/srv/irclogs.ubuntu.com/2010/10/01/#ubuntu-devel.txt

SpamapSI have an old hard drive in storage that is running woody00:13
SpamapSShut down my old home server and never turned it back on..00:14
nh2hi, I'd like to develop new features for rosetta (launchpad translations), but there are quite a lot of branches. Which one should I use as a base?00:52
ajmitchnh2: you'll want to talk to people in #launchpad-dev about that00:54
nh2ajmitch: thanks00:54
=== jjohansen is now known as jj-afk
=== ivoks is now known as ivoks-afk
achianghm, is there something else i need to do while testing X clients inside a chroot? i've exported my DISPLAY=localhost:0.0, and issued an xhost + in the host as well02:28
* achiang decides to re-ask in #ubuntu-x02:29
=== asac_ is now known as asac
=== dendro-afk is now known as dendrobates
=== SolidLiq is now known as solid_liq
dholbachgood morning06:34
kirklandslangasek: what were those two packages?  for windows compilation?06:45
kirklandslangasek: mingw32 and nsis or something?06:45
=== amitk-afk is now known as amitk
kirklandslangasek: found them ...06:58
=== almaisan-away is now known as al-maisan
pittiGood morning07:06
dholbachhi pitti07:14
pittihey dholbach, how are you?07:30
dholbachpitti, great - and you?07:30
pittipretty well, thanks07:31
pittiRiddell: can you please review my apport, kerneloops, and cups uploads?07:46
pittiI'll take a look at the others07:46
pittijames_w: heh, seems we both uploaded a kerneloops for disabling for final07:47
Riddellpitti: ok08:16
yao_ziyuanwhen you release an alpha or beta, you're supposed to build apps with more debugging information, and therefore they would run slower than a final release, right?08:17
pittiyao_ziyuan: no, optimization/compiler settings don't change between alphas and release08:18
pittiyao_ziyuan: we strip off debugging information and publish them as separate pacakges on http://ddebs.ubuntu.com08:18
yao_ziyuanpitti: and this is since when?08:19
pittiyao_ziyuan: compiler switches -> always; ddebs> since 7.04 I believe08:21
yao_ziyuanthanks08:22
YokoZarcjwatson: new icoutils is now in the archive, if you want to utnubu that back to Debian that'd be fantastic08:35
YokoZaryao_ziyuan: individual packages can still tweak them though08:36
tkamppeter_mvo, hi08:39
=== smb` is now known as smb
=== smb is now known as smb`
=== smb` is now known as smb
mvohey tkamppeter_08:57
=== tkamppeter_ is now known as tkamppeter
tkamppetermvo, in your update-manager fix for bug 647460, when you remove foomatic-db-gutenprint, do you replace it with ijsgutenprint-ppds?09:03
ubottuLaunchpad bug 647460 in update-manager (Ubuntu) "10.04 -> 10.10: printconf/ebox-printers/foomatic-db-gutenprint block the upgrade" [High,Fix released] https://launchpad.net/bugs/64746009:04
=== spike_ is now known as spikeWRK
mvotkamppeter: not yet, I prepare a update for this09:10
=== dendrobates is now known as dendro-afk
tkamppetermvo, OK.09:14
* samiz is away: Away09:22
persia!away > samiz09:26
ubottusamiz, please see my private message09:26
persiaHobbsee, Your ideas are stronger than launchpad: ~ubuntu-universe-sponsors refuses to be deleted :)09:37
Hobbseepersia: hah09:39
Hobbseetake that, launchpad!09:39
cjwatsonachiang: [[ isn't POSIX sh, no.  See https://wiki.ubuntu.com/DashAsBinSh which lists it.09:39
cjwatsonYokoZar: thanks, will look after the maverick release09:39
apwcjwatson, have we done anything in our CD images to support EFI for maverick ?09:46
Riddellpitti: did you do two kerneloops uploads?  I accepted one but I don't know if it's the same one I looked at09:49
pittiRiddell: the other is from james_w; we can reject one09:49
cjwatsonapw: yes, supported on amd64 only09:50
cjwatsonapw: (sadly can't really test it on my test machine since am awaiting fixed firmware)09:50
apwcjwatson, there are reports that this is breaking mac's on boot (amd64 only, which now makes sense).  presumably this is something to do with how the CD ismade09:53
apwbug #63398309:53
ubottuLaunchpad bug 633983 in linux (Ubuntu Maverick) "Can't boot Maverick beta amd64 live cd on macbook (EFI)" [High,Confirmed] https://launchpad.net/bugs/63398309:53
apwso i presume this isn't going to be a kernel issue ?09:53
cjwatsonapw: it's hard to say10:01
cjwatsonapw: but, you know, there are reasons I left the i386 CD BIOS-only ...10:02
cjwatsonapw: the only way to prove that it's specifically due to EFI support would be to remaster the CD without that10:02
cjwatsonand there's very little we can do about this without removing EFI support, it's a catch-2210:03
apwcjwatson, and indeed they indicate the 32bit one does boot ok10:03
cjwatsonbut that's not really proof10:03
cjwatsonwould need to remaster the amd64 CD10:03
apwno indeed not anything more than indicative a test is appropriate10:03
apwis that someting one can do anywhere other than on the cdmaster machine?10:04
cjwatsonsure.  loop-mount image on PATH.  genisoimage -r -V 'Ubuntu 10.10 amd64' -o maverick-desktop-amd64-remastered.iso -cache-inodes -J -l -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-load-size 4 -boot-info-table PATH10:04
cjwatsonactually you might need to loop-mount and then copy the whole thing10:04
cjwatsonso: mount -o loop maverick-desktop-amd64.iso old-amd64; cp -a old-amd64 new-amd64; genisoimage -r -V 'Ubuntu 10.10 amd64' -o maverick-desktop-amd64-remastered.iso -cache-inodes -J -l -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-load-size 4 -boot-info-table new-amd6410:05
apwcjwatson, ok so that sounds like something i can try and do ... people are targetting the kernel on this one and blaming me :/10:05
apwthanks10:05
cjwatsonwe could release-note "you have to use i386 on Macs"10:06
cjwatsonnot exactly ideal but ... dear Apple, please fix your EFI implementation10:06
apwcjwatson, well indeed, from the pointeres in the bug it seems even M$ cannot make ISOs which boot correctly on them10:07
apwcjwatson, so i suspect its going to be that in the end, i wonder if we could have ports cds10:07
cjwatsonthe eternal disk space pain10:07
cjwatsonit's possible10:08
cjwatsonprobably, just about10:08
bobbaHi - I'm wondering where I should go to ask how I'd customise the installer? In particular, I'd like to know how I could remove EXT4 as an option for installations as I only wish to support installations on EXT310:28
cjwatson'rm /lib/partman/valid_filesystems/ext4' should be sufficient10:31
cjwatson(I think)10:31
cjwatson#ubuntu-installer is generally a better place to ask though10:32
bobbaand then it should default to a different FS? because EXT4 is the default at the moment10:32
bobbaahhhh fantastic :)10:32
cjwatsoner, yeah, you would need to change the default that's true10:32
cjwatsonare you happy to rebuild packages to achieve this (more work, but more flexible), or do you want to just fiddle about with the live filesystem?10:32
bobba*nod* Well that's getting part way there which is further than I've got by myself ;)10:33
cjwatsonthe default filesystem is in debconf, as the default for the partman/default_filesystem question10:33
bobbaIdeally I'd prefer to fiddle with the live system (of course...) but since I'll be building a fresh ISO then I'm happy to consider anything that's needed10:33
cjwatsonwell in that case you can edit /var/cache/debconf/config.dat and look for partman/default_filesystem10:34
cjwatsonalternatively, add partman/default_filesystem=ext3 to the boot options10:34
bobbaahhh that sounds good - but I'd still have to change the life system to remove it from the valid list10:36
bobbafantastic though - I'm going to try that right away10:36
cjwatsonright10:37
* samiz is away: Away11:01
nigelb!away | samiz11:01
ubottusamiz: You should avoid noisy away messages and -nicks in a busy channel like #ubuntu, or other Ubuntu channels; it causes excessive scrolling which is unfair to new users. Use the command "/away <reason>" to set your client away silently.  See also «/msg ubottu Guidelines»11:01
=== diwic is now known as diwic_lunch
Davieycjwatson: Can I ask you review a debdiff for bug #651027, please?  I just want to make sure i understood your suggestion.11:12
ubottuLaunchpad bug 651027 in dovecot (Ubuntu Maverick) "dovecot-postfix claimed as built in debian/control but not actually built" [High,New] https://launchpad.net/bugs/65102711:12
Davieycjwatson: http://pb.daviey.com/vZXO/raw/11:13
didrockshey, is there a release team member having some time to look at sync request in bug #651266 ? current plugin is broken as the feed change. I can do a fakesync if required11:14
ubottuLaunchpad bug 651266 in totem-plugin-arte (Ubuntu) "Sync totem-plugin-arte 0.9.2-1 (universe) from Debian experimental (main)" [Undecided,Confirmed] https://launchpad.net/bugs/65126611:14
pittididrocks: looking11:14
bobbacjwatson, I've found the 05ext4 file to delete, but I was trying to change config.dat for the partman/default_filesystem setting to set it to ext3 - however, the config.dat doesn't seem to have the setting in? It just says "Name: partman/default_filesystem" "Template: partman/default_filesystem" "Owners: ubiquity".  I'm therefore clearly missing something :)11:14
didrocksthanks pitti ;)11:14
OdyXHrm . How is it that pyside is in "pending publication" state for 3 days ?11:14
bobbacjwatson, sorry - I've figured it out!11:16
bobbathanks11:16
pittididrocks: done; off to lunch now11:16
didrockspitti: thanks a lot! enjoy your lunch :)11:16
=== al-maisan is now known as almaisan-away
cjwatsonbobba: maybe templates.dat, I forget exactly11:18
bobbait is templates.dat - sorry for asking before actually looking deeply enough11:18
cjwatsonDaviey: looks like what I meant, except that you removed "ChangeLog" from the dh_installchangelogs line which I don't think was intentional?11:18
cjwatsonDaviey: I would have provided a diff but I hadn't test-built it and debdiffed the binaries to make sure they were coming out right11:19
cjwatsonOdyX: because it's in the NEW queue11:19
Davieycjwatson: Yeah, that wasn't intentional... My next step was to build and see - but i just wanted to make sure that is what you were thinking before i moved on.11:20
OdyXcjwatson: oh. "just wait" is the remedy to that, right ? :)11:20
cjwatsonOdyX: correct11:20
Davieycjwatson: Unless, you want to carry on with it?11:20
cjwatsonDaviey: no, please go ahead11:20
Davieyok, great.11:20
=== diwic_lunch is now known as diwic
ograi'm looking for an archive admin to review powervr-omap3, opengles-sgx-omap3 and devmem211:50
Davieycjwatson: For info - certainly looks like that debdiff (with ChangeLog added) fixed the issue - http://pb.daviey.com/1uDq/raw/11:59
=== MacSlow is now known as MacSlow|lunch
* Daviey hmmmmmmm's12:02
cjwatsonDaviey: looks good, just check that the new dovecot-postfix's contents look sane and that none of the other packages have changed contents unexpectedly12:13
Davieycjwatson: Yes, fair comment.12:15
Davieycjwatson: Well there is another issue i'm fixing now with the depends of the transitional package.12:16
ogradoko_, the java buildd is ready for you, got instructions for lamont ?12:29
doko_Ng: is working on it12:30
ograoh, ok12:30
lamontthe instructions had better consist of "schedule the build on that box"12:30
Chipzzjava buildd? ^^12:30
cjwatsonChipzz: bug 64234412:31
ubottuLaunchpad bug 642344 in libspring-2.5-java (Ubuntu Maverick) "libspring-2.5-java needs an initial manual build" [High,Confirmed] https://launchpad.net/bugs/64234412:31
lamontChipzz: newer hardware than what we have had12:31
lamontcjwatson: no12:31
lamontopenjdk-612:31
lamontat least I think so12:31
ograright, bug 60504212:32
ubottuLaunchpad bug 605042 in eglibc (Ubuntu Maverick) "[armel] java fails to start with eglibc-2.12-0ubuntu4" [High,Triaged] https://launchpad.net/bugs/60504212:32
Chipzzlamont: when I saw java buildd, I was thinking along the same lines as "i386 buildd"12:32
lamontChipzz: well, java is totally like its own architecture. that much is true12:33
cjwatsonoh ok12:33
Chipzzlamont: well, yeah... when is ubuntu on java scheduled to be released? ;)12:34
lamontcjwatson: libspring was last seen waiting for RC to happen12:34
lamontChipzz: hopefully not before I'm dead12:34
cjwatsonlamont: RC has happened :-)12:34
Chipzz*grin*12:34
Chipzzlamont: my initial reaction to myself was: who has been smoking what kind of crack, and more importantly, where can I get some of that? ;)12:35
lamontcjwatson: yeah - I'm out of the loop on the ticket today12:36
lamonts/the/that/12:36
apwcjwatson, fyi i've built tested (it boots) and posted that remastered CD image for testing on the bug ...12:40
cjwatsonta12:41
cjwatsonbe sure to make it clear that this is not a fix we can simply roll into the distribution (because it breaks other things) but simply an attempt to narrow things down12:42
apwcjwatson, added a rider to it ... that is why i wondered if we might be wanting 'ports' style cd-image.  ones which arn't mirrored etc for the smaller corner cases like this12:44
cjwatsonas I say, yes it's possible if there's no alternative12:44
apwprobabally something that needs discussing in a more formal setting, to see what alternatives we have12:44
cjwatsonwe had an i386+mac port in the past for other reasons; the same kind of thing could be resurrected as amd64+mac12:45
cjwatsonit's not even all that difficult12:45
apwapple sort out your firmware being the prefered option12:45
cjwatsonquite, but Fedora have had the same problem for a while and obviously neither they nor MS have got anywhere12:45
cjwatsonI shouldn't think "fails to boot non-Apple OS" is very high on Apple's priority list12:45
apwi would think "make sure it fails to boot non-apple OS" is high on their priority list12:46
cjwatsonultimately ... your hardware is hostile to other OSes, should have bought different hardware12:46
apwyep ... sadly12:46
cjwatsonwe can certainly try workarounds as long as they don't break other things12:46
apwnot 'damn i can sell this sh*t to people who'll pay for an OS and not use it, sweet'12:46
apwcjwatson, anyway thanks for your help, i learned more than i wanted to about mastering CDs12:47
apwthough one day when we arn't busy [sic] it might be nice to have the fulll incantations for the EFI ones too12:47
cjwatsonsame as before but add '-eltorito-alt-boot -efi-boot boot/grub/efi.img -no-emul-boot' before the path at the end12:48
cjwatson(ordering matters)12:48
apwyeeks thats an incation and 3/412:49
Riddellmvo: any more uploads of update-manager planned?  I just made a small change12:51
mvoRiddell: yes, one for ebox-printers12:57
ScottKmvo: I noticed in http://www.mythbuntu.org/10.10/rc that they are telling people to remove the mythstream package.  Since it's not in the archive anymore, if they upgrade using u-m, then that's done automatically, right?13:00
mvoScottK: yes, it will be suggested for removal at the end, if it causes problems during the upgrade I can make it go away earalier too13:03
ScottKmvo: I guess we should ask superm1 if it needs that.13:04
ScottK(hopefully he's an early riser and we'll know soon).13:04
mvosuperm1: I can make mythstream go away earlier in the upgrade if it has the potential to cause trouble13:04
mvoScottK: yeah, there is one other issues releated to gutenprint that I need to fix today too13:05
GoodiHi! Would anyone know when/if the sun-java6-* packages will be available for Maverick?13:16
diwicGoodi, yeah, I've also been looking for them13:23
=== MacSlow|lunch is now known as MacSlow
Goodidiwic-  I grabbed the lucid java, but that's not a real solution... There seems to very little info about it.13:24
GoodiI hope they keep it somewhere (partner repo is not optimal, but definetly better than nothing) so we don't need to start rebuilding it13:25
ScottKGoodi and diwic: It will be available in Partner13:25
Goodigreat :)13:26
Goodiscottk - any deadlines/schedules (ie quesses) when it will appear there?13:26
diwicScottK, good, do you know when?13:26
Goodiin stereo... where available :)13:27
ScottKNo, you'd have to ask a Canonical person for details.  Partner is controlled by them and not part of Ubuntu development.13:27
Goodiok. Hopefully they'll get it out before the 10.10 release :)13:28
diwicpitti, do you know who I should give a gentle push to make him/her put sun-java-* into the partner archive?13:28
cjwatsonit's in NEW, I'll process it today13:29
cjwatsonno need to push13:29
diwic\o/13:29
diwicfinally13:30
pittiah, ok; I thought archive admins aren't supposed to touch partner packages13:30
diwicpitti, who is supposed to touch partner packages? :-)13:30
* cjwatson has no compunctions about doing so13:30
cjwatsonanyway, nobody else other than archive admins can process partner packages through NEW :-P13:30
pittidiwic: I'm not sure, I think the other day there was some discussion between iamfuzz, bdmurray, and the archive admins that there needs to be some QA step before the AAs should accept it13:31
cjwatsonit's gone through source NEW already ...13:31
pittidiwic: sorry, since I'm not an active archive admin, I'd rather trust cjwatson here13:31
diwicok :-)13:31
* diwic thought pitti was all-knowing :-)13:32
* pitti ponders for a snappy answer about that being a contradiction to itself13:33
pitti"I don't know a puzzle which is so hard that I can't solve it" or something like that13:34
diwicpitti, then you can't be all-knowing, because then you would have known the snappy answer! Hah! :-)13:34
pittidiwic: q. e. d.13:34
diwicall right, let's get back to work13:35
=== ivoks-afk is now known as ivoks
mvocould someone eyeball http://paste.ubuntu.com/503887/ please?13:48
pittimvo: I think you should also export LC_MESSAGES and LC_ALL13:49
pittioh, and $LANGUAGES13:49
seb128pitti, LANGUAGE?13:50
pittimvo: right, sorry; $LANGUAGE13:51
pitti       export LANG LANGUAGE LC_MESSAGES LC_ALL13:51
* pitti sees this snippet being replicated all over the system13:51
cjwatsonI'm surprised that /etc/pam.d/cron doesn't handle this13:53
cjwatsonsession       required   pam_env.so envfile=/etc/default/locale13:53
cjwatsonperhaps only user cron jobs have a PAM session?13:53
cjwatson(if so, crontab(5) doesn't say that)13:54
mvocjwatson: I was suprised as well TBH, but I did test and LANG was not set without that13:55
mvopitti: thanks!13:55
doko_fluidsynth is another "nice" sync from experimental :-/14:01
diwicdoko_, hmm?14:05
doko_diwic: look at the recent build failures of seq24, muse, fluidsynth, zynaddsubfx, xjadeo, lashwrap, ... on armel and powerpc14:07
diwicdoko_, seems like you have a new one in queue?14:10
doko_diwic: are you involved with these multimedia updates?14:12
pittiRiddell: can we have daily CD builds again, to test the new langpacks, oversizedness, etc?14:12
diwicdoko_, I've been developing fluidsynth for a year or two14:12
diwicdoko_, and we don't have anybody upstream (or in Debian, IIRC) which tests armel14:13
diwicso armel is not supported by upstream (i e the FluidSynth community)14:14
diwicas for packaging of it, quadrispro did most of it but I helped some14:15
doko_diwic: maybe, but is this a reason to upload packages two weeks before the release which already fail to build in experimental, or show up as build failures in maverick? had this discussion yesterday with quadrispro ...14:15
doko_diwic: are the feature freeze exceptions for lash and fluidsync documented?14:17
diwicdoko_, depends on which standpoint you have. Knowing what we did to improve stability (for supported platforms) in 1.1.2, I support having 1.1.2 in Maverick,14:17
diwicdoko_, I don't know if quadrispro made any formal FFe:s.14:18
doko_diwic: fluidsync did build before, now it's not. see https://launchpad.net/ubuntu/+source/fluidsynth/1.1.2-2/+build/198239914:19
diwicdoko_, I'll see what I can do about it14:23
doko_diwic: one problem seems to be a missing b-d on libxml2-dev according to the build log14:24
doko_diwic: liblash-dev doesn't depend on libxml2-dev. don't know if devlibs:Depends is supposed to have this dependency. fixing it now directly14:32
diwicdoko_, I was just about to come to the same conclusion, but I don't know much about those ${devlibs:Depends} variables to know if it should be listed there or not14:33
=== ivoks is now known as ivoks-afk
=== almaisan-away is now known as al-maisan
pitticjwatson, dpm, Riddell: new langpacks built, they look fine to me; ready for the flood?14:53
dholbachpitti, dpm is travelling14:54
cjwatsonit's ok by me14:54
pittiI'll accept them as they flow in14:56
=== doko_ is now known as doko
=== bjf[afk] is now known as bjf
ogra_acgah, i'm screwed15:45
* ogra_ac just noticed that tar xzvf arm-rootfs.tgz /mnt/> is not such a good idea 15:50
pittiogra_ac: ? wouldn't that just try to extract "/mnt" from arm-rootfs.tgz?15:55
ogra_acno, it in fact extracted the tgz to /15:55
pitti"oops"15:56
amitkogra_ac: main devbox? ;)15:56
pittiweird that the "/mnt" parameter does that15:56
rsalvetiogra_ac: ouch15:59
ogra_achmpf16:01
=== deryck is now known as deryck[lunch]
=== amitk is now known as amitk-afk
apwanyone know if it is possible to tell apt-get to not check for or worry about disk space ?16:08
cjwatsonapw: doesn't appear to be, from a glance at the code.  LD_PRELOAD statvfs? ;-)16:09
apwcjwatson, damn ... t16:09
apwthanks :)16:10
ogra_acoh, i thought i'm screwed by killing my system disk and only being left with armel HW (and no cdrom) running ....16:10
ogra_acbut i'm screwed even more16:10
ogra_acThe following packages have unmet dependencies:16:10
ogra_ac usb-creator-common : Depends: syslinux but it is not installable16:10
ogra_acE: Broken packages16:10
ogra_acgah16:10
apwogra_ac, do i want to ask how you killed your system disk?16:11
ogra_aci said how ... above16:11
ogra_aci typoed a tar command16:11
ogra_actar xzvf arm-rootfs.tgz /mnt/>16:11
ogra_acnote the last char16:11
* apw is supprised that is not an error in the shell16:12
apwogra_ac, oh hmmm do you have network ?16:12
ogra_acthe tar ended with an error after extractinbg the whole think if that makes you happy :P16:13
cjwatsonerror in the shell> so am I.  perhaps you wewren't using bash16:13
apwhttp://kernel.ubuntu.com/~kernel-ppa/testing/16:13
ogra_aci do have network16:13
cjwatsonbusybox sh doesn't like it either16:13
=== sabayonuser_ is now known as jturek
ogra_aci surely was just using a gnome-terminal16:13
cjwatson-bash: syntax error near unexpected token `newline'16:13
cjwatsondash: Syntax error: newline unexpected16:13
apwhttp://kernel.ubuntu.com/~kernel-ppa/testing/ <-- there are .img files there which are bootable images which are DD able onto a stick16:13
cjwatsonsh: syntax error: unexpected newline16:13
ogra_acstrange16:14
cjwatsonall the shells I've tried fail16:14
apwnot sure if they have any interactive mode thought, but i suspect you can init=/bin/bash them16:14
ogra_acprobably there was a special char behind it, thats all i got from shell history16:14
xax200Excuse me, what information should I know if I want to help implement client side decoration support in a window manager?16:15
ogra_acapw, well, someone said i should be able to just dd the iso to usb key nowadays16:15
sivanghi all16:15
ogra_acand given the download is done in 2min i'll know for sure by then16:15
sivangIs KDar not in the repositories anymore?16:16
apwusb-creator does do something using stuff on your system, whether its needed ... i don't know16:16
ogra_acyes, i know16:16
sivangAFAIR it was in universe.16:16
ogra_acbut there was a mail thread recently about the possibility to dd16:16
* sivang asks in -motu16:16
apwogra_ac, well if it fails i know the .img files have been made such that they are bootable after a dd, some magic using a VM i am told to make them16:17
ogra_acok16:17
ogra_aci'll fall back to them, thanks16:17
apwbut they are special case images, so probabally need to hack them16:18
ogra_acapw, so dd didnt work (indeed) ...16:39
apwpup16:39
xax200I don't usually find that dd'ing iso images to usb works16:41
ogra_acwell, the new syslinux is supposed to make that work16:42
ogra_acbut i guess my BIOS cant read iso9660 from usb16:43
ogra_acor something like that16:43
cjwatsonogra_ac: bug16:44
cjwatsoner16:44
cjwatsonbug 52480316:44
ubottuLaunchpad bug 524803 in Ubuntu CD Images "isolinux hybrid mode should be used - all other major distributions do so since last year" [Undecided,New] https://launchpad.net/bugs/52480316:44
ogra_acah, right16:45
ogra_aci even know that one16:45
* ogra_ac is totally off track after that disaster 16:45
ogra_acerm16:47
ogra_acwhy dont i use a netboot image16:47
=== jam1 is now known as jam
ogra_acheh16:52
* ogra_ac hugs d-i16:52
=== deryck[lunch] is now known as deryck
jibelmvo, cjwatson, I reported bug 651325 against gnome-keyring during a wubi upgrade, the gnome-keyring is fixed but the unbootable problem still remains and can be reproduce.17:10
ubottuLaunchpad bug 651325 in update-manager (Ubuntu) "package gnome-keyring 2.92.92.is.2.30.3-0ubuntu1.1 failed to install/upgrade: /var/lib/dpkg/tmp.ci/preinst: 31: dpkg-maintscript-helper: not found" [Undecided,New] https://launchpad.net/bugs/65132517:11
jibelI'll file a new report.17:11
cjwatsonwas your wubi install on a drive other than C:?17:11
jibelcjwatson, no17:11
cjwatsondunno then, my wubi upgrade recently worked fine17:12
jibelA few lines are displayed very quickly when I select the Ubuntu entry at the ntloader boot menu then there is a soft reset.17:12
cjwatsonvideo them if you can17:13
jibelcjwatson, that's what I'm currently doing17:13
cjwatsondid you 'dpkg --configure -a; apt-get -f install' etc. to get all the packages back into a sane state?17:14
cjwatsonnot that it *should* be needed but ...17:15
jibelfresh 10.04.1, update/upgrade, restart, update-manager -d, restart17:16
=== al-maisan is now known as almaisan-away
cjwatsonjibel: well, I can give it a try here17:25
cjwatsonjibel: nothing special beyond that?17:25
jibelcjwatson, error: unknown command 'loadfont'17:25
jibelerror: file not found. (5 times)17:25
jibelerror: unknown terminal 'gfxterm'17:26
jibelerror: file not found17:26
jibelthen the system restarts17:26
jibelthat's all17:26
cjwatsonhm, those are a bit crap but shouldn't really be fatal17:27
jibelcjwatson, a fresh install of 10.10 on the exact same system is working correctly.17:28
cjwatsonI wonder what changed since my successful test17:28
cjwatsonadmittedly I didn't use update-manager because I wanted to force my local mirror, but that shouldn't have mattered17:29
jibelcjwatson, I can get the upgrade logs and see if I find something interesting.17:29
cjwatsonwell, I can just rerun here and see what happens17:30
cjwatsonalways optimal if a developer can reproduce locally17:31
=== ivoks-afk is now known as ivoks
shtylmansladen: I have already signed the agreement per my work with ubiquity17:39
shtylmanthat may ease some things17:39
pittiRiddell: since we now get new langpacks with empty update packs, ok if I add back some langpacks to the seeds?17:40
Riddellpitti: yes please17:40
sladenshtylman: do you now consider the Kubuntu logo  Author: Existing + Roman Shtylman   Copyright Holder: Canonical Ltd  ?17:41
pittiRiddell: I'll watch the dailies then; (are they re-enabled?)17:42
shtylmansladen: if that is what is needed then sure :)17:43
shtylmansladen: like I mentioned... I don't need to maintain copyright and it would make sense for canonical to have the copyright anyhow imho17:43
sladenshtylman: sooo, "yes" ?17:44
shtylmansladen: yes17:44
jibelcjwatson, I can't even mount the disk image, the file system is not recognized :/17:46
cjwatsonuh17:46
cjwatsonthat's fairly spectacular failure17:46
cjwatsonyou didn't run out of disk space or something?17:46
cjwatsondoes Windows still boot?17:47
Riddellpitti: yes I re-enabled them about midday today17:52
superm1mvo, it won't cause problems earlier in the upgrade, it just will cause mythtv-frontend to not execute properly later on17:53
superm1mvo, so as long as it's removed later when update-manager recommends it to be gone, people should be fine17:53
jibelcjwatson, bug 65313418:09
ubottuLaunchpad bug 653134 in Wubi "Can't boot Ubuntu after an upgrade from 10.04.1 to 10.10" [Undecided,New] https://launchpad.net/bugs/65313418:09
jibelcjwatson, windows still boot18:10
LahtexDEATH TO BAZHANG .. FUCKED UP MORON!18:11
LahtexFreenode ....18:11
jibelcjwatson, and there's 43GB free (and I didn't ask wubi to create a root disk larger than than 43GB)18:12
Lahtexcjohnston: Do something or never come back....18:12
Seeker`!ops | Lahtex18:12
ubottuLahtex: Help! Channel emergency! mneptok, Hobbsee, cjwatson, mdz, lamont, Keybuk, or thom!18:12
cjwatsonjibel: ok, I'll see what it does for me18:14
jibelcjwatson, thank you18:14
shtylmanhahaha18:15
=== rsalveti` is now known as rsalveti
ogra_accjwatson, any idea why the dd that zeroes the swap in user-setup-apply uses a blocksize of 16M ?19:23
* ogra_ac wonders if thats mandatory19:23
ScottKmvo: Apparently extras.ubuntu.com is added to sources.list of servers too, but the keyring isn't seeded there, so one gets errors when updating.19:35
cjwatsonogra_ac: I don't know, but I'm guessing it's for speed19:38
cjwatsonScottK: ouch!  It's not meant to be19:39
cjwatsonmvo: do you remember whether we decided to have apt-setup/extras default to false and be preseeded on for desktops, or default to true and be preseeded off for servers?19:40
cjwatsonmvo: I think the former might make more sense since it should only be included when ubuntu-extras-keyring is19:40
=== You're now known as ubuntulog
=== almaisan-away is now known as al-maisan
=== jjohansen is now known as jj-afk
ScottKcjwatson: I suppose I should file a bug report too.  What should I file it against?19:50
cjwatsonScottK: apt-setup (Ubuntu), and the ubuntu-cdimage project19:51
ScottKOK.19:51
=== ivoks is now known as ivoks-afk
ScottKcjwatson: Bug #65320019:56
ubottuLaunchpad bug 653200 in apt-setup (Ubuntu Maverick) "extras.ubuntu.com added to server installs" [High,New] https://launchpad.net/bugs/65320019:56
cjwatsonta19:57
=== al-maisan is now known as almaisan-away
sladenshtylman: thanks, updated, process and forwarded!20:03
mvocjwatson: I think the initial idea was exclude on server, but in the light of the seeding it makes more sense to set it to true on desktop only and default to false. will that be a problem for derivatives? or do they just inheritc from the generic "desktop"20:32
ogra_acjdstrand, around ? i'm looking for an archive admin20:36
=== jj-afk is now known as jjohansen
paissad_guys, i received this lintian warning 'missing-debian-source-format' ....  i don't know anything about quilt, i don't use it in my package ...i did not use it in the previous versions of my package & here is what i did --> echo '3.0 (native)' > debian/source/format ... do you think this good ?20:57
paissad_thanks in advance for helping20:57
cyphermoxgeser, I noticed you did a rebuild of gtkmathview because of libgdk_pixbuf2.0.la, could you review https://code.edge.launchpad.net/~mathieu-tl/ubuntu/maverick/aiksaurus/lp653217-rebuild or do the same rebuild for aiksaurus?20:57
paissad_btw, i have this other warning:20:59
paissad_W: pms-linux: executable-not-elf-or-script ./usr/share/pms-linux/resources/images/clients/musicpal.png21:00
paissad_i did in my debian/rules --> dh_fixperms /usr/share/pms-linux/resources/images/21:00
paissad_but i don't understand the reason why nothing changed !21:01
cyphermoxpaissad_, afaik, that's good for debian/source/format / missing-debian-source-format.21:01
paissad_i even did dh_fixperms only, but that did not change anything21:01
paissad_weird21:02
cyphermoxpaissad_, as far as executable-not-elf-or-script, did you copy these files from a ntfs or fat partition? this just means that file is marked executable and should not be21:02
jdstrandogra_ac: sure, what do you need?21:02
ogra_acjdstrand, three packages from NEW21:03
ScottKpaissad_: The source format one is safe to ignore.21:03
paissad_cyphermox, those files are retreive (svn export) from a subversion server .. my system is ext3 ....21:03
ogra_acjdstrand, powervr-omap3, opengles-sgx.omap3 (multiverse) and devmem221:03
paissad_ScottK, you mean 1.0 instead of 3.0 ?21:03
ScottKpaissad_: You can use whichever you prefer.  Nothing explicitly defined is 1.0, so if you want 1.0, it's not needed to add anything.21:04
jdstrandogra_ac: these all have FFes?21:04
ogra_acthe bugs are mentioned in the changelog and ScottK looked over the FFes afaik21:05
jdstrandogra_ac: k. I'll take care of it21:05
ogra_acthanks !!!21:05
ScottKjdstrand: I gave him the "If you can find an archive admin to New it" sort of FFe.21:05
ogra_acubuntu ARM loves you :)21:05
jdstrandhehe21:05
=== ivoks-afk is now known as ivoks
ogra_ac(and asac too i guess)21:05
ogra_aclinaro is eagerly waiting for that stuff ... more than us actually21:06
jdstrandso linaro loves me too?21:06
jdstrandturning into a rather nice day :)21:06
ogra_acjdstrand, yeah, the whole arm developer community will love you, these are the GL drivers for beagleboards, nobody ever shipped them in a painless way :)21:09
jdstrand\o/21:09
jdstrandsounds cool21:09
ogra_acand TI managed to release them in a redistributable way just in time to make maverick the last minute21:09
ogra_acyeah21:09
gesercyphermox: will do21:16
cyphermoxgeser, thanks!21:19
paissad_debian-watch-file-in-native-package21:21
paissad_what's a native package .. i don't really understand this ^^21:21
kklimondapaissad_: native packages are packages that debian/ubuntu is an upstream for.21:23
paissad_kklimonda, so created by debian ubuntu developers ?21:24
kklimondapaissad_: if you are packaging an upstream project you have to append a debian revision to it's version.21:24
kklimondapaissad_: yeah21:24
kklimondaits*21:24
paissad_kklimonda, i'm confused because, when i did not have the debian/source/format file, i had this warning21:25
paissad_'missing-debian-source-format21:25
paissad_and now i created this file, i have this lintian warning21:25
paissad_debian-watch-file-in-native-package21:25
paissad_^^21:25
paissad_well, i will remove the watch file then21:25
paissad_and this too -> native-package-with-dash-version21:26
paissad_really complicated & confusing the 'debian rules for packaging' .. that changes too often21:27
kklimondawhat is the version of your package?21:27
kklimondayou can also run lintian with --info for a long description of this errors/warnings21:27
paissad_kklimonda, pms-linux-1.20.409+svn41121:27
paissad_kklimonda, in the changelog , i have -> 1.20.409+svn411-121:28
paissad_kklimonda, i did lintian -iIEmv --color always --pedantic21:28
paissad_so i have explanations21:29
kklimondapaissad_: you should have 3.0 (quilt) in debian/source/format and an .orig.tar.gz in the parent directory.21:29
paissad_kklimonda, actually, i did (native) instead of (quilt) ...21:29
paissad_i will try quilt then21:30
jdstrandogra_ac: I'm not keen on this udev rule in opengles-sgx-omap3: KERNEL=="pvrsrvkm", GROUP="users", MODE="0666"21:31
ogra_acrsalveti, ^^^^21:32
ogra_acjdstrand, what would you like 0660 ?21:32
jdstrandrsalveti, ogra_ac: is that required? wouldn't KERNEL=="pvrsrvkm", GROUP="video", MODE="0660" be better?21:32
ogra_acdo we actually still have a video group ?21:33
jdstrandogra_ac: I would, yes, and a group that makes sense21:33
rsalvetiif the video group is still working, np to use it21:33
jdstrandvideo:x:44:21:33
ogra_aci guess ricardo took rthat over from an upstream doc or from openembedded21:33
jdstrandwell, I am not up to date on the proper group21:33
jdstrandusers seemed weird21:33
rsalvetiI couldn't find a proper group to allow a normal user to use the 3d acceleration21:34
ogra_acbut that means tinkering with groups21:34
rsalvetididn't think that video would fit that21:34
ogra_acyeah, video is rather for video devices21:34
kirklandslangasek: hey, upstart question for you ....21:35
jdstrandhmm21:35
jdstrandkees: ^21:35
jdstrandwe should probably look at what nvidia does21:35
kirklandslangasek: so asking upstart for the "status" on a given process really just checks if the process is still running, right?21:35
ogra_acyeah21:35
rsalvetido we have a /dev for nvidia?21:35
jdstrandI thought we did21:35
* jdstrand doesn't have nvidia hardware21:35
* rsalveti neither21:36
kklimondayes, we do21:36
kirklandslangasek: i need to add some additional smarts, beyond whether the process is running, i need to check if it's actually doing something21:36
kirklandslangasek: ie, that it's not wedged in a while-true loop21:36
kirklandslangasek: what's the best way to do that?21:36
jdstrandI'll try to pull the sources and find it21:36
* kees reads21:36
kirklandslangasek: in sysvinit script, the status action can be anything, in its own stanza in the case statement21:36
kirklandslangasek: how could i get something similar out of upstart?21:36
jdstrandkees: looking for the proper group for a udev rule for 3d graphics21:36
ogra_acsigh21:37
jdstrandkees: it was 0666 with group 'users', I thought 0660 with something else, but am not up on groups for 3d21:37
ogra_acthe nvidia source is 50 meg21:37
keesjdstrand: the regular dri uses "video"21:37
* ogra_ac twiddles thumbs21:37
jdstrandah21:37
jdstrandI thought it did21:37
keesjdstrand: but gdm start-up set facls on console users too21:37
ogra_acah21:37
keescrw-rw----+ 1 root video 226, 0 2010-09-30 22:28 /dev/dri/card021:37
kenvandineev, ping21:37
ogra_acthen 0660 video it is21:37
rsalvetihm, then video it is21:38
rsalveti:-)21:38
jdstrandkees: thanks21:38
keesjdstrand: sure thing21:38
jdstrand$ ls -l /dev | grep video21:38
jdstrandcrw-rw----  1 root video    10, 175 2010-10-01 10:35 agpgart21:38
keesit still might need facl-style stuff though21:38
jdstrandtoo easy21:38
ogra_aclol21:38
kees$ getfacl /dev/dri/card0 | grep kees21:39
keesuser:kees:rw-21:39
keesgdm or somebody does that, since I don't think the default users is in "video" any more...21:39
kees*user21:39
ogra_acright21:39
jdstrandah, seems not21:39
jdstrand(I am not)21:39
ogra_acthats why i said above it would require tinkering with groups21:40
keesprobably best to ask RAOF about it21:40
kees<unhelpful>"users" 0666 is certainly wrong though. :P21:40
jdstrand0666 makes me pretty uncomfortable though21:40
trijntjeis there a way to see in LP who imported a package (and translations) from upstream into LP?21:43
slangasekkirkland: 'status' returns the information upstart has about the status of the job; if you need some other kind of status check, I would recommend a standalone tool similar to apachectl21:44
ogra_acjdstrand, uploading ubuntu221:44
kirklandslangasek: :-/21:44
slangasekkirkland: not in an init script, because that will collide with all of the auto-mapping support between sysvinit and upstart21:45
slangasekkirkland: why do you have a server that can get wedged? :)21:45
kirklandslangasek: hmm, well, hard to answer that question in a politically correct manner21:45
slangasekkirkland: does this status check need to be integrated with something like heartbeat, though?21:46
kirklandslangasek: in any case, is it safe to say that moving a sysvinit script to upstart loses the ability to do a "smart" status check?21:46
kirklandslangasek: yes, something like heartbeat/pacemaker/etc21:46
kirklandslangasek: okay, so let's say I add something to service(8)21:47
kirklandslangasek: perhaps, "status+"21:47
slangasekkirkland: "safe to say" - I wouldn't say it that way because it casts it as an init system problem instead of a service problem, which is what I think bears the responsibility for not getting itself wedged21:47
kirklandslangasek: service <FOO> status+21:47
jdstrandogra_ac: accepted21:56
jdstrandkees: thanks again21:56
keesjdstrand: np21:56
ogra_acjdstrand, awesome21:57
ogra_acthzanks !21:57
andreserlkirkland, just for clarification, but prolly you already know this, pacemaker has various ways of monitoring a service. One, would be by the use of sysvinit scripts, using the return codes of the scripts. Second, using upstart status output, which was recently added by Ante, Third, using OCF scripts, that what they basically do, is to monitor the application itself, however this is customizable  as long as you follow the standard for creating21:57
andreserlsuch scripts :)21:57
jdstrandhehe21:58
rsalvetijdstrand: cool, thanks :-)22:07
jdstrandrsalveti: sure :)22:08
paissad_guys, here is my debian/rules http://pastebin.com/LCb0Ggb9, there is dh_fixperms into it ... but i don't understand the reason why i get this warning with lintian -->executable-not-elf-or-script <-- AND indeed those files have their permissions not changed in the package22:25
paissad_where am i wrong ?22:25
paissad_i'm really confused .. i even added 'dh_fixperms' in other targets of the debian/rules file ... but nothing changed22:26
paissad_the build finished successfully ... but i have 3 warnings about executable files in /usr/share/pms-linux/ ....22:27
cjwatsondh_fixperms isn't entirely magic - if something is making the file executable, shove chmod -x in there22:29
paissad_cjwatson, i'm quite sure that dh_fixperms failed ^^22:30
paissad_i even can say i'm certain of this !22:30
cjwatsonno22:30
paissad_cjohnston, it did not change the perms22:30
cjwatsonread its documentation22:31
cjwatsonit's not supposed to change *everything* - it fixes *certain problems*22:31
paissad_cjohnston, i did read the manpage already (before posting)22:32
paissad_it's a minimum :)22:32
cjwatsonI am not cjohnston22:32
paissad_oh sorry !22:32
cjwatsonthe manual page does not say anywhere that it makes files under /usr/share non-executable22:32
cjwatsonexcept doc22:32
cjwatsonit's very specific22:32
cjwatsonjust insert chmod -x commands22:33
cjwatsonyou can also run with DH_VERBOSE=1 set in the environment to see exactly what debhelper commands are doing22:33
paissad_cjwatson, i thought about the chmod -x command ... but i just wanted to know the reason why dh_fixperms did not work in my case22:33
cjwatsonbecause it can't realistically guess everything - files in /usr/share/PACKAGE are often meant to be executable22:34
paissad_cjwatson, ok22:35
ari-tczewcjwatson: are you a Debian Developer?22:39
cjwatsonari-tczew: yes, have been for ten years, why?22:41
cjwatson(well, thinking about it, that's exaggerating by a few months)22:42
* maco rolls eyes22:42
ari-tczewcjwatson: I think whether you can apply this patch in Debian: debian bug 50608322:43
ubottuDebian bug 506083 in palo "palo: open(O_CREAT) needs a mode argument" [Normal,Open] http://bugs.debian.org/50608322:43
cjwatsonDebian doesn't normally work like that.  You need to ask its package maintainer22:43
cjwatsonoh, it's QA22:44
cjwatsonyeah, guess I could, not tonight though22:44
ari-tczewcjwatson: would be nice if you add this one to your to-do ;)22:44
cjwatsonit had a proper maintainer when I filed that bug22:44
kirklandRoAkSoAx: are you here?22:52
RoAkSoAxkirkland: yes22:52
kirklandRoAkSoAx: i know about those 3 methods...  there's just a "deficiency" we perceive in upstart's ability to test the real actual status of a service22:53
kirklandslangasek: okay, here's what I'm thinking ...  adding something to the service(8) command;  perhaps call it "status+", which first runs the status from sysvinit or upstart, but then executes (if available) something in /etc/status/$NAME, which is able to do "smarter" or "deeper" testing of the practical status of the service22:55
kirklandslangasek: for something like lighttpd, it's great that there's a pid for the process, but can wget retrieve a webpage from it22:55
kirklandslangasek: thus, you can define the real status+ of a service with something in /etc/status/lighttpd (as an example)22:56
slangasekkirkland: why under /etc/status instead of, say, /usr/lib/service/status/ ?22:56
slangasekputting it under /etc implies packaging overhead that seems avoidable here22:57
kirklandslangasek: sure -- that's fine22:57
RoAkSoAxkirkland: "pid for the process, but can wget retrieve a webpage from it" > that's exactly how the "monitor" option in a cluster would work, instead of just verifying if the service is running or not, which doesn't mean that it is receiving request in your example22:57
kirklandslangasek: i was thinking of symlinks in /etc, to stuff in /usr/lib, actually22:57
kirklandslangasek: that someone could break, or modify to their liking, if necessary22:57
=== yofel_ is now known as yofel
slangasekkirkland: I wouldn't introduce the complexity of /etc/status until there was a demonstrable demand for it22:58
kirklandslangasek: ack, understood22:58
kirklandslangasek: what do you think of the nomenclature?  "service <FOO> status+" ?22:58
slangasekkirkland: otherwise, that seems ok - it'd be nice if this wasn't needed, using something like what RoAkSoAx suggests in its place, but I don't feel strongly about this22:58
slangasekkirkland: sees ok22:59
slangasekseems22:59
kirklandRoAkSoAx: okay, so how is your way done generically, for anything started by upstart or sysvinit?22:59
RoAkSoAxkirkland: from the cluster point, generically, it is just monitoring the return codes of sysvvinit, for start, stop, and status, and something similar is done for upstart23:01
RoAkSoAxbut for examplke, if using OCF resource agents, we could monitor if the webserver is respoding to request23:01
RoAkSoAxfor ex, from the apache OCF resource agent: " monitor:  return TRUE if the web server appears to be working. For this to be supported you must configure mod_status and give it a server-status URL.  You have to have installed either curl or wget for this to work."23:01
kirklandRoAkSoAx: right... okay;  here's the problem -- upstart's "status" is NOT sufficient (ie, it's not "smart" enough) for the test we need to perform23:01
ebroderHmm...how do I make mountall interact nicely with encrypted swap? Right now I have a line in /etc/crypttab, and a line in /etc/fstab, but cryptsetup hasn't run before mountall does, so I get the "unable to find <my swap device>" error23:02
RoAkSoAxkirkland: if you need it for monitoring within an HA cluster, I'd suggest creating OCF resource agents23:02
RoAkSoAxkirkland: or providing the ability to upstart to do something similar, or either a tool that could do just that, monitor something specific23:03
ebroderOh wait...maybe I screwed up my fstab23:03
kirklandRoAkSoAx: okay23:03
RoAkSoAxkirkland: is this for HA clustering?23:04
slangasekebroder: yes, sounds like a busticated fstab; for encrypted swap, cryptsetup should run /well/ before mountall, i.e., in the initramfs23:05
ebroderslangasek: Yeah, I left out the second column in my swap line. My bad. Although I thought cryptsetup only ran for the root device in the initrd23:13
RoAkSoAxkirkland: if you need to take a look to some OCF resource agents, they are in "cluster-agents", inside heartbeat folder. Pacemaker also provides other ocf resource agents I believe23:14
slangasekebroder: root device and swap, because you might hibernate/resume to swap23:14
slangasekto encrypted swap, I mean23:14
ebroderslangasek: Huh. That's clever23:18
paissad_damn it, i have to build a package for amd64 and for i386 archs, but i don't have currently an i38623:37
paissad_^^23:38
paissad_is there a way to build a package for i386 from an amd64 arch ?23:44
Davieypaissad_: yes.. you can use amd64 to build i386, but tricky to do the other way around23:47
Davieypaissad_:  if you use pbuilder-dist.....  $ pbuilder-dist maverick i386 create   -> then $ pbuilder-dist maverick i386 foo.dsc23:49
YokoZarArneGoetje: ttf-baekmuk and xfonts-baekmuk look like they contain the same thing -- should the latter package be dropped?23:52

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