-tcaslangasek: nothing23:30
slangasekcharlie-tca: output of find /lib/modules -name 'vbox*.ko', please23:31
cjwatsonI'm sure there was a bug about allowing pkgsel/include to install recommends23:31
cjwatsonoh, I can't find it because Timo fixed it :) bug 31536323:32
ubottuLaunchpad bug 315363 in pkgsel "add an option to install "Recommended" packages too" [Wishlist,Fix released] https://launchpad.net/bugs/31536323:32
cjwatsonHobbsee: on jaunty, pkgsel/install-recommends=true tweaks it; on earlier releases, no23:32
charlie-tcaslangasek: 2 lines, '/lib/modules/2.6.22-15-generic/misc/vboxvfs.ko'23:32
slangasekcharlie-tca: the other line?23:33
charlie-tca'/lib/modules/2.6.22.14-generic/misc/vboxvfs.ko'23:33
Hobbseecjwatson: ah, excellent, thanks.23:33
charlie-tca'/lib/modules/2.6.22-15-generic/misc/vboxvfs.ko'23:33
slangasekcharlie-tca: yep.  So, you need to rebuild your peul kernel module for the new kernel23:33
charlie-tcasame line with -1423:33
slangasekcharlie-tca: presumably you had to do this before for the -14 -> -15 update23:34
charlie-tcaI sure don't remember it, but you are most likely correct.23:34
slangasekcharlie-tca: so it's not actually a regression in the linux security update, but the vbox folks should provide an updated virtualbox-ose-modules package for gutsy as well; I've fixed up the bug to reflect that23:35
charlie-tcaThank You. I'll go rebuild it then.23:35
* Hobbsee invalidates a bug, claims it for 5-a-day23:36
* directhex opens 5 bugs for Hobbsee to close23:37
Hobbseedirecthex: darn you ;)23:37
RAOFHobbsee: Like to add making f-spot installable with monodevelop, banshee, et al your next bug?23:46
HobbseeRAOF: not overly, but nice try ;)23:46
pwnguinhow 'bout eclipse 3.4? ;)23:47
Hobbseenot that crazy...23:47
RAOFAll it needs is for you to apply a debdiff.  Nice and simple :)23:49
pwnguinRAOF: speaking of closing bugs, what should we do about bugs against nv?23:50
pwnguinask them to retest on nouveau?23:50
RAOFnv is still maintained by a guy at nVidia.23:50
RAOFFeel free to ask them to check on nouveau; that'd probably be useful/nice.23:51
RAOFBut we're not replacing nv with nouveau for Jaunty.23:51
pwnguindoes nv support anything nouveau doesn't? obviously it might be buggy, but that "guy at nVidia" seems to dislike features23:51
pwnguinso long term i'd imagine a replacement is going to happen. (certainly not for jaunty)23:52
RAOF"That guy at nvidia" really doesn't like features, no.23:52
RAOFnv probably supports some newer cards than nouveau does; nouveau apparently doesn't support nv9x chips at the moment.23:52
slangasekwho can blame him, no good has ever come of features23:52
slangasek;)23:52
pwnguinRAOF: if you have a nv9x chip, it's probably connected to a display that nv wont drive properly23:53
* RAOF rather likes the fact that nouveau actually dithers for his laptop's 18bit LCD properly.23:53
pwnguinbut there's not really a point to debating this23:54
pwnguinim not arguing for the removal of nv23:54
RAOFDebating?  Feel free to suggest people test nouveau.23:54
RAOFIf it drives their card, it'll almost certainly be better than nv.23:54
pwnguinRAOF: what if a bug comes in against nv, is confirmed broken, the tester tries nouveau and drops nv entirely?23:55
pwnguindo we leave the bug open?23:55
RAOFI'd say yes.  It's still a bug in nv, and nv is still our default nvidia driver.23:55
pwnguini kinda wonder how often nv is used in practice23:56
RAOFEverytime someone with nvidia hardware loads a livecd, certainly.23:56
pwnguingood point23:56
pwnguinRAOF: perhaps an invitation to test should be published, so people know it's available?23:57
RAOFThat sounds like a job for ReleaseNotes™23:58

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