[00:11] bdmurray: update-manager -d on trusty fails for me with 'Failed to download repository information' - any idea why? [00:30] slangasek: try with DEBUG_UPDATE_MANAGER=1 set [00:31] bdmurray: seems fairly nondescript: http://paste.ubuntu.com/7455113/ [00:32] slangasek: edit /etc/update-manager/release-upgrades.conf from lts to normal [00:33] bdmurray: hmm; surely -d should override that? [00:33] (so I think that's a bug, even if editing it works around it) [00:34] lts sets lts in the url, so -d with lts just means you want the devel version of the next lts [00:34] ah, that's nonintuitive [00:35] anyway, same error [00:35] (assuming that /etc/update-manager/release-upgrades is the right file, since that's the one that exists / is owned by the package) [00:36] http://paste.ubuntu.com/7455126/ [00:36] new dist: [00:36] so that looks right to me [00:36] yeah... still gives the error about failing to download repo info [00:36] and does not offer me an upgrade [00:37] so that seems to be because I have some repos in sources.list that are not currently accessible [00:37] ah yeah that sounds familiar [00:40] I can't find the bug right away though [00:43] hmm [00:43] but after disabling them, I no longer get the error but I also don't get offered to upgrade [00:43] possibly because it doesn't like me for continuing to postpone the "you must reboot NOW" dialog :) === doko_ is now known as doko === shadeslayer_ is now known as shadeslayer [10:31] doko: Are the powerpc builders just on manual in order to schedule gcc-4.9 on sagari? If so can they be re-autoed now? (Guessing as to somebody who might know ...) [10:31] cjwatson, yes, reenabled [10:32] Thanks [10:32] would take 24+h else [10:33] Sure, I have no problem with shoving stuff on sagari, just wanted the rest of the queue to clear :) === jhodapp|afk is now known as jhodapp [12:37] so who's going to work on 12.04.5? [12:52] hi. out of the trusty's unapproved GStreamer updates, if you can work only on one of them, please try to get gst-libav approved which alone fixes a very visible crasher when seeking h.264 files (including any totem thumbnailer usage since it seeks first) [14:12] infinity, SpamapS, Daviey, slangasek, arges: hey, is there any chance some of you could spend some time on the trusty SRU queue this week? We have some important fixes waiting for 3 weeks :/ [14:12] seb128: sure i'll spend some time today [14:13] arges, thanks! [14:20] seb128: so bug 1312305 bumps gst* packages to 1.2.4 from 1.2.3. some of these haven't landed in utopic yet. should we wait until these land there and get some testing before updating trusty? [14:21] arges, well, we can do that, I was expecting those to be approved to proposed and pocket copied to utopic [14:21] which we often do early in the cycle [14:22] Laney, ^ I think the ones with changes are yours [14:22] (gst and base are synced to utopic already) [14:22] hi all, also following up on the previous comment, could someone help with approving the Qt upload in the Unapproved queue? It fixes a Qt regression that affects running and developing the Reminders app on the desktop. https://launchpad.net/ubuntu/trusty/+queue?queue_state=1&queue_text= - ScottK, perhaps? [14:38] arges: I thought they would be copied up when I did them [14:38] (IIRC) [14:38] I can upload them too at the same time, would appreciate simultaneous verification in proposed though if possible [14:44] Laney: I'm just not 100% sure how to handle these, so I'd like to get input from infinity, slangasek, bdmurray, etc about reviewing the gst* packages [14:45] wait [14:45] arges, just because of the utopic upload thing? [14:45] arges, because e.g gstreamer1.0 has 1.2.4-1 in utopics [14:45] same for -base [14:45] didn't I upload them all already? [14:45] to U, that is [14:46] Laney: there are some that are fix committed in 1312305 [14:46] seems you did [14:46] Laney, you probably didn't update the bug status [14:46] oh I guess because sync [14:46] yep [14:47] those are mine :p [14:47] arges, sorry about the confusion, everything is in utopic, I just failed to set to "fix released" the one that got autosynced from Debian [14:48] done now [14:49] thanks [14:49] Laney: seb128 so has any testing been done, or that will be done after it lands in -proposed? [14:49] in trusty for 1.2.4 [14:49] well, I'm running it locally since it got uploaded [14:50] which is as much testing as we do before sending things to proposed [14:50] not for SRU purposes, that'll be done by people when they test proposed and give feedback to the bug [14:50] that's why we have the "1 week in proposed" to get things tested [15:43] infinity, whats the story on -Werror=cast-align ? libnice started using it, but glib-2.0 is definitely not alignment safe for armhf (and perhaps for other arches I don't care about). https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=743233 - Is it OK to just turn it off and continue take the performance hit for alignment traps ? [15:45] rtg: For arches that do alignment fixups (like armhf), it's just a performance penalty, so safe enough, but still bugs worth fixing. [15:46] infinity, I can fix some, but the glib-2.0 stuff requires deeper thought then I can give it. [15:46] plus, the fixes are probably less performant then just taking the trap [15:48] bdmurray, just after you accepted my ceph upload for 14.04 upstream released a 0.80.1 [15:49] bdmurry: I'm doing some testing via PPA and I'd like to upload that version in the next day or so if that's OK with you [15:50] jamespage: sure, just let me know [15:50] bdmurray, ta [15:50] rtg: Yeah, if the "fix" is wasting a few hundred thousand cycles on mangling things in C to not misalign, you've lost the game. But sometimes, it's as simple as padding a struct correctly or something, and you're done. [15:51] rtg: Of course, for libraries, fixing alignment on public structs also breaks ABI, so sometimes it's pretty much unfixable without upstream buy-in and an SOVER bump. [15:51] infinity, thats what I'm thinking is the case for glib-2.0 [15:52] infinity, it looks like upstream added tihs test, and then never built with a compiler that implements it. [15:52] rtg: Of course. :P [15:52] doh! === dpm is now known as dpm-afk [17:06] bdmurray: Hrm. Why doesn't sru-release have an option to set phasing? Do I just need to do it manually and hope to avoid override races? [17:07] bdmurray: (Have a tzdata upload to do that's critical because the changed timezone takes effect in two days, so phasing it to 100% pretty much will have to happen) [17:59] hey guys, we have another unity8 release that needs a version bump. thanks! [18:00] robru: On it. [18:00] infinity, thanks! [18:00] Man, the new colourful adt results in excuses are distracting. [18:03] haha [18:07] infinity, hey. so looking at excuses, we have 4 different things that are blocked by a regression in ubuntu-purchase-service. I just published the fix for that though, so once it lands do I need to do anything special to get those tests to re-run? or will it see he new release, re-run, and then allow those 4 blocked things to migrate? [18:08] robru: The latter, modulo potential bugs in the system. [18:08] infinity, sweet, ok thanks, will keep an eye [18:08] Yeah, I believe that should work. [18:32] infinity: tz change with 2 days notice? o_O [18:33] michagogo|cloud: It's possible I've cursed a little bit this morning. [18:34] infinity: reminds me: have you seen https://www.youtube.com/watch?v=-5wpm-gesOY yet? [18:35] This specific point seems relevant: https://www.youtube.com/watch?v=-5wpm-gesOY&t=4m18s [18:35] michagogo|cloud: Nope. But given that (a) he's English, and (b) it's about timezones, I can only assume it'll be full of colourful language. ;) [18:35] infinity: nah [18:35] Shame. [18:35] Tom's videos are mostly pretty clean [18:36] Be glad that it's not about something which has already happened [18:36] He's done a few for Computerphile, but he also does his own videos on his own channel [18:36] Laney: hmm? [18:38] infinity: Got the OpenSSL exception in. ↑ :) [18:38] jpds: Snazzy. [18:38] jpds: Did you have problems explaining why that was necessary, or was it a low drama thing? [18:39] infinity: A little bit. Apparently libpam-mount doesn't have the exception. [18:39] infinity, hey, you don't want to do some trusty SRU reviews by any chance? ;-) [18:40] jpds: Fun. That should be fixed. [18:41] infinity, I guess it's a no? :-( [18:42] seb128: I'll do some, yeah. [18:42] thanks [18:42] sorry for being nagging, but we have important fixes in the queue for almost 3 weeks [18:42] seb128: If this is specifically about gst and webkit, we've been chatting about those internally. [18:43] those are indeed the oldest one [18:43] but gst-libav fixes some of the most reported e.u.c issues [18:43] (totem thumbnailer hitting segfaults on building image for mkv files in nautilus) [18:45] seb128: Right, will have a poke today. [18:45] thanks [18:49] michagogo|cloud: Oh man, I didn't know about the West Bank split timezone. That's utter madness. [18:50] infinity: is it normal to have versions like "1.2.4-1~ubuntu1" (looking at the gst* stuff for trusty) I see 1.2.4-1 is already in utopic... [18:51] arges: Normal? No. But it sorts correctly, so it's certainly accetable from the archive's POV. [18:51] acceptable, too. [18:52] infinity: ok yea I understand that... wasn't sure if we normally have versions with ~ [18:52] The security team (and I) would probably argue that ~ubuntu0.14.04 is easier to understand, but meh. [18:53] arges: We use ~ all over the place. It literally means "just before the preceding version atom". [18:53] infinity: yeah, there's a related darwin award [18:53] one sec [18:53] arges: So, -1 is >> -1~ but -0whee is << -1~ [18:53] infinity: cool. yup I use it as well.. just didn't want ot get the archive all dirty [18:54] infinity: http://darwinawards.com/darwin/darwin1999-38.html [19:48] seb128: And when I said "I will look at it", clearly what I meant was "Chris will beat me to it". But I'll look at more of the queue later to see if we can empty it (or close). [19:48] infinity, thanks ;-) [20:06] infinity: could you please look at urfkill in proposed? I'd like to fix the version number if possible ;) [20:09] cyphermox: Which series? [20:09] utopic, sorry [20:09] I should just upload a new version on top without the tilda [20:10] cyphermox: Erm, kay. Nothing for me to look at there, there's no queue involved. [20:11] no, there isn't, sorry === beidl_ is now known as beidl [20:35] trying to understand why dbus-cpp is still not promoted from proposed, and I see a failure at http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt but don't really know what that means [20:37] robru: any idea? [20:37] rsalveti: well, there's an soname change; have the reverse-dependencies been rebuilt? [20:38] slangasek: not the entire rev-dep, no [20:38] is that needed for promotion? [20:38] right, all the revdeps need to be rebuilt first so the packages can transition together [20:38] rsalveti: What's needed is that upgrading it doesn't cause those other packages to become uninstallable. [20:38] looks like libconnectivity-cpp0 is the only one missing [20:38] (based on the last error, at the bottom of the file) [20:39] right, but would libdbus-cpp2 be removed from the archive after the update? [20:39] yes [20:39] right, got it [20:39] thought it'd stay until nobody is depending on it anymore [20:39] It does stay. [20:40] mmm? I thought we dealt with NBS in -proposed these days [20:40] did cjwatson hack britney to do something different? [20:40] Britney assumes it'll be deleted. [20:40] It's still removed manually. [20:40] right [20:41] This makes things less painful in some cases. [20:41] But the transitions should mostly still happen in -proposed due to britney testing as if the binary was removed. [20:41] Which is good. [20:41] ah; yet britney calculates the migrations using the assumption that it will be deleted [20:41] * slangasek nods [20:42] cool, thanks guys [20:42] Anyhow, just looks like a straight rebuild of connectivity-api is needed. [20:42] Or whatever that was. [20:42] yeah, checking that [20:44] rsalveti: In future, the easiest way to read update_output is to search for the source you care about, and then search backwards (so you get the last mention in the file). [20:44] yeah, I saw the failure in there, but thought that having the old package around would already be enough [20:44] rsalveti: That'll show the autohinter attempts which, in this case, shows: [20:44] leading: dbus-cpp,media-hub,location-service,platform-api [20:44] * i386: libconnectivity-cpp-dev, libconnectivity-cpp0 [20:44] Much less scary than the first hit you see, that tells you half the distro is broken because of dbus-cpp. ;) [20:45] right :-) === Ursinha is now known as Ursinha-afk [20:56] slangasek: right, partly this is because -proposed being a partial suite makes it tremendously painful to do anything else; but fortunately we get constrained into something that's roughly sane. ish === Ursinha-afk is now known as Ursinha [21:46] infinity: I could add a phased_update_percentage switch if you want [21:46] bdmurray: Well, for today, I just hacked it for my needs. [21:47] slangasek: what do you think about release apport early? It'll mean slightly less database reapir work [21:47] bdmurray: But a switch for either initial-phase or just phased-on/off toggle would work. [21:47] slangasek: from trusty-proposed [21:47] bdmurray: That logic could be reused so that when calling with --security, we also don't phase. [21:47] bdmurray: (Since I assume we don't phase security, so it doesn't make sense to phase the updates copy) [21:47] infinity: we don't phased when called with security [21:48] bdmurray: Oh, no? Kay. [21:48] bdmurray: That's not how I read the code. [21:49] bdmurray: The way I read it, it'll phase to -updates, but not -security, which makes less sense. [21:49] Oh, no. [21:49] I missed that if. [21:49] if (release not in ['lucid', 'precise', 'quantal'] and [21:49] not options.security): [21:49] Yeah. [21:49] Derp. [21:49] I missed that cause it was also the series check. [21:51] bdmurray: how well exercised is the fix for bug #1282349? That's the only one that looks intrusive enough to make me worry about early release [21:54] well its been in utopic for a couple of weeks now and the fixed package version doesn't appear in the errors bucket - https://errors.ubuntu.com/bucket/?id=/usr/share/apport/apport-gtk%3AEOFError%3A%3Clambda%3E%3Acollect_info%3Aexc_raise%3Arun%3Athread_collect_info%3Aadd_gdb_info%3Agdb_command%3Awrite%3Aread%3A_read%3A_read_eof%3A_read_exact [21:55] bdmurray: ok, +1 [23:18] infinity: https://code.launchpad.net/~brian-murray/ubuntu-archive-tools/set-phased-update-percentage/+merge/219444 [23:42] can bijiben and gnome-boxes be dropped from -proposed? they auto-synced but require gtk 3.12