=== robbiew is now known as robbiew-afk [00:23] mdke: eww, right, we must strip gnome help files, otherwise we'll install the files in both places; sorry, I'm afraid we need to fix the symlinking instead [00:23] slangasek: ^ [00:23] does anyone know what other distros use apport, if any? [00:24] PovAddict: OpenSUSE does [00:25] pitti: does that mean the solution is known and in progress? [00:26] slangasek: not yet, I'm afraid, it basically turned up yesterday [00:28] slangasek: that's the cdbs task on that bug [00:32] kees - are you interested in working on bug 446395? [00:32] Launchpad bug 446395 in gnome-screensaver "Screen lock unlocks after 5 wrong attempts" [Critical,Triaged] https://launchpad.net/bugs/446395 [00:32] i'll work on that tomorrow if nobody else has fixed it in the meantime [00:34] news about the loop module? still =y? [00:35] mdke: yes, it will need a langpack rebuild to pick up the new gnome help files [00:35] slangasek, mdke: got a DSL reconnect (I really ought to sleep at this time..), did you answer? [00:35] slangasek: hm, the .deb is 450 kB; what do you mean exactly with "baloon"? [00:36] pitti: apt says upgrading ubuntu-docs here will take up 64MB more space [00:36] mdke: ignore the "we must fix symlinking" bit, I was confused; fdupes is unrelated to translations [00:36] ?!? [00:37] wow, 446395 looks serious [00:37] pitti: installed-size: 169084 vs. 231408 [00:38] * pitti looks at debdiff ubuntu-docs_9.10.9_all.deb ubuntu-docs_9.10.10_all.deb [00:38] pitti: hah - something is calculating installed-size *before* pruning files from the tree (!) [00:38] (AFAICT) [00:38] hmm, no [00:39] slangasek: it seems the new version replaced all the relative symlinks with absolute ones [00:39] pitti: which shouldn't change the package size at all, really [00:39] well, slightly [00:39] but certainly not in the MB range [00:39] not much after compression [00:40] how many symlinks are we talking about? [00:40] PovAddict: only if you count on your screensaver for your security, luckly [00:41] pitti: right, in both cases the installed-size is way overestimated - the actual unpacked size is 17-19MB, the installed-size figures are > 160MB. I think it's being calculated before the tree is trimmed [00:41] PovAddict: 5270 [00:41] pitti: relative and absolute symlinks should still take up exactly one block on disk, surely? [00:41] oh true, blocks... yes, it should take the same space [00:41] slangasek: right; I just purged and reinstalled ubuntu-docs, and df says before/after is a difference of 20.9 MB [00:41] unless you have a crazy-long path somewhere [00:42] slangasek: uncompressed yes, just the .deb gets a tad bigger (not much presumably) [00:42] pitti: right - apt is reporting the installed size [00:42] ok, so I think this is a red herring [00:42] which shouldn't differ for absolute vs. relative symlinks [00:42] pitti: yeah; it's a bug in its own right though [00:42] I'll report it [00:43] * pitti -> bed, really this time [00:43] 'night! [00:44] night pitti! [01:01] directhex: gasp do I see banshee tagged? [01:05] ArneGoetje, pitti: was language-support-extra-eu meant to be removed from the archive? (No rdeps, depends on NBS myspell-eu-es) [01:06] jdong, 1.5.1? yes [01:06] directhex: indeed [01:07] directhex: heh a friend and I are now whining over how banshee psychically stole our smart playlist shuffler XD [01:07] directhex: we chose to build a student's t-distribution modeling each song's preference score though ;-) [01:11] jdong, the code in libmono-system-corlib2.0-cil which steals the user's brainwaves also serves to send all user-made extensions upstream silently [01:12] directhex: I KNEW IT! THEYVE BEEN WARNING US FOR YEARS ABOUT THIS! [01:12] * jdong registers boycott(2n+1)novell.com [01:14] jdong, you want to boycott boycott-boycottnovell? [01:15] :) [01:15] well the site for n <=2 has been taken, hasn't it? ;-) [01:17] dunno dude, i'm just a packager [01:49] slangasek: all language-support-extra-* packages got removed from the archive. Instead of myspell-eu-es, hunspell-eu-es gets installed with the language-support-writing-eu package. [01:49] ArneGoetje: well, language-support-extra-eu didn't get removed :) [01:50] ArneGoetje: is there a bug # I can reference for a removal request? [01:50] anyone having trouble upgrading gdm? failing in configure scripts [01:51] ArneGoetje: in fact, I see 11 language-support-extra-* in karmic presently [01:51] slangasek: eh? I thought pitti removed them already... [01:53] ArneGoetje: apparently not - could I ask you to file a quick bug report requesting their removal and explaining (briefly) why they should go away, and I can knock them out of the archive? [01:53] slangasek: ok [01:53] slangasek: against "ubuntu"? [01:54] ArneGoetje: against one of the source packages [01:54] slangasek: ok [01:55] nm, figured it out. gdm doesn't like being upgraded as root [01:55] needs to be done as sudo'd user [01:57] spotter: um [01:58] spotter: that's not something gdm is entitled to have an opinion on; could you please file a bug report (with error messages)? [01:59] the GDM upgrade issue is already reported isn't it? (the failures due to the su calls in the postinst script) [01:59] ah, ok [01:59] slangasek, bugs already filed [01:59] (is the bug targeted to karmic?) [01:59] not by me [01:59] by many others [01:59] I just commented on one [01:59] slangasek - yeah, that's the one [02:00] its been this way for a few weeks [02:00] been ignoring it as had other things, and laptop still worked [02:00] what bug? [02:00] here's the bug I commented on https://bugs.launchpad.net/bugs/438789 [02:00] Launchpad bug 438789 in gdm "package gdm 2.28.0-0ubuntu8 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,Confirmed] [02:00] might be a duplicate of others [02:01] that's a slightly different issue but it's the same root cause [02:01] ah; looks like that should be marked as a duplicate of 438561? [02:01] the stuff in the postinst script causes a gvfs daemon to be spawned in the GDM session [02:02] slangasek - yeah, you're right [02:02] ok [02:02] thanks for confirming :) [02:02] i need to speak to seb128 tomorrow and see what he wants to do with this one [02:03] i've got an idea for fixing it but it involves a non-trivial patch to g-s-d [02:04] btw, I didn' (and don't currently) have a .gvfs file in my /var/lib/gdm directory [02:04] nto before it refused to configure and not after I got it configured w/ sudo [02:05] spotter - it's still most likely a related issue. the su calls seem to cause multiple problems at the moment [02:07] ok [02:10] slangasek: https://bugs.edge.launchpad.net/ubuntu/+source/language-support-extra-de/+bug/451802 [02:10] Launchpad bug 451802 in language-support-extra-de "[karmic] please remove from archive" [Undecided,New] [02:22] ArneGoetje: thanks! [02:42] slangasek: i think m2crypto is ok without the testsuite enabled === Whoopie_ is now known as Whoopie === akgraner_ is now known as akgraner [04:09] cjwatson: seems the installer still doesn't know about the Chinese package split... it still installs the -zh metapackages, instead of only -zh-hans or -zh-hant for Simplified and Traditional Chinese respectively. Against which package should I file the bug? pkgsel? [04:26] cjwatson: ah, ok. the latest daily Live CD doesn't have the latest ubiquity and language-selector yet... updating and retry... [04:26] * hyperair notes the wonderful clicking sounds his soundcard makes when resuming from powersave [04:29] zul: well, that's for the MIR team to say [04:31] * hyperair pokes dtchen [05:05] cjwatson: good news! The Edubuntu DVD seems to be working again! Thanks so much, you rock [05:22] does root-on-dmcrypt result in the root partition not being unmounted properly? [05:24] oh great, now my /etc/init/gdm.conf went missing. wtf? [05:25] lolwut now encoding problems [05:25] win [05:35] hyperair: gdm.conf is no longer used [05:35] i meant /etc/init/gdm.conf [05:40] oh [05:40] ext4's driving me up the wall [05:40] even after a clean reboot, fsck complained endlessly about all kinds of things, and whacked up /var/lib/dpkg [05:41] ugh === YDdraigGoch is now known as Richie [05:43] gasp ext4 loses tons and tons of data in the writeback cache? :) [05:46] jdong: why, is it a known issue? [05:46] well i could probably blame myself for running an rc kernel, but i didn't think things would get *this* bad with a filesystem that's supposedly stable [05:51] TheMuso: ubuntustudio-desktop depends on gstreamer0.10-schroedinger, but that package is NBS (and I'm removing it) - can you update ubuntustudio-meta? [05:56] slangasek: IIRC it moved to a different gstreamer package and we didn't sync the one it moved to. [05:57] * ScottK recalls no details though. [05:57] slangasek: RIghto thanks [05:57] ScottK: hmm, looking [05:58] It's also 1AM here and I'm very tired. I could be totally full of crap. [05:58] Just so you're warned. [05:59] * TheMuso waits to see if further updates are necessary [06:00] ScottK: right, confirmed; the changelog said it moved to -bad, and I took it at face value [06:01] TheMuso: ^^ so please hold off wrt uploading, at least === dendrobates is now known as dendro-afk [06:02] slangasek: Right, I agree, since I have bad installed and /usr/lib/gstreamer-0.10/libgstschro.so is not in bad [06:02] * slangasek nods [06:02] where does upstart's umount code go? [06:02] or does it not unmount filesystems at all? [06:03] hyperair: scripts are still in /etc/rc[06].d/ === porthose is now known as porthose|afk [06:03] ah i see [06:05] ScottK: has anyone taken a position wrt the right way to fix this for karmic, that you're aware of? [06:06] I see bug #436350 [06:06] Launchpad bug 436350 in gst-plugins-bad0.10 "Please Upgrade Gstreamer Bad and Multiverse Plug-ins To 0.10.14" [Undecided,Confirmed] https://launchpad.net/bugs/436350 [06:06] slangasek: Not that I know of. [06:06] and I don't see slomo on IRC :/ [06:24] slangasek: ok meta is ready fr upload, waiting for your call in the event something needs changing [06:26] TheMuso: ok; everything I see here currently tells me we should drop the dep from the meta packages and pull in a new upstream version of plugins-bad, because the current upstream version of schroedinger doesn't even have the plugin in the source anymore [06:26] TheMuso: I would advise you to go ahead with uploading [06:27] slangasek: ok [06:34] slangasek: you're wanted. [06:34] er whoops [06:34] slomo: [06:34] works well enough, though ;) [06:34] haha [06:35] slomo: hi, you seem to have synced the new upstream version of schroedinger which drops the gst plugin, but there doesn't seem to be any work done on getting the new upstream version of gst-plugins-bad in that takes it over? [06:39] hmm it seems that banshee hasn't been seeing karma support for quite some time [06:42] slangasek: blame seb128, i asked him to sync gst-plugins-bad0.10 0.10.14 more than a month ago :) unfortunately i didn't notice that he forgot to do it [06:42] slangasek: sorry [06:45] what was the tool used to do reverse build dep? [06:46] hyperair: apt-cache rdepends? [06:46] that's r *binary* dep [06:46] hyperair: grep-dctrl on apt's Sources files [06:46] i'd like reverse build-dep [06:46] ah ok [06:46] slomo: ah right thanks [06:47] hyperair: good work on updating banshee btw :) does it already use the external hyena? [06:47] hmm so it seems libkarma-cil only has one reverse build-dep [06:47] slomo: no it doesn't. that's still stuck in binary new [06:48] slomo: also, i don't see anything in the build system that looks for hyena [06:48] so i guess not [06:49] hyperair: reverse-build-depends script in ubuntu-dev-tools wraps grep-dctrl and makes it easy. [06:50] ah cool [06:50] thanks [07:00] slangasek: sorry, I don't know what has caused that, the deb is the same size [07:02] whose brilliant idea was it to make https://launchpad.net/ubuntu/+source//+filebug point to a wiki page instead? [07:02] >=( [07:02] i know what information i want to include and don't need apport damnit [07:02] * hyperair grumbles [07:06] hyperair: #ubuntu-bugs is the best place to provide feedback on that decision. [07:07] alright, i'll do that when i figure which channel to leave [07:19] ArneGoetje: probably a bug with tasks on pkgsel and ubiquity, yes; although I'm afraid we may not have time to get that dealt with this cycle === thekorn_ is now known as thekorn [07:48] shite, my firefox-3.5 crashes 100% on start now, even when removing the profile directory.... [07:48] amd64 [07:49] asac: you signed of the latest release a couple of hours ago. Can that be related? [07:54] reboot helped with the firefox crash issue, sorry for bothering [07:57] good morning === tkamppeter_ is now known as tkamppeter === thekorn_ is now known as thekorn [08:20] slomo: ah - ok, I'm following up on bug #436350, then [08:20] Launchpad bug 436350 in gst-plugins-bad0.10 "Please Upgrade Gstreamer Bad and Multiverse Plug-ins To 0.10.14" [Medium,Confirmed] https://launchpad.net/bugs/436350 [08:20] mdke: bug #451764 filed, with analysis [08:20] Launchpad bug 451764 in ubuntu-docs "Installed-Size header in ubuntu-docs is broken" [Low,Triaged] https://launchpad.net/bugs/451764 === slangasek changed the topic of #ubuntu-devel to: Archive: Final Freeze! | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-jaunty | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs === ion_ is now known as ion [08:52] Hobbsee: What do you think about bug 359934? Is this ok with you and can we get this into karmic? [08:52] Launchpad bug 359934 in ubuntu-restricted-extras "Java is not included in ubuntu-restricted-extras" [Undecided,Confirmed] https://launchpad.net/bugs/359934 [08:57] doh [08:57] slangasek already froze the archive [08:57] I was expecting an extra day to get work done ;-) [08:57] seb128: I work European hours on Thursdays now ;) [08:58] slangasek, still I was expecting freeze at the end of the day [08:58] seb128: I'll mail the press letting them know that we'll delay by a day, so we can get more new gnome crack! [08:58] heh [08:59] * dholbach hugs seb128 [08:59] dholbach, we plan to get GNOME 2.28.1 anyway as every cycle ;-) [08:59] * seb128 hugs dholbach [08:59] which is on monday next week [08:59] seb128: if you need help with sponsoring, let me know [08:59] dholbach, would be nice thanks! [08:59] can't let my old desktop buddy down :) [09:03] ;-) [09:03] simon-o: ie, reverting version 29? This seems to be a policy decision that keeps switching, which is why i've been trying to ignore it. mvo: your thoughts? [09:04] mvo: 451428> hrm, that's not a bug in grub, that's a bug in whatever called grub to be installed... [09:06] slangasek: so you think he is installing kernels from some location that is not ubuntu? [09:06] Hobbsee, mvo: Yes it is. Right now the java plugin is installed on amd64 but not on i386 which is inconsistent. I would like to keep the plugin in u-r-e until the open source plugin is stable. [09:06] mvo: no, I think something removed grub-pc from his system and replaced it with grub on upgrade [09:07] simon-o: it seems like pitt did this change in 29 [09:09] slangasek: aha, I misread the bug then, that makes sense. I will ask for more info [09:09] mvo: already done, just wanted to get a heads-up since you see these bugs before I do [09:09] mvo, pitti: yes he did. But it seems like icedtea6-plugin isn't stable enough yet, what do you think pitti? [09:09] s/get a/give you a/ [09:09] slangasek: I had one report about a jaunty grub2 being removed on upgrade, but was not able to reproduce yet as my VM did not boot with grub2 in jaunty [09:09] slangasek: thanks [09:10] right; there've been a few reports from people saying grub-pc went away in favor of grub, I /think/ these are all actually fallout from old (fixed) bugs but am chasing them up to be sure [09:21] slangasek: of course there are still some cases where we need to use grub rather than grub2, e.g. dmraid [09:21] cjwatson: sure - those aren't the people that have been filing bugs, though :) === dholbach_ is now known as dholbach [09:34] Good morning [09:35] <\sh> moins pitti [09:35] ArneGoetje, slangasek: hm, I wasn't aware of -extra; I just removed language-support-translations* [09:35] good morning pitti [09:35] pitti: ah - well, they're knocked out now [09:36] good, thanks [09:36] simon-o: icedea-plugin? what? [09:40] pitti: sorry, I didn't provide any context. We were talking about bug 359934 and if u-r-e should contain sun-java6-plugin or not (currently this is the case on amd64 but not on i386). icedea-plugin is maybe not stable enough to replace the sun plugin [09:40] Launchpad bug 359934 in ubuntu-restricted-extras "Java is not included in ubuntu-restricted-extras" [Undecided,Confirmed] https://launchpad.net/bugs/359934 [09:45] simon-o: hm, I'm afraid I have no idea how stable it is; but given that we're about to remove sun-java6, I'd rather point people to openjdk TBH [09:45] doko might have a better opinion [09:49] pitti: I think the wish to include the sun plugin came because of bug 344705 [09:49] Launchpad bug 344705 in openjdk-6 "IcedTea Plugin Doesnt Work" [High,Confirmed] https://launchpad.net/bugs/344705 [09:50] Keybuk: could you please give me a hand with bug #451556 ? [09:50] Launchpad bug 451556 in update-manager "fails to reboot during recent 9.04 -> 9.10 upgrades" [High,Confirmed] https://launchpad.net/bugs/451556 [09:50] Keybuk: it seems to be upstart releated [10:01] cjwatson: ok... [10:02] cjwatson: BTW: Evan's libwebkit-1.0-2 fixes the crash issue for me. [10:03] ArneGoetje: can you post a comment to that effect on 451980? [10:04] evand: ah, you are online... :) yes, I will post a comment [10:05] ArneGoetje: thanks! [10:08] pitti: is the fact that bug 401381 didn't pick up the core dumps for grub a bug in apport, or more fundamental to how apport works? If it's the latter, do you see any use for rules that say something like, "if ubiquity crashed, and the following processes also have crash dumps, include them in ubiquity's crash report"? [10:08] Launchpad bug 401381 in ubiquity "ubiquity crashed with InstallStepError in configure_bootloader()" [Undecided,New] https://launchpad.net/bugs/401381 [10:09] (the grub crash appears near the bottom of the syslog) [10:24] evand: that bug is against ubiquity with a python traceback; I don't quite understand why it should have a core dump? [10:24] Jul 19 13:00:50 ubuntu grub-installer: Floating point exception (core dumped) [10:24] Jul 19 13:00:56 ubuntu last message repeated 4 times [10:25] (from UbiquitySyslog) [10:25] evand: well, of course the package hook could do some magic to pick up the related .crash reports, but if you just attach them you wouldn't get auto-duplication and retracing [10:26] evand: the .grub crash ideally gets submitted to LP separately, and gets retracing/duplication love [10:30] Perhaps it could offer to file both at once (so there's a greater chance the user will actually do it), and link the two together with a comment? [10:30] evand: FYI, just updated bug 451263; unblocked now [10:30] Launchpad bug 451263 in ubuntu-artwork "Ubuntu logo is not redistributable" [Critical,In progress] https://launchpad.net/bugs/451263 [10:30] hooray [10:30] evand: would be nice indeed; unfortunately right now we have no control about the actual filing process in the browser; [10:31] pitti: ah, unfortunate. Do you want a wishlist bug for this anyway? [10:32] evand: sure; don't hold your breath for it, though, it requires large changes, also in LP (be able to file bugs through launchpadlib, replicate the bug filing interface in GTK/KDE, etc.) [10:32] and looking for duplicates, etc. [10:34] understood [10:44] bryce_: something strange in the diff of your latest -evdev upload; a patch to configure.ac to make it think it's 2.2.2 instead of 2.2.5? (+ some undocumented source changes) [10:45] bryce_: I'm looking at this to fix the FTBFS, but the other stuff looks notable [10:45] slangasek, hrm, I'll doublecheck [10:46] hmm, didn't show up in my debdiff [10:49] bryce_: well, somehow you also skipped -1ubuntu3 (or rather, seem to have renumbered -1ubuntu2 to be -1ubuntu3 (!)), the previous upload was -1ubuntu2 - all very confusing [10:49] uploading -1ubuntu4 in the meantime for the build fix [10:50] oh crap, I know what I did [10:51] slangasek, let me do the -1ubuntu4 if you've not done it already [10:51] bryce_: ok, go ahead [10:54] slangasek: can I upload a fix for nvidia-common (bug #292606)? It should fix a rather nasty problem with dist-upgrades [10:54] Launchpad bug 292606 in nvidia-common "dkms - error when installing custom kernel" [Medium,New] https://launchpad.net/bugs/292606 [10:54] tseliot: yes, no need to ask before uploading :) [10:55] slangasek: ok, thanks [10:56] slangasek, can I re-upload fixed -1ubuntu4 or does it need to be -1ubuntu5? [10:56] bryce_: I haven't uploaded a -1ubuntu4 yet, just use that number [11:00] slangasek: I just pushed a new partman-uboot to fix the initial implementation of the mkfs.ext2 call in there; Marvell had told me in the past that -I 128 should be used (not -b 4096), and they just confirmed in https://bugs.launchpad.net/bugs/450906 [11:00] Launchpad bug 450906 in adana "u-boot is not tolerant of unclean ext2 file systems" [High,Incomplete] [11:01] slangasek, fixed package uploaded [11:04] bryce_: cheers [11:04] lool: will see it as I process the queue, I'm sure :) [11:06] slangasek: Oh yeah sorry; I guess you should poke if you have any question rather than me poking you whenever I upload, sorry [11:07] lool: no worries, and if you have anything that's you need pushed through the queue urgently, just shout [11:08] Ok I dont [11:08] Thanks [11:08] slangasek, I got an error trying to upload -1ubuntu4 [11:09] oh? [11:09] er, wait [11:09] lool: where is mutter-moblin? is it still sitting in some queue? [11:09] lool: the other sync requests build-depend on libmoblin-applet-dev or whatever it's called [11:09] bryce_: sorry, -1ubuntu4 was the one you already uploaded? Sorry, I was talking nonsense when I told you to reuse that number [11:09] ahh ok [11:10] bryce_: -1ubuntu4 was already in the archive, so you have to increment; was getting my wires crossed [11:10] slangasek, ok -1ubuntu5 uploaded [11:13] bryce_: thanks [11:30] dholbach: mutter-moblin is in Debian; I thought paulliu would request a sync for it [11:31] paulliu: Hey ^ [11:31] oh, I thought it was sitting in some queue already === porthose|afk is now known as porthose [11:39] paulliu: usr/share/mutter-moblin/theme/moblin-panel-myzone is empty (not sure if that's empty) and src/mnb-scale-group.* are LGPL-2 (not sure if that's an upstream oversight) [11:39] other than that the package looks good to go [11:39] I need some help with brainstorm [11:39] lool: ^ [11:39] dholbach: thanks for the review [11:40] no worries [11:40] I updated http://brainstorm.ubuntu.com/idea/21684/ but would like to subscribe to it [11:40] Also, I'm not sure of the good way to link a brainstorm idea with a bug report [11:40] I just put a link in the solution description which is not ideal [11:40] lool: jcastro will know, but he's not awake yet, I guess [11:41] dholbach: thanks [11:41] and ndeschildre does not seem to be around either [11:45] Keybuk, moo ... [11:47] Keybuk, why did you move the framebuffer hack in usplash into the C code ? i would like to change the script to use fb1 instead of fb0 (my external monitor with way higher resolution) moving the code into C means i need to recompile the whole thing [12:07] ogra: i dont think the language of the implementation should matter; the issue is rather with usplash not handling multiple screens which is well known [12:08] lool: urm. you're supporting the hardcoding of configuration options? (!) [12:09] lool: hardcoding a setting that might need changing is different from simply ensuring it has a good default [12:10] sladen: fb0 was not configurable either [12:10] it was just easier to change in the shellscript ... it wasnt meant to be configurable at all though [12:11] sladen: ogra's critic was that he couldn't update the /usr/share/initramfs-tools script to set fb1 easily as a hack to workaround usplash picking the wrong res [12:11] The real bug is usplash handling of multiple screen [12:11] indeed [12:11] well, deeper even i guess [12:11] In fact we shouldn't be reading the size from /sys but just let bogl figure out the best size [12:11] the kernel seems to pick the res for my laptop LCD in kms [12:11] But that's orthogonal to fb0 versus fb1 [12:12] instead of using the external one [12:14] putting it in C code let us simplify the upstart job quite a bit [12:14] if it needs to be configurable for now, have the C code read it from usplash.conf [12:16] * sladen blantently assumed the script in question was under /etc already [12:18] cjwatson, it doesnt need to [12:18] tseliot: I don't understand why in nvidia-common you aren't just making sure the stdin/stdout of the nvidia-detector call are on clean fds? === MacSlow is now known as MacSlow|lunch === mpt_ is now known as mpt [12:25] mvo: should bug #432819 be un-targeted as well as un-milestoned, or do you consider this a candidate for SRU? [12:25] Launchpad bug 432819 in software-center "software-center doesn't respect X-AppInstall-AlwaysOnTop" [High,Confirmed] https://launchpad.net/bugs/432819 [12:27] mvo: could you also comment on bug #434173? [12:27] Launchpad bug 434173 in language-selector "[karmic] Regression from 9.04 in getting fully translated Ubuntu installation" [Undecided,Fix committed] https://launchpad.net/bugs/434173 === apw` is now known as apw [12:56] slangasek: cleans fds such as? [12:56] s/cleans/clean/ [12:58] tseliot: use shell redirection to control what goes in and out of the process? [12:59] kirkland: around ? [12:59] dholbach: hmm. I'll upload a new package for those changes. [13:00] kirkland: committed a euca -0ubuntu3 with a hotfix for the wrong IP issue and the web UI changes from Gustavo. I waited for the CD spin to go with the euca_rootwrap change so that it can be tested independantly [13:01] ogra: do you get pulsating ubuntu logo on livefs boot now? it works for me now [13:01] but the wrong IP issue was a pita that needed to be avoided asap and the ui changes had to go in urgently [13:02] pitti, i havent tested live images since a while [13:02] slangasek: patches are welcome ;) [13:02] slangasek: 432819> the alwaysOnTop should be removed for karmic. I don't think it can be done in a way suitable for a sru [13:02] slangasek: I have a look at the language-selector one now (I have to add that l-s is no longer officially my baby) [13:02] ogra: would be nice to confirm [13:03] tseliot: shall I reject the upload in the queue, and put together a patch? [13:03] pitti, will do once the board is available (it's busy atm) [13:03] slangasek: with my solution I made sure that things don't crash in any case [13:04] slangasek: so, please accept it and if you have the time to work on a patch, I'll be glad to review it [13:04] mvo: the language-selector bug has a task open on update-manager, and cjwatson commented in the report that ArneGoetje should talk to you [13:05] kirkland: note that the wrong IP hotfix doesn't address the question of "adding a linklocal address confuses ifconfig" which needs to be discussed more. [13:07] mvo: do you have any observations on bug 452090 ? davmor2 failed to upgrade from hardy [13:09] Riddell: I just commented on it === The_Company is now known as Company [13:10] mvo: hum, anything that can be done about it? [13:12] Riddell: not sure, was this a stock kubuntu hardy setup? [13:12] mvo: I expect so, davmor2? [13:12] * mvo tests in a VM [13:14] tkamppeter: ah, the buildd hostname fix is being worked on === MacSlow|lunch is now known as MacSlow [13:26] pitti, I got this mail from slangasek, too. [13:26] So probably we can upload 1.4.1-8 today and it will get built. [13:27] pitti: It is unusual for a package investigating the build server (reading IP, searching for printers and available PPDs, ...) during build. [13:29] tseliot: I'm looking at this nvidia-common bug report and the changes, and I don't see how this could ever work... the script you're calling this from is invoking debconf, which means stdout of this script is *always* attached to the debconf frontend - flushing output doesn't help anything? [13:32] slangasek: I can't reproduce the problem here. The script won't crash though, which is a big win. [13:33] slangasek: if the script gets garbage instead of the proper output it won't give it to debconf as a reply [13:38] mvo: yup [13:39] tseliot: I really can't figure out why you've gone with this solution, or which fd it is you think has the garbage on it (which is generally confusing enough when debconf is involved...) [13:39] mvo: stock + proposed for the adept update [13:39] tseliot: why should nvidia-detector ever need to be run in a loop? [13:40] mvo: sorry forgot + updates [13:40] tseliot: nvidia-detector doesn't appear to read from stdin; if it did, you could work around this with LATEST=$(nvidia-detector < /dev/null); and the stdout of nvidia-detector is already being captured [13:42] slangasek: if all goes well, it is called once, otherwise try again and, if this goes wrong, just give up [13:42] yes, why do you think it should *ever* need to be called more than once? [13:42] slangasek: no, it doesn't read from stdin [13:42] this doesn't seem to have anything to do with the bug= [13:43] I'm not going to accept something I don't understand during final freeze... [13:45] slangasek: as you can see in the bug report, it can happen that my script doesn't get the output of nvidia-detector because dkms leaves something like " * Running DKMS auto installation service for kernel 2.6.27.2" on stdout [13:46] and LATEST will be assigned that value [13:46] tseliot: eh, I don't see that at all from the bug report, or from the code [13:46] slangasek: https://bugs.launchpad.net/ubuntu/+source/nvidia-common/+bug/292606/comments/4 [13:46] Launchpad bug 292606 in nvidia-common "dkms - error when installing custom kernel" [Medium,New] [13:47] slangasek: what I know for sure is that nvidia-detector can give print "nvidia-glx-$VER" or "none" [13:47] tseliot: that shows the bufferred stdin being sent *to debconf*, not preventing reading the correct output from nivdia-detector [13:49] slangasek: how? [13:50] that's the point [13:52] tseliot: the parent process has already started the debconf frontend before calling this script; it looks like the garbage has already been sent to the debconf frontend before the nvidia-common script has ever been called, and nvidia-common is just the first process to read back the resulting error code from the pipe [13:52] (i.e., I agree with Mike Stroyan's analysis in the bug) [13:52] oh, yes, this is like malloc debugging [13:52] it's often the next thing along that gets to encounter the problem === marjomercado is now known as marjo [13:54] slangasek: ok, so are you suggesting that I simply prevent the script from failing without trying again? [13:55] or do you have a different solution in mind? [13:55] tseliot: yes - and that the way to do this is by fixing dkms, not nvidia-common [13:56] what's starting debconf, though? [13:56] slangasek: ouch, again. superm1 has a few objections though (to fixing that in dkms) and this is why I tried to deal with the problem in nvidia-common [13:56] kernel postinst? [13:56] cjwatson: not that I see [13:56] yep ^^ [13:57] hmm, it used to I think, obviously no longer [13:57] /etc/kernel/postinst.d/ [13:57] 2.6.27.7-k8 though, that's kind of old [13:57] and looks custom [13:57] nvidia-common installs the script there [13:58] this is meant to help users who dist-upgrade with apt from the command line [13:58] as Update Manager already uses nvidia-common to get rid of obsolete drivers [13:59] kernel image packages generated using make-kpkg still load the debconf confmodule in the postinst [13:59] right [14:00] so kernel postinst starts the frontend, dkms inherits the fd and throws garbage at it, nvidia-common catches it [14:00] yes, it could be [14:01] and while Ubuntu images of that same vintage apparently didn't, /etc/kernel/postinst.d/ in general probably has to assume that stdout might be debconf [14:01] so this is a dkms bug, and superm1 will just have to accept that :) [14:03] mvo: I've commented on the bug as to what I had done. [14:03] slangasek: ok, so in the mean time I'll modify nvidia-common so that it's capable of detecting garbage but doesn't call nvidia-detector more than once. Right? [14:04] tseliot: that sounds better to me, yes - how are you going to detect the garbage? [14:04] i.e. I'll get rid of the while loop [14:05]