[00:05] micahg: is this ok for ubuntu-restricted-addons? http://paste.debian.net/38606/ === salem_ is now known as _salem [01:03] jbicha: I don't see the point of pulling in chromium codecs everywhere [01:07] because chromium is widely used and we can't have anything else recommend the codecs [01:09] jbicha: I don't see the need to have that in Xubuntu, if Ubuntu wants that, I guess that's fine for them [01:09] I think people install u-restricted-extras thinking that it will get them all the codecs they'll ever need [01:09] its purpose isn't one-off codecs, but ones that work in many applications [01:11] also, what happened to -plugins-ugly? [01:12] it's still there, it just didn't show in the diff http://paste.debian.net/38624/ [01:12] jbicha: also, adding that codec will pull in chromium on everyone's system [01:13] no, we fixed that today :) [01:13] I think that's wrong [01:14] the codecs do depend on chromium specifically to work [01:14] it was bug 1208518 otherwise things wouldn't be right for Lubuntu [01:14] bug 1208518 in chromium-browser (Ubuntu) "Don't have the codecs depend on chromium-browser" [Undecided,Fix released] https://launchpad.net/bugs/1208518 [01:15] that's wrong, the codecs should depend on chromium and the restricted package shouldn't [01:16] a pal is getting package authentication errors, http://pastebin.com/jgms472S -- I can't replicate with my 12.04 VM.. any ideas? [01:16] restricted extras is for system wide codecs [01:16] Riddell: is it fine if kubuntu-restricted-addons recommends the chromium extra codecs? [01:17] sarnold: I think that can happen if the user and mirror are updating at just the right time [01:18] micahg: ah :) [01:18] but I"ll let someone else verify [01:18] micahg: so, something like the Releases is updated but the Releases.gpg isn't? === freeflying is now known as freeflying_away [01:18] I would think if those files were updated before the new debs were or vice versa [01:19] hrm, that's not right [01:19] sorry, don't remember exactly why ATM, just recall it being transient for me on occasion [01:20] micahg: funny you say transient, that's what he's reporting -- it works correctly now for him. go figure. [01:20] micahg: thanks :) [01:22] overly strict dependencies cause problems for Ubuntu flavors, like bug 1191522 and bug 1222155 [01:22] bug 1191522 in libsignon-glib (Ubuntu) "Don't have libsignon-glib1 depend on signond" [Undecided,Fix released] https://launchpad.net/bugs/1191522 [01:22] bug 1222155 in lightdm (Ubuntu) "Installing gnome-session-flashback also installs Unity" [Undecided,Fix released] https://launchpad.net/bugs/1222155 [01:24] and I really don't think people that install *buntu-restricted-extras care if they have an extra 700ish KB installed on their computer [01:26] jbicha: libraries depending on programs is usually wrong, since they can be consumed by many things, in this case, this codec can only be used by chromium [01:26] s/by/with/ [01:27] that seemed to be the sane fix in both the cases above [01:29] actually, the codecs package should enhance: chromium-browser, not depend TBH [01:30] so, bottom line is that I think it violates the spirit of the package, if the Ubuntu desktop team and other flavors want it, fine, I don't think I want it for Xubuntu [01:33] ok, I'll keep xubuntu separate for now and add the gstreamer 1.0 codecs to the xubuntu package [01:33] jbicha: thanks [01:36] I'll file a bug to add Enhances: chromium-browser on the codec packages === _salem is now known as salem_ === freeflying_away is now known as freeflying [03:06] cjwatson: Thanks. === salem_ is now known as _salem [03:32] is your driver not coming until 9:30? === freeflying is now known as freeflying_away [03:32] Tower 8 7b [03:32] lol, wrong channel === sj__ is now known as sj === freeflying_away is now known as freeflying [04:31] pitti, hallyn: so https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=65aafb1e7484b7434a0c1d4c593191ebe5776a2f finally got merged into Linus' branch. A bit late for 13.10 but for 14.04 we'll have apport working properly with containers (the kernel will send the crash to apport within the container). === gavin__ is now known as gavinguo [06:39] 3 [06:40] good morning === smb` is now known as smb [07:39] g'day [07:40] oh btw i think the lts-s repo is complete, so feel free to test and find any bugs :P [07:58] cool [07:59] slangasek, Got the mountall bug updated. Sorry, I was gone already yesterday when you asked about stuff. [08:23] slangasek: re: ifupdown, reblasted - wait and see what happens. [08:26] slangasek: re: blasting livefs to usb - kpartx -a setup loopmounts of the cd and blast just the contents (first iso partition) of the cd onto the usb partition you want and then copy across the EFI folder from the other one, that should just work. [08:37] doko_: debian #688251 is interesting [08:37] Debian bug 688251 in debian-policy "Built-Using description too aggressive" [Normal,Open] http://bugs.debian.org/688251 === seiflotfy_ is now known as seiflotfy === gusch is now known as gusch|lunch [10:41] chrisccoulson: firefox-testsuite needs to be updated to depend on fonts-liberation, not ttf-liberation - I think I have commit access but which branch(es) do I need to commit to? === freeflying is now known as freeflying_away === MacSlow is now known as MacSlow|lunch === freeflying_away is now known as freeflying === tkamppeter_ is now known as tkamppeter === freeflying is now known as freeflying_away [12:26] Good morning [12:27] stgraber: oh, great! that's a trivial patch, too late to cherrypick? [12:27] stgraber: but anyway, not that urgent I guess; nice to have it for the LTS === _salem is now known as salem_ === MacSlow|lunch is now known as MacSlow [12:48] stgraber: yeah i saw akpm forwarded that yesterday [12:49] \o/ === rickspencer3_ is now known as rickspencer3 === freeflying_away is now known as freeflying === didrocks1 is now known as didrocks [13:58] cjwatson: hey, so I just updated my laptop, got a new grub and I'm now booting to a grub prompt [13:59] cjwatson: any known issue there before I go dig into what happened? === kentb-out is now known as kentb [14:01] cjwatson: hmm, so nothing obvious, /boot/efi/EFI/ubuntu/grub.cfg points to the right UUID, grub.cfg exists and appears correct, so I'm not sure why it didn't get loaded... [14:02] cjwatson: grub itself works fine besides the fact it didn't load the config as I could boot my machine by giving it the location of the kernel and initrd just fine [14:02] ... and you are not using saucy-proposed? [14:02] * stgraber reboots to run /boot/efi/EFI/ubuntu/grub.cfg by hand, see if that's enough [14:02] xnox: I'm definitely not [14:02] it migrated [14:02] Laney: oh, ok. [14:02] I know we had a broken grub in proposed but I hope that's not what we let through to saucy... [14:02] * stgraber reboots [14:04] cjwatson: so calling "configfile /EFI/ubuntu/grub.cfg" makes it work just fine [14:05] cjwatson: so it looks like it's the builtin logic to find the grub.cfg on the EFI partition that's broken [14:06] stgraber: Huh, that did change but it worked fine for me [14:07] No doubt r2344 but how ... [14:07] stgraber: What's $prefix? [14:07] cjwatson: http://paste.ubuntu.com/6097168/ [14:07] stgraber: Also, grub.cfg should not be in /EFI/ubuntu/ [14:08] It's meant to be read directly from /boot/grub/ - no? [14:09] cjwatson: my understanding was that grub2-signed would load a minimal grub.cfg next to it on EFI which tells it where to look for the real grub.cfg in /boot/grub on the root partition [14:09] anyway, that's how things got automatically setup on this machine [14:09] Oh, blah, I missed a bit in grub-install [14:14] Bug #1220894 (found in beta1 testing) has been fixed and commited. Can we get lp:ubuntustudio-default-settings rereleased/uploaded? [14:14] bug 1220894 in ubuntustudio-default-settings (Ubuntu) "ubuntustudio-default-settings does not make sure a grub rebuild is called after install." [Undecided,Fix committed] https://launchpad.net/bugs/1220894 [14:15] This must have always been sort of broken - I don't see how this ever worked in the case of, say, SB with /boot on LVM [14:16] Because then AFAICS there wouldn't be a /boot/efi/EFI/ubuntu/grub.cfg [14:16] arges: check out qemu-kvm -ppa2 version for precise in ppa:serge-hallyn/lucid-kvm-test. Our qcow2 code is just getting scarier and scarier [14:17] hallyn_: what changed? [14:17] stgraber: When you're at the GRUB prompt, can you tell me what prefix is set to? [14:17] cjwatson: thankfully I don't have /boot on LVM on that box (I do on a UEFI precise box where I found that d-i bug which would install grub-pc instead of grub-efi in such case :)) [14:17] cjwatson: sure, rebooting quickly [14:17] arges: somebody was getting kvm corruption with heavy load on qcow2 [14:18] the commit in rhel5 to fix it has (as usual) 10-20 patches it needs before it === kentb is now known as kentb-afk [14:18] hallyn_: ouch. is there a bug # to track? [14:19] arges: bug #12239071223907 [14:19] Error: Launchpad bug 12239071223907 could not be found [14:19] feh [14:19] bug #1223907 [14:19] bug 1223907 in qemu-kvm (Ubuntu Precise) "qemu-kvm crashes in qcow2 code" [High,Confirmed] https://launchpad.net/bugs/1223907 [14:19] wow : ) [14:19] cjwatson: (hd0,gpt1)/boot/grub [14:19] stgraber: ok, and what's GRUB's name for the EFI system partition? [14:19] (hd0,gpt1) or something else? [14:20] (hd0,gpt1) is indeed the EFI boot partition [14:20] but /boot/grub doesn't exist on it as that's a SecureBoot system [14:20] It wouldn't anyway [14:21] ah right, the modules are in /grub when booting an unsigned grub, not /boot/grub (if my memory is vaguely correct, only have one UEFI non-SB system around) [14:21] The modules are still in /boot/grub unless you have a separate /boot partition - but they're not on the EFI System Partition [14:21] hallyn_: so if we have a definitive test (we could use the same one in bug 1189926), we could try to bisect again [14:21] bug 1189926 in qemu-kvm (Ubuntu Quantal) "data corruption in storage attached to VM using KVM" [High,Fix released] https://launchpad.net/bugs/1189926 [14:22] hallyn_: or are you planning on backporting those 10 or so patches? [14:22] arges: no, for right now i just hand-applied the spirit of the patch that fixed it in rhel [14:22] as you'll recall we tried backporting all those patches before [14:23] we just end up at a newer release altogether (or with broken patch hunks) [14:23] yeaa [14:23] thing is my impression has alwasy been that upstream felt qcow2 was fragile. [14:23] too bad it has such great features so ppl keep using it :) [14:24] i wonder if we can push ppl onto qed using the cloud archive [14:24] which is good, hopefully more bugs get ironed out by people using it === blitzkrieg3 is now known as jmleddy [14:25] well let's see if this fixes it for that guy. some of the other patches which i didn't backport sounded frivolous, but some didn't [14:25] arges: have you seen those compiz patches from nvidia at all? [14:25] arges: I know you're not assigned to the case anymore [14:25] stgraber: I'll just revert the bit that broke you for now, and figure out the rest later [14:25] cjwatson: ok, thanks [14:25] hallyn_: fyi. also working on the qemu-kvm perf issues after migration. I think i have a patch and I can reproduce it. so hopefully this is a fix for that [14:26] arges: I'm confused. which "this" and which "that"? [14:26] jmleddy: so I've actually tested teh latest saucy and the corruption with powermizer at teh lower setting seems to be gone [14:26] arges: oh, great news [14:26] jmleddy: so we think something was fixed between 0.9.9 and 0.9.10 [14:26] arges: you've also been following http://lists.gnu.org/archive/html/qemu-devel/2013-07/msg01856.html ? [14:26] arges: of compiz? [14:26] jmleddy: yea compiz. [14:26] (i thought you'd been involved in it, but i don't see you in the thread list) [14:27] hallyn_: yea that's where i foudn the patch [14:27] okay, that much is mark's problem ;) [14:27] hallyn_: somebody identified a patch that seemed to fix perf for him [14:27] arges: ok :) [14:27] but I'll report that later compiz fixes the problems to Nvidia [14:28] jmleddy: cool thanks [14:29] stgraber: uploaded, thanks for the heads-up [14:29] cjwatson: np, thanks for the quick fix. I'll review the upload in Unapproved as soon as it gets in there. [14:35] Hi! Can someone look at this bug - https://bugs.launchpad.net/nautilus/+bug/885989. Why nobody wants to approve working patch to fix this bug? [14:35] Launchpad bug 885989 in nautilus (Ubuntu Precise) "white screen on second monitor when using two xsessions" [Low,Triaged] [14:48] cjwatson, hi, just installing 13.10 onto a new MacbookAir and after install and reboot we get dropped to a grub prompt. Any ideas how to fix [14:48] cjwatson, we are using grub-efi [14:48] ChickenCutlass: see conversation above with stgraber [14:48] probably the same thing [14:49] ChickenCutlass: for now, run "configfile /EFI/boot/grub.cfg" that should make it boot until you update to the fixed package (should be available in ~1h) [14:49] stgraber, ah ok [14:50] stgraber: bit more, it'll need a matching grub2-signed in order to migrate [14:50] ChickenCutlass: hmm, that's wrong [14:50] ChickenCutlass: "configfile /EFI/ubuntu/grub.cfg" [14:50] stgraber, trying now [14:51] stgraber, that worked [14:51] cjwatson: yeah, I assumed as much (grub2-signed) but didn't check how long grub2 takes to build so that's why my estimate is a bit off === kentb-afk is now known as kentb === mzanetti is now known as mzanetti|food === mzanetti|food is now known as mzanetti [15:16] xnox: so I've noticed recently that some of these give-backs are resulting in existing branches getting tags moved; do you know the story behind that? [15:17] xnox: kpartx -a> check. That's roughly what I wound up doing, though I rather wish there were a nicer way to edit in-place :) [15:19] slangasek: define "in-place" - as in somehow dd the iso onto the usb stick, yet still have a partition table and partitions that one can edit? [15:19] slangasek: about tags getting moved, no idea how or why. [15:24] xnox: yeah, dd iso to the usb stick, then edit the partition table... doesn't work because none of the tools are happy with overlapping partitions [15:27] slangasek: quite, e.g. everything complains that it's a broken GPT table and doesn't let to do anything with them. =/ i typically have to dd from dev/zero to get usb back into partitionable condition. [15:27] oh, it doesn't seem that bad here :) [15:29] slangasek: well - Disks gui, parted, gparted, fdisk all don't like amd64 iso dd'ed onto the usb stick =/ [15:29] am I doing something very odd?! [15:29] xnox: ah; sure, I used the commandline tools and they complained about it, but I was able to bypass the warnings [15:31] gotcha. I don't like those warnings [15:49] YokoZar: Do you care about keeping dssi-vst in the archive? It's been removed from Debian and seems to have no rdeps in Ubuntu. [15:49] cjwatson, i kind of feel like i have ot tell you, after reading and understanding germinate more, germinate is really cool and helpful. [15:49] made it easy to figure out "what if i add this package, what does that pull in" [15:50] smoser: Cool. It can be a bit obscure :-) [15:50] i guess the one comment i'd have is in pretty table output. [15:50] is nice for humans. [15:50] Yeah, I've been meaning to do a machine-readable output format for some time. [15:51] yeah. [15:51] if there was a prettytable reader [15:51] (which surely someone has done) [15:51] One of these days ... Nowadays LP just uses the underlying libraries directly [15:51] then it'd be ok. but the totals at the end would make it at least tricky [15:51] Though partly that's because it's significantly faster for its needs to do it that way [15:52] (Not because the table formatting is expensive or anything, but because it was duplicating a lot of work in the multiple runs it used to do) [15:52] oh. its was quite reasonably performant for me. ran on a caonnistack instance (so fast mirror access) but the whole run would only take maybe 30 seconds. [15:54] yeah, for only one run it's fine [15:54] LP does * [15:59] ogra_: can you do something like image stats [15:59] ogra_: thta just shows the diff from /current to /current ? [15:59] asac, which kind of image stats ? [15:59] ogra_: QA would like to do that [15:59] sure [16:00] cjwatson: didrocks and ricmm seem to had delays in publishing to proposed [16:00] my script can take any version number you want [16:00] (as long as it exists) [16:00] cjwatson: not sure if they talked to you ... just want to ensure that no infraa parts art down etc. [16:00] asac: done now [16:00] ogra_: http://people.canonical.com/~ogra/touch-image-stats/ [16:00] yeah [16:00] asac: hence the no ping, while we speak, it happened [16:00] ogra_: you called it image stats :) [16:00] ogra_: i want the same for current (not proposed) [16:00] asac: monitoring the migration to release now [16:00] ogra_: that will help QA to do focussed efforts [16:00] asac, they should just grab the script and call it with the two image stamps they want to compare [16:00] didrocks: allright [16:00] cjwatson: ignore [16:01] ogra_: can you set it up so it automatically happens? [16:01] as long as there is a manifest it will show the diff [16:01] ogra_: i know tey can do it, but i want to have a reusable link [16:01] that we can share and everyone sees the same [16:01] asac, tricky, since i dont have a history about which image was when /current [16:01] we would have to store that somewhere [16:02] ogra_: ok ... give the script to jfunk please [16:02] in a bzr branch (not a paste) [16:02] :) [16:02] ogra_: or give it to jfunk and pitti :) [16:02] pfft ... it is in the dir :) [16:02] then i feel its understood [16:02] no idea [16:02] just tell jfunk and pitti so they can sort it :) [16:03] pitti, jfunk, http://people.canonical.com/~ogra/touch-image-stats/ see the README and the compare-manifests script [16:04] FWIW I see nothing particularly untoward in publisher runtimes or anything recently === lfaraone_ is now known as lfaraone [16:10] cjwatson, does mark-current long its actiona anywhere ? [16:10] *log [16:10] kees: please could you review https://code.launchpad.net/~dannf/cpu-checker/arm-cortex-a15/+merge/171667? Looks like dannf has updated his MP following your review. Could you check/merge if appropriate, please? [16:10] so that i could have an overview which images were /current [16:12] *actions ... sigh [16:18] ogra_: cdimage/log/mark-current.log [16:18] * ogra_ hugs cjwatson [16:18] thanks ! [16:29] cjwatson: Re: dssi-vst. It is hard to make it a dep or include on an ISO because it is effectively 32 bit only. So to use it on a 64 bit machine it has to be installed after the ISO is installed and then of coarse it pulls in all the required 32 bit libs (and wine I think)... windows :P [16:30] cjwatson: That of course says nothing about if it should be kept or not. Hopefully falktx's stuff will get put in debian and replace it. [16:31] OvenWerks: I'll keep it for now, but if you folks decide you want to keep it in Ubuntu, it'd be nice if somebody could take up maintenance in Debian too [16:31] Since I think it was just removed due to build failures that we'd fixed ... [16:32] Ah, good to know. I don't know if anyone uses it to be honest. [16:53] slangasek: I think the pycompile patch in ubuntu-release-upgrader (for bug 689615) could also be dropped. agreed? [16:53] bug 689615 in python-defaults (Ubuntu Maverick) "pycompile fixes needed for maverick" [High,Fix released] https://launchpad.net/bugs/689615 [17:02] bdmurray: yes, since precise is the oldest version this upgrader will be used from, that can be dropped. I was being conservative in dropping fixes for upgrades from releases that you could *theoretically* still use this upgrader for; but if you want to be more aggressive, we should draw the line at anything older than precise [17:04] slangasek: okay, I'll keep that in mind thanks [17:23] barry, ScottK: Do you know how important it is to ship foo.egg-info/ bits in python* packages? [17:24] I just noticed that the python-autopilot doesn't have it, and nothing is blowing up apparently === bfiller is now known as bfiller_lunch [17:28] pitti: in practice (as you saw) it's probably not that important, unless you have entry_points (e.g. console_scripts) or other things that the runtime might want to introspect [17:28] barry: ok, thanks [17:40] Hello! I'm about to dput a new checkbox package for Ubuntu, it contains only bug fixes. Is this OK since we're past FF? (just wanted to double-check before messing things up) [17:52] roadmr: unless you consider things like "change the UI" and "we are missing this feature" as bugs, then yes [17:53] pitti: heheh :) I do have one minor string change, my rationale is that it's OK since we're not past UI Freeze yet [17:53] roadmr: ah, yes [17:54] pitti: other than that, I'm sure there are no new features or UI/behavior changes [17:54] roadmr: that's fine then [17:54] pitti: thanks, I'll proceed with the upload then! === bfiller_lunch is now known as bfiller === xxiao is now known as ausjke === sam113101 is now known as sam113101_afk === sam113101_afk is now known as sam113101 [20:50] slangasek: do you know how the verification for bug 1179781 may have happened before your comment regarding uploading it to precise? [20:50] bug 1179781 in curl (Ubuntu Raring) "If-Modfied-Since undhandled case causes apt lists corruption with https repositories" [Medium,Fix committed] https://launchpad.net/bugs/1179781 [20:51] sil2100: please don't touch the saucy platform, misc, or unity8 stacks for now, I'll handle them [20:51] we did some changes that will majorly break stuff if it's touched [20:52] cyphermox: ACK, doing now packaging and coding anyway [20:53] kirkland, roaksoax: Do either of you have some time to talk about testdrive [21:15] stgraber, I just built my own kernel -- how do I sign it to get the efi.signed version? === ChickenCutlass_ is now known as ChickenCutlass [21:26] ChickenCutlass: you can't [21:27] ChickenCutlass: you'd have to sign it with the Ubuntu SecureBoot key which you obviously don't have access to [21:27] ChickenCutlass: but grub2 will be happy to boot an unsigned kernel, so just do that [21:27] stgraber, ok, right - that is what I am doing [21:32] stgraber, what file do I edit to make grub used the non signed kernel permanently [21:32] stgraber, so I don't have to edit each boot [21:33] ChickenCutlass: depends how you made your kernel. If it's a .deb that you installed, I'd expect grub to just do the right thing. [21:33] ChickenCutlass: You mean your kernel is a lower version than others, and thus isn't "entry 0" on the list? [21:34] ChickenCutlass: I had a quick look at the code and it'll only use .efi.signed if the file exists, otherwise it'll use the standard vmlinuz-... name [21:34] ChickenCutlass: If update-grub put it elsewhere than 0, you can edit /etc/default/grub and change GRUB_DEFAULT=0 [21:34] stgraber, so I built my own kernel deb -- installed it. but grub.cfg still points to the .signed [21:34] stgraber, ah [21:34] Oh, if there's a .signed that matches your kernel version but isn't your kernel, that would be a problem. :P [21:34] stgraber, so just delete the signed [21:35] bdmurray: 1179781> looks like the verification tag was added in error [21:36] ogra_: Wait, I was only half-listening. Uploaded "just now" to where? The archive? [21:36] slangasek: okay, that was my thought too [21:36] but you had at least a funny expression on your face :P [21:36] ogra_: he's been like that all day :) [21:36] I've been arguing with computers and attempting to listen to meetings at the same time. It's hard work. [21:37] infinity, unity-mir just finished building, livecd-rootfs , lxc-android-config and ubuntu-session-touch were just uploaded right now [21:37] ogra_: Oh, I thought you mentioned something abot upstart-app-launch, which is what I was looking for. [21:37] infinity, i said, that it might not make sense if you wait that long [21:38] And ubuntu-session-touch doesn't exist. :P [21:38] upstart-app-launch will be pocket copied from the daily-next PPA [21:38] then re-order :P [21:38] ubuntu-touch-session :P [21:38] * infinity goes back to arguing with arm64 buildds instead. [21:39] * ogra_ always has a hard time remembering the package names he made up when drunk :P === salem_ is now known as _salem === kentb is now known as kentb-out === lifeless_ is now known as lifeless [22:39] jbicha: the debian/copyright for gnome-weather is incomplete; it's missing license/copyright info for all of data/*. It also doesn't cover libgd/*, but I'm hoping you're not using that embedded copy at build-time... [22:46] slangasek: libgd is included in GTK 3.9; projects like gnome-weather and gnome-documents bundled the library before then as it was still under heavy development (it's not designed to be packaged separately by distros) [22:47] gnome-weather was rejected in Debian new because the images it uses several CC-BY 2.0 and CC-BY-SA 2.0 images [22:48] since CC-BY is generally *more* permissive than CC-BY-SA I would think Debian should accept art with that license but I never braved debian-legal to try to get that official === freeflying is now known as freeflying_away [23:59] Can tests have cross-arch dependencies?