=== robert__ is now known as nudtrobert [01:52] ^^ this gets the cross-compiler in sync with gcc-4.7-source for release, which is probably a good idea; I actually want them in sync to try to pin down an issue where gcc-4.7 is reportedly behaving differently for certain code for native vs. cross building [01:53] (and the native compilation is giving buggy results, but that's beside the point... we don't know that the compiler itself is at fault) [01:54] oh, is the final freeze still advisory for universe? ok then [03:03] Is there some reason that there arent yet Ubuntu GNOME RC builds? [03:05] darkxst: We created the milestone, and are just letting them cron in at their usual time. [03:06] darkxst: Which for ubuntu-gnome, appears to be 15:32 UTC... A bit of a wait. [03:06] darkxst: I can do a manual spin for you, if you have confused testers. [03:08] darkxst: Spinning you a set now. === henrix_ is now known as henrix [04:40] infinity, thanks ;) [04:51] infinity: Thanks. [06:04] Can someone do a quick review of that gdb upload I just made? [06:33] infinity: doh - right, I forgot about the efi binary new, didn't I [06:33] infinity: gdb> looking [06:33] or not [06:49] slangasek: Does anything read /lib/init/fstab other than mountall, and ever before /etc/groups is available? [06:50] infinity: well, I had a conversation with stgraber and hallyn about possibly making lxc read it, but no. Also, how can /etc/group ever not be available? /etc must be on the rootfs [06:50] slangasek: Well, it could be unavailable if something pulled that file and mountall into an initrd, for instance. [06:51] slangasek: (Which doesn't ever happen currently, I assume?) [06:51] slangasek: Anyhow, was just asking because of the "gid=tty" instead of "gid=5" in there, which would go awry if one couldn't resolve "tty" to anything useful. [07:03] infinity: yes, no mountall in initramfs [07:04] I'm beggining to think maybe I should revert the pt_chown change for now. [07:04] And maybe fix mount to add gid=5,mode=0620 to the "defaults" set. [07:05] There's this curious feature where, even if your base system is booted and mounted correctly, if you "mount -t devpts devpts-foo foo/", all instanced of devpts (including the base system) pick up the new defaults. [07:05] And lose the gid/mode bits. [07:05] Which has always been true, but then pt_chown picked up the slack for this misconfiguration in the past. [07:09] slangasek: On the other hand, Fedora already pushed this out as a security update a couple of months ago, and basically said "if stuff breaks, mount /dev/pts correctly, derp" in their announcement. :P [08:27] infinity: if you are still around, the mir one is for this touch image ^ (it's just a build-dep of xorg, but the fix is for touch only), not seeded anyway [08:28] stgraber: poke poke [08:28] didrocks: I'm not here, and you can't prove it. [08:29] didrocks: Oh, also: congrats. Where was my invite? :) [08:29] infinity: ahah, thanks a lot! :-) [09:29] hey, i accidentially triggered an ubuntu-gnome build, hope i didnt cause havoc with this [09:32] slangasek: Probably ought to be changed, yes. Can you file a bug and I'll do it next cycle? I don't think I should mess with it now for 13.10 [09:35] ogra_: They'll probably live. There's going to be a cronned one today too, anyway. [09:39] infinity: well i have two installer bugs fixed now, and if/when ubiquity upload will happen today. I'd love to respin all ubiquity images and test them before departing for the weekend. [09:39] unless too late. [09:40] xnox: Not too late. I'm sure this won't be the last installed upload of the cycle (though one can always hope). [09:40] s/installed/installer/ [09:40] xnox: The freeze pretty much explicitly doesn't apply to installers, since this is when we get all the good/juicy bugs. [09:42] infinity: oh and base-files are not in yet, so yeah we will be respining =) [09:42] Indeed, I'll upload that on the weekend, probably. [09:43] Or maybe today sometime. [09:43] Since I lost a bunch of my weekend to flying in the wrong direction. [09:43] Silly timezones. === tkamppeter_ is now known as tkamppeter [10:19] ↑ and ↓ indicators + unity stacks, fully tested on desktop, best autopilot results ever so seems pretty nice. indicators also tested on touch (manually) for the parts that are affected. [10:20] of course the best AP tests comes down to unity7 team fixing AP false alarms, but still, makes for nicer numbers [10:21] Mirv, so are there any issues with landing it? [10:22] mhr3: not that I know of, as mentioned good AP results, manual testing on touch fine. [10:23] Mirv, cool, pls ping me if there's any problem with it [10:23] mhr3: sure. if release team has any worries, they'll tell us here. [10:30] cjwatson: http://paste.ubuntu.com/6221814/ [10:31] those are components that didnt make it it seems. [10:31] cjwatson: what are the options? what do you need from us? [10:34] It was only 15 minutes ago :) I'll review [10:34] cjwatson: ok cool. let me know. especially if we need to divert one of those to -updates or wait much longer [10:34] thanks [10:35] (I'm mostly fixing a problem with the publisher right now, so if anyone else can help ...) [10:35] cjwatson: what kind of help? I assume that is "anyone from release team"? [10:35] Yes [10:35] asac: Is ido one of yours as well? [10:35] cjwatson: so thats all that is comnig from CI [10:35] we unfortunately have still desktop stuff in CI [10:35] fortunately actually, but in this case a bit awkward [10:36] asking didrocks [10:36] ido has a desktop bug fix [10:36] it doesn't impact touch [10:36] cjwatson: ido is also coming from CI, but its pure desktop only [10:36] seb128: Do you know if the fix in ido affects anything in screenshots? [10:37] It's fine otherwise [10:37] cjwatson, it means the indicator-session avatar icon being properly aligned if you use a non default font size [10:37] I doubt we have any screenshot of the bug [10:37] Ah, so not screenshots [10:37] so yeah we missed ido [10:37] means->makes [10:37] asac, I'm glad you uploaded it :p [10:37] that bugs annoys me ;-) [10:38] asac, btw "unfortunately have still desktop stuff in CI" ... does it mean you plan to drop CI for desktop components? [10:38] cjwatson, thanks [10:41] /37 [10:57] asac,didrocks: all accepted now [10:57] cjwatson: thanks a lot :) one worrying things less :) [10:57] cjwatson: awesome! [10:57] what an amazing day [10:57] i will have to buy flowers for my wife still :P [10:58] * asac puts that on his "nice to have TODO list" :) [10:58] asac: "nice to have" only? [10:58] I think qtdeclarative-opensource-src is unseeded, right? [10:58] * didrocks checks [10:59] can't be unseeded if it's in those packagesets [10:59] but it looks fine, is it OK from the touch POV? [10:59] cjwatson: yeah, it's fixing the "unity8 is taking 100% of GPU" :) [10:59] CPU* [11:00] thanks colin! [11:00] ok, now touch-only uploads, all tested, fine, let's publish them [11:01] didrocks: yeah i should prioritze that more appropriately :) [11:03] heh [11:12] infinity: Did you retry the failures due to lp-buildd vs. pt_chown? [11:13] cjwatson: The one or two I knew about. Were there many? [11:13] I only knew of the ones you told me about :-) [11:14] Fair. I did those. :P [11:44] thanks seb128 cjwatson asac didrocks :) [11:45] I even got to eat proper food today, so I'm even more happy [11:46] Mirv: lucky you! :) [11:48] Mirv, yw ;-) [13:16] openstack just pushed out a second rc for glance with a number of critical and high bug fixes - zul preparing and reviewing now [14:07] glance [14:07] oops [14:43] I think platform-api is what is freezing location-service ^ === doko_ is now known as doko [14:45] ubiquity ^ diff will be large, but it's really one bugfix per commit on lp:ubiquity. Also note that majority of the diff is debconf-updatepo, which means unfortunately that many strings were not possible to translate up to now. [14:48] cjwatson: any reason why kubuntu-active ISO's aren't built? [14:51] (thanks to whoever accepted location-service ;)) [14:51] shadeslayer: I believe kubuntu-active was marked disabled in the ISO tracker, because someone didn't want them to be in milestones/release. Has that position changed? [14:53] shadeslayer: Active wasn't in any of the alphas or betas, only dailies. [14:53] shadeslayer: check the build failures [14:53] shadeslayer: it's been failing for a while [14:53] I don't know where :( [14:53] I hadn't had a chance to look [14:53] Oh, did you literally mean "not built" rather than "not posted to the tracker"? [14:53] http://people.canonical.com/~ubuntu-archive/livefs-build-logs/ somewhere [14:54] infinity: yes [14:54] ahhh [14:54] contour [14:54] cjwatson: infinity could you trigger a build now? [14:54] should be fixed [14:55] as of when? today? [14:55] I pushed fixes early today [14:55] https://launchpad.net/ubuntu/+source/plasma-mobile/1:0.5-0ubuntu2 [14:55] doing [14:55] thx [14:56] I don't see how that would fix it. [14:57] shadeslayer: If both contour and plasma-active are being installed, you've not improved the situation by adding an unversioned Breaks. [14:57] infinity: contour is no more [14:57] dropped it from the seed [14:57] removed from archive [14:57] https://launchpad.net/ubuntu/+source/kubuntu-meta/1.291 [14:58] https://launchpad.net/ubuntu/+source/contour/+publishinghistory [14:58] I guess we'll find out [14:58] shadeslayer: Ahh, that was a bit more informative than the plasma-mobile link. ;) [14:58] :) [14:59] sorry about only providing half the info [15:18] Hey, I've added a block ubuntuone-credentials hint to test it while it's in -proposed (it's not under cu2d CI) [15:18] (dobey is the uploader and has tested it "upstream" / in trunk) [15:18] (it's only seeded in touch) [15:19] thanks [15:33] stgraber: updating ubuntu-meta does "* Added shotwell to desktop-recommends [powerpc]" yet i was just removed "* Removed shotwell from desktop-recommends [powerpc]" by your last upload. is that correct? [15:36] xnox: Looks like it was removed from powerpc at one point and then restored [15:36] cjwatson: ok. uploaded. [15:36] maybe? [15:37] it's recommends and powerpc.... maybe it was removed/FTBFS at the time of last upload?! [15:37] Yeah, https://launchpad.net/ubuntu/saucy/powerpc/shotwell, deleted from release entry #7 [15:38] cjwatson: otherwise it's a FTBFS fix. [15:38] hi [15:38] 17:36 -queuebot:#ubuntu-release- Unapproved: ubuntuone-credentials (saucy-proposed/universe) [13.10-0ubuntu1 => 13.10-0ubuntu2] (ubuntuone) [15:38] what does the ubuntuone package set mean? [15:38] does it have special rules for reviews or something? [15:38] I think it's for upload perms [15:38] can i bring the case of bug #1206268 to your attention again [15:38] Launchpad bug 1206268 in webbrowser-app (Ubuntu) "[MIR] unity-webapps-qml" [Undecided,Fix committed] https://launchpad.net/bugs/1206268 [15:39] don't think so, just uploads perms & release-bug tracking / owners. [15:39] Heyo folks! There is a change that dbarth and folks would like to make to the ubuntu-sdk package. ^ This would involve dropping a Recommends from that seed, and then changing qtmultimedia to stop building a certain package only referenced by the sdk metapackage. How doable is that (we have branches, just looking for approval) at this stage? [15:40] mterry: are all changes required reviewable somewhere? is it all the branches attached on the bug report? [15:40] mterry: can you comment links with merge proposals against seeds and/or all packages involved? [15:40] (comment on the bug that is) [15:41] stgraber: There might be some possible improvements to the autoaccept script if you like to: ubuntuone-credentials is only seeded in touch, but is stuck in queue because it's in a special upload packageset (ubuntuone) [15:41] cause it's hard to see what the scope of changes is.... [15:41] xnox, sure. let me link [15:41] lool: I'll whitelist that one [15:42] mterry: should Mirv also comment on the bug directly, with the plan he proposed earlier in the thread? [15:42] stgraber: thanks [15:43] lool: added to whitelist now [15:43] can someone please let livecd-rootfs through ? [15:44] (first is adding click packages to the manifest, second is a fix for a serious touch bug) [15:45] dbarth, the plan is basically "land these three branches" [15:45] looking [15:45] xnox, three branches added to bug in comments [15:47] robru, ^ btw [15:47] ogra_: I'm a bit concerned by that manifest change, won't that break Ursinha's script and anything that expects packages listed in the manifest to be avaiable from the archive? [15:47] stgraber, it wont break my script, Ursinha's is still not merged afaik [15:47] stgraber, and jibel picks his data from my location [15:48] stgraber, in any case we need that info in the manifest [15:48] well, I'd argue we should have a separate manifest file for click [15:48] sure, we can do that later [15:48] cjwatson: I'd like your opinion on this if you have a minute? [15:49] stgraber, this is the quick variant to get it in before release [15:49] stgraber, i'l happily teach cdimage about a new file and how to publish it in T [15:49] stgraber: I recommended the approach ogra_'s taken - I think manifest consumers should be prepared for the packages to be either .deb or .click [15:50] mterry: right [15:50] If it breaks anything then surely it'll be straightforward to fix [15:50] try: except: [15:50] the format is the exact same too [15:50] cjwatson: is there an easy way for something reading the manifest to determine whether it's a .deb or .click? [15:50] [15:50] stgraber: "does it exist in the archive as a .deb" [15:51] I mean, it has to go looking for the changelog anyway, it can just check, no? [15:51] stgraber, most likely the namespace will be com.ubuntu.foo.bar.* [15:51] ogra_: that's definitely not a reasonable thing to check for in the consumer [15:51] if it would be helpful to prefix the package name with "click:" or something, that would be possible [15:51] sure, it's just going to be tricky to know whether something is a click package or a package coming from a PPA (even though we're not supposed to ever use PPA, the past as showed that we sometimes do) [15:52] I dunno, I guess a separate file would be possible, I'd just like to avoid a proliferation of output files [15:52] I think it's confusing [15:52] in any case a separate file is nothing for the week before release imho [15:52] I think I'd be fine with a click: prefix as it shouldn't break ogra_'s script and will let other parsers easily detect click and either act on it or ignore them [15:52] i dont want to do big plumbings in cdimage that close to release [15:56] xnox: do we really have to introduce a new string in ubiquity that close to release? [15:57] xnox: hmm, well, it doesn't really matter since none of the ubuntuone strings were translated anyway... I guess it's going to be one of those releases... [15:57] stgraber: note that none of the other strings introduced this cycle in ubiquity didn't make it into .pot and were not translated due to not running debconf-updatepo this cycle at all. So in fact I am introducing new strings anyhow in that plugin. [15:58] stgraber, so do you let it in ? i urgently need the second fix [15:58] oh [15:58] stgraber: it only affects ubuntu images, as nobody else seeded ubiquity-ubuntuone plugin. [15:58] * ogra_ notes the mail [15:58] sigh [15:58] ogra_: I'll accept your /system/bin/sh change and reject the other [15:58] k, thanks :) [15:58] stgraber: yeah, it is shame that translations will be bad..... (note U1 api only returns back errors in english =/ ) [15:59] ogra_: hmm, actually I can't since you stacked the /bin/sh one on top of the manifest one... (I was hoping it was the other way around and I could simply pull the old one from the reject queue) [15:59] ogra_: if you re-upload adding a click: prefix to the manifest entries, I'll accept it. Alternatively, just upload the /bin/sh fix and I'll accept that. [16:00] ok that has to wait 1h then ... [16:00] ogra_: why? [16:00] * ogra_ has meetings now [16:00] ah ok [16:00] i'll re-work the click part ... [16:00] since we need that [16:00] thanks [16:01] (i doubt any of the scripts can handle columns in the package name though) [16:01] ogra_: ping me once it's back in the queue if I don't notice and I'll make sure it's reviewed quickly [16:01] ogra_: well, either they really care about the source name existing and they'll fail either way or they don't and a click: prefix won't make any difference [16:07] ^ didrocks: signon-ui as requested [16:07] I'm reviewing it now [16:07] stgraber: thanks [16:09] stgraber, cjwatson http://paste.ubuntu.com/6223092/ does that look ok to you ? [16:10] cyphermox, is the ball rolling for webbrowser-app ? [16:10] ogra_: did you mean "while read line"? [16:11] alex-abreu: best I can tell you is it *did* for image 94, I don't know more [16:11] stgraber, not really, i tested that variant locally [16:11] ogra_: hmm, actually, both do the same, ignore me ;) [16:11] but i can tiurn it into a while loop :) [16:11] it's just the first time I see it done with for I believe ;) [16:11] stgraber, heh, now that i think about it, me too :P [16:12] ogra_: anyway, yes, that looks good to me [16:12] great, committing and uploading 2.194 then [16:13] ogra_: you know that you can just squash all the changes and upload 2.192 right? as none of those got accepted into the archive [16:13] cyphermox: are you familiar with the actual changes in that upload? [16:13] stgraber, i'm to lazy to roll back the bzr branch [16:13] ogra_: ok :) [16:13] its upstream committed [16:13] :) [16:14] stgraber: I tested that you could still create accounts and such, and ran the autopilot tests [16:14] cyphermox: I'm not entirely confident that this is a bugfix only upload based on the wording of the changelog and the number of qml changes, does that upload cause any kind of user visible change? [16:15] thanks cyphermox! [16:15] stgraber: not that I noticed [16:16] didrocks: do you know more? ^ [16:17] it's a bit hard to figure it out from the diff with all of those qml files being renamed/split/merged... [16:18] ogra_: just to confirm, that click change is in a touch specific section of auto/build, right? [16:18] stgraber, yep. ubuntu-touch and armhf even [16:19] ok, thanks [16:20] ogra_: there you go ^ [16:20] * ogra_ hugs stgraber [16:20] thanks :) [16:21] ogra_: "for line in read" runs the loop body once with $line set to "read" [16:21] ogra_: that needs to be a while loop instead [16:21] (I tested this ...) [16:21] cjwatson, eeek [16:21] will fix in a minute [16:21] $ (echo foo; echo bar; echo baz) | for line in read; do echo $line; done [16:21] read [16:22] I would be very surprised if any POSIX shell did otherwise [16:22] http://paste.ubuntu.com/6223144/ [16:23] stgraber: only on touch AFAIK [16:23] cyphermox: ^ [16:23] ogra_: those don't look prefixed to me [16:24] ogra_: that only works at all by sheer coincidence [16:24] And indeed will never be prefixed [16:24] http://paste.ubuntu.com/6223153/ [16:25] ogra_: What that does is cause the shell to evaluate the output of "click list" as a file name, attempt to redirect stdin from it, and then emit an error message saying that that (ridiculous) file name doesn't exist [16:25] ogra_: still broken [16:25] ogra_: you need: [16:25] ogra_: FWIW I tend to prefer: Chroot chroot "click list | while read line; do ...; done [16:25] Chroot chroot "click list" | while read line; do ...; done [16:25] ok [16:26] And indeed your second version is broken for the same reason - <"$(...)" is basically nonsensical shell :) [16:26] right and cjwatson's will actually work since he didn't forget a " :) [16:26] cjwatson: ISO seems to have built \o/ [16:26] shadeslayer: cool [16:26] ok, final attempt http://paste.ubuntu.com/6223157/ [16:27] should work yes [16:37] thanks stgraber [17:06] we got a test failure on ppc https://launchpadlibrarian.net/153485520/buildlog_ubuntu-saucy-powerpc.mediascanner_0.3.93%2B13.10.20131011-0ubuntu1_FAILEDTOBUILD.txt.gz [17:06] however, mediascanner is currently not utilized by any desktop, it's only used for phone where it works [17:07] so, could we ignore the test error under those circumstances? [17:07] We can remove the binaries on powerpc [17:07] Sorting that out now [17:08] It just needs unity-scope-mediascanner removed too [17:09] cjwatson: thanks [17:09] done now pending the next publisher cycle [17:13] hmm, interesting, pre-release freeze isn't considered as development by Launchpad? [17:13] >>> list(lp.distributions['ubuntu'].getDevelopmentSeries()) [17:13] [] [17:14] (I noticed because I had to pass -s saucy to a few of our tools recently and finally got around to checking why) [17:14] stgraber: correct [17:18] stgraber: .current_series [17:18] stgraber: I fixed ubuntu-archive-tools for that fairly recently - you may want to update [17:18] (r775 last week) [17:22] cjwatson: haha, I was indeed a bit behind on ubuntu-archive-tools! thanks for the fix [17:24] looks like ubiquity 2.15.23 has landed, can we respin all ubiquity images? [17:25] we're still on cronned run [17:25] stgraber: i know, but it has so many fixes and duplicate bugs are kept being filed. [17:26] stgraber: can ubiquity images that have already been build today be respun? e.g. ubuntu-desktop at least? [17:26] xnox: ok, I'll trigger a rebuild of all images listed on the manifest [17:26] stgraber: thanks. [17:27] done (selected all the desktop images listed under the Final milestone) [18:04] infinity, re: bug #1236666 - should I request a sync ? get the source deb and upload manually, or go away and never mention it again ? [18:04] Launchpad bug 1236666 in msr-tools (Ubuntu) "[Feature] update msr-tools package to revision 1.3" [Low,New] https://launchpad.net/bugs/1236666 [18:31] can I upload a fix for python-xlib? (LP: #1231453, and LP: #1221514). Looks like it's seeded on the edubuntu dvd. [18:31] Launchpad bug 1231453 in python-xlib (Ubuntu) "get_full_property() throws exceptions" [Undecided,Confirmed] https://launchpad.net/bugs/1231453 [18:31] Launchpad bug 1231453 in python-xlib (Ubuntu) "duplicate for #1221514 get_full_property() throws exceptions" [Undecided,Confirmed] https://launchpad.net/bugs/1231453 [18:34] stgraber: ^ [18:35] rtg: Sell me on why we want a new version 6 days before release, and what it won't break. [18:40] mdeslaur: sounds reasonable [18:40] stgraber: thanks, uploaded [19:34] infinity, why is msr-tools in main ? the new 1.3 version doesn't look like it has any compelling new features to warrant uploading right now. 14.04 ought to be soon enough. [19:35] rtg: It's in main for cpu-checker [19:35] rtg: Which, in turn, is in main for qemu. [19:36] rtg: Which is in main because we hate our users and refuse to support Xen. [19:36] QED. [19:36] infinity, ok, I'll take option 3 (i.e., go away) [19:36] rtg: Sure. It's not modified in Ubuntu, so it'll autosync when we open 14.04. [19:36] yep [21:05] infinity: fyi, got approval from #ubuntu-ci-eng for the apparmor upload [21:05] infinity: http://paste.ubuntu.com/6224151/ [21:06] $ apt-cache depends apparmor|grep python [21:06] Depends: python3 [21:06] before: [21:06] $ apt-cache depends apparmor|grep python [21:06] Depends: python [21:06] infinity: so, does your comment in #ubuntu-devel mean I can upload? [21:10] * jdstrand uploads [21:11] if there is a problem, reject it [21:12] hmph, apparently ubiquity slideshow translations weren't uploaded since a few weeks. [21:27] knome: hmm, that's unfortunate... let me do that quickly now [21:27] stgraber, cheers :) [21:42] jdstrand: Seems fine. [21:42] knome: uploaded [21:43] Oh, someone got it already. [21:43] yeah, I did [21:43] stgraber, ScottK: thanks! :) [21:44] ^ that diff will likely be quite massive, it's a translation update from LP, so I'd recommend just checking that I'm not doing anything else :) [21:45] stgraber, thanks again! :) [21:48] hint-unblocked ubuntuone-credentials (touch only) [22:11] stgraber: diff looks fine, but is kind of late? Why didn't the translation updates get done before final freeze? [22:11] and do we know for sure that we're having another set of images rolled, so that we should definitely accept this now? [22:19] lool: what unblocking is needed? the fact that it's touch only means it should go through automatically, and anyway I don't see it uploaded [22:19] slangasek: just a followup to: 17:18 < lool> Hey, I've added a block ubuntuone-credentials hint to test it while it's in -proposed (it's not under cu2d CI) [22:20] lool: ah [22:20] slangasek: I did my own blocking and unblocking basically, but wanted to keep the RT in the loop for hints [22:20] ack :) [22:28] slangasek: right, we apparently dropped the ball on that one and nobody updated the translations in weeks... knome noticed a bit earlier today [22:28] stgraber: so we're planning full ISO respins to pick this up? [22:29] slangasek: we're still running on cron at the moment and infinity is planning to turn off cron probably tomorrow/sunday [22:29] ok [22:29] accepting, then [22:29] thanks [22:30] slangasek, ta