[00:10] kirkland: you around? [00:17] achiang: howdy man [00:17] achiang: wassup? [00:17] achiang: wanna meet me in Aspen in a couple of weeks? :-P [00:17] achiang: i have a saturday/sunday on the slopes coming up in a few weeks :-) [00:22] kirkland: ooh [00:22] kirkland: what weekend? i'm actually skiing silverton over president's day [00:26] kirkland: i have a relatively fresh maverick install, just installed virt-manager, and trying to create a VM, running into: http://pastebin.ubuntu.com/556023/ [00:28] achiang: hmm, interesting [00:28] achiang: what does 'id' say about your user [00:29] achiang: i'm interested in what groups its a member of [00:29] achiang: specifically, you should be in the kvm and libvirtd groups [00:29] achiang: the packaging should have added that for you [00:29] kirkland: i'm in libvirtd, but not kvm [00:29] achiang: but if not, sudo usermod -a -G kvm $USER [00:29] achiang: but if not, sudo usermod -a -G libvirtd $USER [00:30] achiang: then you'll need to logout of all sessions and log back in (or just reboot) [00:30] achiang: that's the most common mistake i've seen [00:30] ok, back in a jiffy [00:30] kirkland: but the packaging didn't add it [00:30] kirkland: should be easily recreatable [00:30] back in a sec [00:30] achiang: right, well, the packaging adds libvirtd, maybe/maybenot kvm [00:30] achiang: k [00:30] You don't need to be in the kvm group. [00:31] achiang: right, as soren is about to tell you, the kvm thing is now handled automagically by PolicyKit [00:31] Um. No :) [00:31] If you use the systemwide libvirt daemon (which you do if you're a member of libvirtd), you don't need to be a member of kvm. [00:32] Besides, I think the time is ripe for chmod 666 /dev/kvm. It doesn't give you any privileges anymore that are worth containing. [00:32] back [00:32] trying again [00:33] and just read scrollback. :) [00:34] kirkland: soren: hm, adding myself to kvm group and restarting seemed to help on one machine [00:34] achiang: rock [00:34] on another machine, i went into advanced options, deleted the console device, and it too could start [00:34] (without being in kvm) [00:35] so... is this a bug i should file? [00:35] achiang: okay, glad to hear you're sorted out [00:35] achiang: i'm not sure I understand the situation clearly enough to say so [00:35] achiang: to be on the safe side, i'd say "yes" [00:35] kirkland: ok. i'll try and see if i can reproduce it [00:35] achiang: mdeslaur maintains virt-manager for the most part [00:35] achiang: cool, thanks. [00:36] achiang: You basically need to be a member of either libvirtd or kvm. libvirtd is cooler, because it lets you fiddle with networking. [00:36] ...but to just run virtual machines, either works. [00:36] kirkland: ok, i'll just file it against virt-manager, but won't assign it to anyone [00:36] achiang: as for skiing, i'm out there the weekend of Feb 5/6 [00:36] kirkland: another weird thing, i thought i had qemu-kvm installed, but when trying to create a machine, it says kvm is not installed [00:37] achiang: ah, well, then, that'll make it hard to run vm's :-) [00:37] http://pastebin.ubuntu.com/556025/ === releaselogger is now known as apachelogger [00:58] soren: uh, doesn't /dev/kvm basically get you the ability to muck with the whole kvm subsystem, or did that change? i don't like the idea of giving random users access to that without the admin opting in [00:58] just like i don't want /dev/fuse to be 666 [01:07] err...wait...when did /dev/fuse become 666? [01:08] so what's a guy gotta do to get the whole natty archive rebuilt? ;) [01:09] (or is it safe for /dev/fuse to be 666 because fusermount has finer controls?) [01:09] now that this gcc bug has been fixed... === dendro-afk is now known as dendrobates === asac_ is now known as asac === bjf is now known as bjf[afk] === v is now known as stalcup [08:07] good morning [08:12] good morning [08:13] K'day [08:15] ebroder: What's so interesting about the kvm subsystem that you want to limit access to it? [08:15] morning dholbach [08:16] hey davmor2 [08:16] ebroder: I can't consume any more resources through it than you could otherwise do just with malloc or a tight loop. It's not like there's a limited supply of /dev/kvm or anything. Nor can you screw up other people's VM's. There's really no reason to limit access to it. [08:17] ebroder: Once upon a time, having access to /dev/kvm meant you could lock memory pages in RAM. So a malicious user could create a VM with as much memory as the host, and you were DOS'ed. [08:18] ebroder: Nowadays, KVM memory can be paged out, so it's no different from just malloc(). [08:23] soren: i'm more concerned about security issues. i basically trust unix file permissions. not sure whether i want to trust kvm any more than i have to [08:24] ebroder: I'm not sure what you mean? [08:24] in a nutshell, principle of least privilege [08:24] ebroder: The memory of a virtual machine is as easy or difficult to get to as the memory of any other process. [08:25] you're missing my point. i'm not concerned about the security of the guests. i'm concerned about the security of the host [08:27] so am I. And I don't believe /dev/kvm exposes anything that compromises the host's security. [08:27] you're assuming kvm is bug-free [08:28] Not really. [08:28] ..but the other 99.99% of the kernel isn't either. [08:28] Yet we don't limit people's access to open sockets, for instance. [08:28] sure, but it's all about minimizing the attack surface [08:28] We could. [08:28] but we don't. Because it's silly. [08:29] and we are starting to patch out autoloading of obscure protocols and stuff [08:29] a) kvm is not obscure [08:29] b) there are bugs in non-obscure parts of the kernel, too [08:30] I just don't subscribe to the idea of amputating my system to the point of uselessness in the name of security. [08:31] I think this is a symptom of "oh, wow, there's a device node for this stuff! Let's limit access to it!" [08:32] uselessness? it's not useless to reduce my attack surface by a whole subsystem when i personally run all of my vms managed by the system libvirtd [08:33] I understand that we've worked around these limitations for kvm. I know. I did it. [08:33] limitation? i'm not doing that because i think i have to [08:34] Well, whatever you want to call it. [08:34] We've worked around the fact that access to /dev/kvm has historically been restricted. [08:35] If kvm had been implemented by way of a syscall interface, noone would have thought of trying to impose artificial limitations on access to it. [08:35] (syscalls other than ioctl, that is) [08:36] Anyway, this is going nowhere. [08:36] and fast. [08:39] query Sarvatt [08:47] siretart: please could you have a look at the xine-lib build failure? http://launchpadlibrarian.net/61920076/buildlog_ubuntu-natty-i386.xine-lib_1.1.19-2ubuntu2_FAILEDTOBUILD.txt.gz === hungerWork_ is now known as hungerWork [08:59] I am getting error while trying to "branch" in bazzar .Error is "bzr: ERROR: exceptions.NotImplemented Error: should resend request to http://feeds.edge.launchpad.net/bazaar/, but this isn't implemented ".Can anyone provide help on this?? [09:00] Good morning [09:01] hey pitti === _LibertyZero is now known as LibertyZero [09:04] bonjour didrocks, comment vas-tu? [09:07] pitti: ça va bien, merci, et toi? :) [09:07] tres bien, merci! [09:12] I just ran into something rather odd. By way of Mozilla's "Game On" gallery page, I found a game which seems to copy a pile of content from the Ubuntu homepage and dummy it out, either due to copying parts of Ubuntu's page as a template or perhaps trying to get Google juice for random search terms. Who might I go about reporting that to? [09:14] Specifically: https://gaming.mozillalabs.com/games/ led to https://gaming.mozillalabs.com/games/118/the-humanity-project which points at http://farsides.com/ ; check out the links in the green box at the bottom. [09:20] JoshTriplett: contact webmaster@canonical.com (if nobody else gives you a better advice) [09:21] geser: Fair enough. [09:21] geser: Thanks. [09:24] geser: Mailed. === zyga is now known as zyga-efika === almaisan-away is now known as al-maisan [10:02] hi chrisccoulson, thanks for merging danilo's po2xpi branch. Do you think you could now look at automatically uploading the xpi's with every package upload, as we discussed? The easiest way I can think of is simply fetching them from http://releases.mozilla.org/pub/mozilla.org/firefox/releases/4.0b9/linux-i686/xpi/, and I guess it shouldn't be too different to what you are already doing for getting the translations for the Thunderbird locale packages [10:05] Could that be semi-automatic, rather than automatic? It seems like it could go very wrong or be very awkward in the case of a tiny SRU patch, or similar. [10:21] !regression-alert [10:21] cjwatson, jdong, pitti, slangasek, ScottK, mdz, kees, ttx, marjo, seb128: reporting regression in a stable release update; investigate severity, start an incident report, perhaps have the package blacklisted from the archive [10:21] ubuntu-docs 10.10.4 which was SRUed as Bug #677998 also contains unrelated changes, http://launchpadlibrarian.net/60441147/ubuntu-docs_10.10.3_10.10.4.diff.gz [10:21] Launchpad bug 677998 in ubuntu-docs (Ubuntu Maverick) "openldap-server instructions have a small bug" [Undecided,Fix released] https://launchpad.net/bugs/677998 [10:21] This causes regression Bug #690248 which shows using 11.04 in spite of using 10.10 [10:21] Launchpad bug 690248 in ubuntu-docs (Ubuntu Maverick) "In Maverick 'About Ubuntu' displays Natty info" [High,Triaged] https://launchpad.net/bugs/690248 === MacSlow is now known as MacSlow|lunch [11:01] didrocks: we're blocking on bug 702026 [11:01] Launchpad bug 702026 in dcmtk (Ubuntu) "[MIR] dcmtk" [Undecided,New] https://launchpad.net/bugs/702026 [11:03] Riddell: I'll process it before EOD [11:08] seb128, poke ... where is that rhythmbox upload you talked about ? do i need to unseed it to get images ? [11:09] ogra, it's coming today, I ran into other issues yesterday and didn't get it done [11:09] ugh, seems that today's CDs grew by 70 MB, due to LibO? [11:10] seb128, k === csurbhi is now known as csurbhi-lunch [11:10] pitti, switch to arm :) we still have space [11:11] the language pack rebuild saved 50 MB [11:11] and LibO added 130 [11:11] ah, it now pulls in Java [11:11] openoffice.org-common (17.0 MB) [11:12] libreoffice-common (36.6 MB) [11:14] ah, the langpack rebuild saved 8.5, sorry [11:22] Riddell: could you please reject mumble from lucid and maverick-proposed, it will be handled by a security upload [11:24] * ogra sighs about 160 bugs in his bugmail [11:24] bah [11:24] and most of them seem to be expired installer bugs ... i wonder if thats ok [11:26] debfx: ok [11:27] ogra: those are the best kind, no action necessary :) [11:27] pitti, i really doubt that, they were all marked incomplete by a script from some random guy while we were in florida [11:28] i'm sure cjwatson would have liked to keep some of them open === al-maisan is now known as almaisan-away [11:37] Hi cjwatson , I would like to know if its possible to package partman.udeb inside initrd so I don't have to setup a mirror with my modified partman udeb, thks in advance ? [11:38] right now to test partman with modifications I need to setup a mirror, because my installation method is only by network (netboot). [11:46] hi ev, may I ask you to export translations from LP and add them to ubiquity before A2? This way everyone can start testing them. [11:47] dpm: sure thing [11:47] ev, awesome, thanks [11:53] <\sh> guys, did anyone saw this on lucid in the past few weeks: cron[4326]: segfault at 7f77c036a970 ip 00007f77c036a970 sp 00007fff81bd5bb8 error 4 in libnss_files-2.11.1.so[7f77e768b000+c000] ? [11:53] hello everyone! i need some help (not support) figuring out if this is some bug with plymouth worth filing: [11:53] http://img.xrmb2.net/images/787502.png - note the big gap after ureadahead finished, plymouth-debug.log: http://paste.ubuntu.com/556129/ [11:53] this is a tailored 2.6.38-rc1 with no swap and apparmor disabled (, i get similar boot times with ubuntu's kernel [11:57] htorque, does that graph say that plymouth is in an uninterruptible sleep ? [11:58] apw, the plymouthd bar it's slightly red, so that should mean it === almaisan-away is now known as al-maisan [12:03] i'll be off for an hour, i'll check irclogs, so if you need me to do something, go ahead :-) === MacSlow|lunch is now known as MacSlow [12:07] htorgue, not seeing that on my system here, am seeing other issues but not that long gap, i suspect plymouthd is printing the dots hense the blips [12:08] the pause is likely something else === oubiwann is now known as oubiwann_ === oubiwann_ is now known as oubiwann [12:15] do we have any idea how to fix bug 688732? [12:15] Launchpad bug 688732 in pywebkitgtk (Ubuntu) "package no longer has WebView attribute after transition to python 2.7" [High,Triaged] https://launchpad.net/bugs/688732 [12:22] elif: I'm not going to change the initrds we ship to do that, if that's what you mean. there's documentation in the d-i build system for how to add packages to it if you want to do it locally. === csurbhi-lunch is now known as csurbhi === cking is now known as cking-afk [12:25] cjwatson: I meant locally (my installation). I didn't find the docs you mentioned, could you point me ? [12:26] cjwatson: its d-i docs ? [12:28] build/README in the debian-installer source package [12:28] easiest way is to list the names of udebs you want to include in build/pkg-lists/local [12:29] cjwatson: thank you. [12:40] dholbach: no idea where that file came from. Submitter's system was installed with 10.10, which used python-support for pywebkitgtk, it still does. That location is a python-central/dh_python2 location. [12:43] dholbach: aah, 1.1.8-1 looks like the culprit [12:45] Any one here? [12:46] !ask | phenom [12:46] phenom: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) [12:46] Not a question [12:46] Wubi installer is broke [12:46] Spits pyrun.exe says "no disk" [12:46] and then says Can not downloan metalink, and therefor the ISO [12:46] And dies [12:47] I assume it runs on win 7 [12:48] are you asking if it runs on win7 or saying you are running on something other than win7 (and assume it works on win7) [12:48] I'm saying, I tried to run it on win7 and it gives me those errors [12:48] and dies [12:49] it spits the pyrun.exe says "no disk" error about 15 or so times [12:50] and then continues through the installer, after I input the password if spits it can not download the metafile and therefor can not download the iso [12:50] which version? [12:52] rev 190 [12:52] the one from http://www.ubuntu.com/desktop/get-ubuntu/windows-installer [12:52] ev: ^- can you look at that? [12:53] cjwatson: yup === hungeWorkr__ is now known as hungerWork [12:58] phenom: can you pastebin %TEMP%/wubi-10.04.1-rev190.log [13:06] kees: hey, if you can do a security review on https://bugs.launchpad.net/ubuntu/+source/dcmtk/+bug/702026, that will be nice :) [13:06] Ubuntu bug 702026 in dcmtk (Ubuntu) "[MIR] dcmtk" [Undecided,Incomplete] [13:10] ev, pm [13:12] phenom: thanks, I'll have a look after lunch [13:13] Sure, thanks [13:13] didrocks: #ubuntu-hardened is good place for that question :) [13:14] ari-tczew: I just ask for a security check on a MIR process, that can take some days, so I'm not sure I'll stay in #ubuntu-hardened there I think I hilighted the right person about it and it's done :) === cking-afk is now known as cking === sconklin-away is now known as sconklin [13:40] apw: thanks for the feedback. i will keep an eye on it with mainline and ubuntu kernels once they are out - maybe it's just a misconfigured kernel on my end. [13:40] htorque, ubuntu kernels will be a while as 32 don't boot on my test boxes [13:42] apw, yeah, i had to revert that commit about relocatable kernels you mentioned (it was not enough to just disable that option) [13:42] i had to do that indeed, which worked for 64 bit, but my 32 bit box is failing to enable the timer [13:42] and exploding === al-maisan is now known as almaisan-away === oubiwann is now known as oubiwann_ [14:17] @pilot in === udevbot changed the topic of #ubuntu-devel to: Archive: Open | Development of Ubuntu (not support, not app development) | #ubuntu for support and general discussion for dapper -> maverick | #ubuntu-app-devel for application development on Ubuntu | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs | Current Friendly Patch Pilots: smoser [14:18] * dholbach hugs smoser [14:22] pitti: FYI, I'm wrestling with my last couple of bugs for the lucid SRU freeze ... expect uploads over the next hour or two [14:22] trying to make USB stick upgrades not suck [14:25] ack === russellb_ is now known as russellb [14:55] smoser: have you got upload access for main? [14:55] no [14:55] what's the process for dealing with regression-proposed bugs? [14:55] otherwise i would sponsor your changes. (i guess you're looking for sponsor for 705294) [14:56] good afternoon btw. [14:56] hello kklimonda [14:56] smoser: nope, I'm looking for sponsor for main [14:56] (the process after marking sru bug as verification-failed, and the new one as regression-proposed, and nominating it for release) [14:57] smoser: then you're weak patch pilot if you don't have upload access for main [14:57] i agree i'm a weak patch pilot [14:57] ari-tczew: you were also a patch pilot other day [14:57] i go through the sponsor list and see if i can make job easier for others. [14:57] kklimonda: yes === smb` is now known as smb [14:58] bug 705294 [14:58] Launchpad bug 705294 in openmotif (Ubuntu) "[FTBFS] Source openmotif 2.3.3-5 Natty" [Undecided,In progress] https://launchpad.net/bugs/705294 [14:59] smoser: what is wrong with above? ^^ [14:59] ari-tczew, nothing. i just assumed thats what you had a fix for and were asking for a sponsor of. === almaisan-away is now known as al-maisan [15:00] smoser: I still understand. [15:00] don't understand * [15:01] I'm sponsoring 2 bugs for udienz right now. [15:01] never mind then. i was confused. [15:01] I'm looking for main sponsorship for 2 things which I reviewed. [15:01] ah. [15:01] https://code.launchpad.net/~barry/ubuntu/maverick/computer-janitor/bug-665740/+merge/46039 [15:01] https://code.launchpad.net/~om26er/ubuntu/maverick/nautilus/nautilus-fix-630512-maverick/+merge/45048 [15:01] https://code.launchpad.net/~clint-fewbar/ubuntu/karmic/mysql-dfsg-5.1/mysql-sru-343870/+merge/42667 [15:02] update, 3 things [15:02] and I don't understand why patch pilots didn't find main sponsor for these patches [15:02] doko_: do you know about the "pystack" gdb feature? === yofel_ is now known as yofel [15:10] ari-tczew: well, for one, the "fix" for https://code.launchpad.net/~barry/ubuntu/maverick/computer-janitor/bug-665740/+merge/46039 looks OBVIOUSLY incorrect [15:11] you should fix the python-dbus package instead of working around a bug in it [15:12] Chipzz: perhaps, but would that have too much impact for an sru? [15:12] Chipzz: hehehe, you're looking for hole in the whole [15:12] the subject is a response for core-dev === Quintasan_ is now known as Quintasan [15:13] * Chipzz looks at his screen and tries to decipher what ari-tczew is saying [15:13] no offence meant, but your English is quite poor [15:16] Chipzz: nobody is perfect, maybe you? === kirkland_ is now known as kirkland [15:17] barry: there's the question of what you should fix in an SRU in the first place; ignoring wether that bug is a candidate or not, fact is you should try to keep the changes minimal, which among other things means trying not the fix the same issue in different ways; and since you most likely want to fix the issue in the development release too, and you want to do it the proper way there, I think the best thing to do would be to fix it in python-dbus [15:17] ari-tczew: sigh :p [15:18] ari-tczew: I was looking at your sentence for 10 seconds, trying to understand it. then I looked again. and then I gave up trying to figure it out [15:18] I'm not sure why python-dbus should depend on dbus? [15:18] btw, python-dbus depends on libdbus-1-3 [15:19] which makes sense, but i think soren has a point [15:19] and libdbus-1-3 recommends dbus [15:19] I don't see why computer-janitor should depend on dbus actually [15:20] Now that is a different matter. I don't know. [15:20] Chipzz: because the core functionality is in a dbus service (in >=maverick) [15:20] csurbhi: hey ... quick question. mkfs.btrfs has no -U option. are uuids not supported for btrfs? [15:20] barry: errr, I'm just making a wild guess here, but how about "No Scott"? :p [15:21] barry: you don't have every pygtk application depending on libgtk now, do you? [15:21] Chipzz: now you're not making sense :) [15:21] barry: if your app talks to dbus via python-dbus, then the dependency is on python-dbus, not dbus [15:22] c-j also ships its own dbus service; it isn't merely a client [15:23] Chipzz: i think it's a different case. libgtk is a library but dbus is a service [15:23] cjwatson: then it's a matter of how the dbus service interfaces with dbus, isn't it? [15:23] at least some other packages that ship dbus system-services depend on dbus directly (e.g. consolekit) [15:23] there is clearly precedent [15:23] system-tools-backends [15:23] avahi-daemon, network-manager [15:23] etc. [15:24] barry: imo they're both bindings, which is why the 2 cases should be treated the same [15:24] Chipzz: i think the analogy would be to libdbus though, not dbus [15:24] depending on dbus directly seems correct to me. it's relying on the dbus daemon directly and the dependency should be direct. [15:25] I'm happy to sponsor that SRU as written [15:25] (and will do so shortly [15:25] ) [15:25] cjwatson: thanks. Chipzz and ari-tczew thanks also for the feedback. :) [15:26] cjwatson: if you think it's correct, I'll stop arguing about it, but it still looks wrong to me [15:26] Chipzz: :) :) :) [15:26] but I have to go offline anyway [15:26] I think it's correct and that your analogy is faulty given the specifics of this package, yes [15:26] Chipzz: They're not both bindings. [15:26] Chipzz: dbus is the daemon. [15:26] if you were talking about libdbus, the analogy would be sound [15:26] but that's not (only) what's at issue here [15:27] cjwatson: my reasoning is that the dependency chain should likely be app -> python-dbus -> libdbus -> dbus [15:27] that reasoning is faulty, because this application uses dbus in ways other than via python-dbus [15:27] since it makes little sense to have an app linked against dbus (be it client or service) without having dbus available [15:27] specifically by shipping a dbus system-service [15:27] anyway gtg === bjf[afk] is now known as bjf [15:31] Chipzz: think of optional plugins [15:32] Chipzz: e. g. bzr-gtk annouces commits over session dbus, but it's only a bonus feature; so bzr-gtk would certainly depend on python-dbus (to not crash), but not on dbus (since it works without) [15:34] You can also use libdbus without the dbus daemon. [15:34] upstart does, for instance. === dendrobates is now known as dendro-afk [15:35] right, and in any case inserting an extra dependency further down the stack has *more* risk of regressions, not less, so I'd argue it'd be less appropriate for an SRU even if it were correct (which I don't think it is) [15:35] *nod* [15:35] cjwatson: right, that was my thinking too [15:36] mvo: ping http://launchpadlibrarian.net/61917866/buildlog_ubuntu-natty-i386.update-manager_1%3A0.145.10_MANUALDEPWAIT.txt.gz [15:38] siretart: ping [15:40] so, here's an unrelated, but general question. i've seen this a few times with packages i'm doing test builds with (latest example: python-webkit). i branch the source from ubuntu:python-webkit, then bzr bd -S (or debuild -S). dpkg-source exits w/status 255 because "Version number suggests Ubuntu changes, but Maintainer does not have Ubuntu address". this is true, but what's the right way to handle this for local builds? i've just [15:40] been running update-maintainer before doing a local build, but why hasn't that already been done for a -XubuntuY version? [15:41] pitti: libreoffice-common: is this still true with the -2ubuntu2 version? [15:41] probably because the last person who changed it in Ubuntu doesn't have @ubuntu.com in their DEBEMAIL so dpkg-source didn't error out on that [15:41] for quick local builds, I usually do DEBEMAIL=something@else debuild ... [15:42] (I might mean debuild -eDEBEMAIL=something@else ..., I forget) [15:42] doko_: seems to be the same size as ubuntu1, according to https://launchpad.net/ubuntu/+source/libreoffice/1:3.3.0~rc3-2ubuntu2/+buildjob/2183350 [15:43] cjwatson: ah, i didn't realize dpkg-source looks at DEBEMAIL (not obvious from manpage afaict). thanks [15:44] it was a hack to try to get people to DTRT in their uploads without impeding *too* many people's local builds when people get it wrong === dendro-afk is now known as dendrobates [15:45] * barry loves social change by technological hack :) [15:45] DEBEMAIL= debuild works, yeah [15:45] yep, wfm, thanks! [15:45] it at least means that most uploaders remember, even if the contributor does not [15:47] my other dodgy hack along those lines is that I occasionally do debuild -ePATH=$HOME/fake-debconfpo:$PATH -S [15:47] where $HOME/fake-debconfpo is a directory containing only a symlink from debconf-updatepo to /bin/true [15:47] Heh. [15:47] this is helpful in cases where packages call debconf-updatepo on build and I don't want them to (e.g. because I'm building for lucid on natty and gettext's default header has changed) [15:48] just about anything can be worked around if you try hard enough ... [15:53] zul, daviey, kirkland: who's picking up bugs for cobbler now its in archive? [15:54] JamesPage: server team is already subscribed to the bugs so I guess all of us [15:54] JamesPage: all of us :-) [15:54] * kirkland knuckle bumps zul [15:54] :-) === oubiwann_ is now known as oubiwann [16:06] pitti: I see, -common has the images_*.zip for all styles, see bug #696527 [16:06] Launchpad bug 696527 in libreoffice (Ubuntu Natty) "LibreOffice - Human icons theme disabled, patch needs an update" [High,Confirmed] https://launchpad.net/bugs/696527 [16:06] doko_: aah, thanks [16:06] doko_: so I guess we'll need to keep that around for now, and just drop the -writer recommends for JDK? === beuno is now known as beuno-lunch [16:09] pitti: casper and grub2 uploads in {lucid,maverick}-proposed, fixing sabdfl bugs ;-) [16:10] csurbhi: did you get anywhere with that milestoned util-linux bug? [16:10] cjwatson, i am still working on the installer btrfs thing [16:10] i will get at the bug soon and shall try to do it before Sat [16:11] pitti: yes, I'll do that. [16:11] doko_: danke sher [16:11] how urgent is this? [16:11] csurbhi: oh - as I said in yesterday's meeting, the lucid SRU freeze (for 10.04.2) is today [16:11] doko_: by alpha-2, I think (since currently we have a 62 MB oversizedness which we can't fix by langpack dropping) [16:11] should I find somebody else to look at it? [16:12] doko_: alternatively we can just unseed it for a2, and put it back later [16:12] cjwatson, i am not sure i will be able to finish by eod [16:12] but i can try it [16:12] ok, thanks [16:12] ok [16:12] since there is a patch to the bug already [16:12] let me know if somebody in a more western timezone needs to pick it up [16:13] cjwatson: (talking to folks at the hackfest ATM, will look at it in a bit) [16:13] cjwatson, no let me first try doing it [16:13] i will get back in an hour or two [16:13] if i think i cant [16:14] ok [16:14] pitti: ok, this seems to be doable. I'd like to avoid dropping -writer [16:14] pitti: np, thanks [16:17] pitti: just want to wait for the armel builds [16:17] sounds good === dendrobates is now known as dendro-afk [16:19] didrocks: if you haven't already, please assign ubuntu-security to the mir, and I'll pick it up when I start work this morning :) [16:20] kees: excellent, will do, thanks:) [16:34] Riddell: could you have a look at the telepathy-qt4 build failure? [16:46] doko_: I'll add it to my todo list === diwic is now known as diwic_afk === james is now known as Guest26472 [17:03] kirkland: hi, can i bug you again? [17:03] achiang: sure, hit me [17:03] kirkland: just wondering why i'm seeing this: http://chizang.net/alex/tmp/Screenshot.png [17:04] kirkland: and i do have kvm module loaded... [17:04] achiang: natty? [17:04] achiang: what is the output of 'kvm-ok' ? [17:04] maverick [17:04] achiang: ^ [17:04] kirkland: ah. says i don't have it enabled in BIOS [17:05] kirkland: so i'll go do that [17:05] thanks [17:06] when exactly does .postinst get called? I assume it's after 'binary_arch: install' from debian/rules? [17:08] hallyn: well after - when the user calls dpkg --configure (perhaps via dpkg -i, perhaps via apt-get, etc.) [17:09] so on a completely different machine from that running debian/rules in most cases === dendro-afk is now known as dendrobates [17:09] oh. hm. [17:10] so if files are being installed which should belong to a new groupname, if that groupname is created in the .postinst, is that good enough? I was thinking the system would refuse to create files owned by nonexisted group [17:11] achiang: yup [17:12] hallyn: if the files are being created at run-time (e.g. in the postinst, or when the program is run), then the postinst is good enough [17:12] (and Depends: adduser) [17:12] hallyn: if you're shipping these files in the .deb, then you need to run adduser in the preinst, and Pre-Depends: adduser [17:12] kirkland: hm, BIOS claims that VT-D and VT-X are turned on, yet kvm-ok disagrees [17:13] hallyn: http://www.debian.org/doc/debian-policy/ch-files.html#s-permissions-owners [17:13] kirkland: is there a better channel we could debug this in? seems a little OT for here [17:13] cjwatson: thanks, that was my guess, except i had no idea about Pre-Depends [17:14] achiang: either here, or #ubuntu-server, or #ubuntu-cloud; i'm in all of these, it's okay in any of them, IMHO [17:14] cjwatson: ok, lucid SRU queue empty; need one reupload from Micah (mailed him), rest looked o [17:14] k [17:14] kirkland: how about #ubuntu-virt? [17:14] hallyn: is this for kvm? [17:14] good night everyone! [17:14] pitti: yay, thank you [17:14] achiang: #ubuntu-virt -> #ubuntu-cloud [17:14] ah [17:14] pitti: I'll do a CD build pass tomorrow [17:15] cjwatson: we have a hackfest group dinner now, so I'm afraid you or slangasek need to handle the other stragglers today [17:15] cjwatson: cool, thanks [17:15] what's the process for dealing with regression-proposed bugs? [17:15] (the process after marking sru bug as verification-failed, and the new one as regression-proposed, and nominating it for release) [17:16] kklimonda: by default, the previous uploader should upload a fix [17:16] if it's a serious regression, we can also remove a package from proposed [17:16] pitti: ok, thanks [17:17] pitti: I bump the version, add a new changelog entry and create source package in a way that displays both changelog entries in the changes file? [17:17] kklimonda: right, using debuild -vVersionInUpdates === beuno-lunch is now known as beuno [17:18] * pitti needs to run now === sebner_ is now known as sebner [17:20] kirkland: yeah [17:20] kirkland: this is why on first-ever qemu installs, /dev/kvm is owned by root :) on second attempt, when we ask them to re-try it works fine :) [17:20] (someone complained last week about the same thing, right?) [17:26] hallyn: i've seen recent complaints; is this something that got dropped in the last merge? [17:26] hallyn: as i don't think we've seen this before that [17:28] that's the weird thing, i don't think this ahs changed. at least i don't remember playing with it. but let me check the maverick sources === al-maisan is now known as almaisan-away [17:29] you know, for all the compalints ppl have about compiz, one great thing about it is that vnc doesn't insist on redrawing every time i put a window over it or switch viewports [17:33] :) [17:33] kirkland: hell - have you restarted the machine since installing kvm? [17:33] hallyn: nope [17:34] hah, THAT's it [17:34] cool.. they're going to film a movie on my street on Saturday [17:34] hallyn: ? really [17:34] it's a udev rule, so next time it'll work jsut fine. [17:34] hallyn: hmm, do we need to trigger a reload or something postinst? [17:34] i'm just moving group creation to preinst [17:34] hallyn: because "reboot your machine after installing kvm" is kinda silly [17:34] :) [17:34]