[03:44] Good morning [03:45] rodrigo_: wasn't input support recently discussed upstream at least for ibus? that should solve most use cases anyway === tkamppeter_ is now known as tkamppeter === zyga-afk is now known as zyga [06:41] Good Morning. [07:13] hi pitti [07:14] hey rickspencer3, how are you? [07:14] pitti, I am doing well [07:14] I am excited today [07:14] looking forward to the increased manual testing cadence [07:14] jono told me that they are going to go for complete test runs every 2 weeks! [07:15] pitti, how are you doing? [07:15] nice! [07:15] rickspencer3: I'm great, thanks! Making good progress on our jenkins autopkgtest tests [07:15] yeah [07:15] nice to see us continuing to make progress on quality in 12.10 [07:17] pitti, is it true that we go for weeks without ARM images? [07:17] ogra_, ^ [07:17] I'm not sure [07:17] I haven't really followed them so far [07:18] but last Friday I got a Panda board delivered, I was going to set it up today [07:19] ok [07:19] so, I don't know why we would think that we need to wait for milestones to fix that [07:19] http://cdimage.ubuntu.com/daily-preinstalled/20120619/ seems current, though [07:19] I don't consider this to be a good state of affairs if we are going days without an image, we should figure out the problems and solve them [07:19] ok [07:19] I was wondering if it was a point in time thing [07:19] rickspencer3: I think infinity, ScottK, you, and me are mostly talking past each other [07:20] pitti, well, I tried to clarify on my last response [07:20] rickspencer3: yes, that's the point of the stable+1 team [07:20] it seems we all violently agree [07:20] I think there were a few discussions mixed together [07:20] and the main point of dissent is the nomenclature [07:20] well, I really just wanted to know about not freezing, to be honest [07:20] (for the milestones) [07:20] IMHO, when we all rally around testing on certain points in time (say, every two weeks) to shake out bugs found in manual testing, why not just call this a "milestone"? [07:21] because that's exactly what milestones have been [07:21] pitti, right [07:21] we can call it differently, of course [07:21] but we can't do that if we freeze the archive and everything [07:21] we haven't actually frozen the archive for alphas in years [07:21] well, we sort of do [07:21] it was a soft-freeze, so people still backed up a bit [07:22] but we need that if we want to be able to put out a known-good image [07:22] well, people tell me that it caused them a lot of extra work, and it went slow [07:22] using -proposed by default shoudl help a lot here [07:22] I think Thierry's idea was ideal [07:22] just use the "last known good" image [07:22] which I think we can easily do these days [07:22] "a lot of extra work" and "slowing down developers" should be, quite frankly, blatant lies [07:22] oooh [07:22] you can queue up stuff in bzr, in PPAs, work on bugs, work in upstream trunks, etc. [07:23] well, that all sounds like extra work and slowing down development to me [07:23] lol [07:23] except for the release team, of course, who do have to work fulltime on the images [07:23] yup [07:23] rickspencer3: not to me; we do stuff in bzr, at upstream, etc. anyway [07:23] the only thing that you need to hold back a bit is uploading to the dev release [07:24] but that's not causing any extra work [07:24] it's just causing some delay of landing stuff [07:24] well, I think that delay is lost velocity [07:24] you can upload to a PPA if you want other folks to test [07:24] well, you can't have both [07:24] can't have both what? [07:24] we can't produce a known-good image for public testing _and_ have uploading go on at full speed [07:24] I think we produce lots of known-good images [07:24] good enough for public testing [07:25] I think we do it *most* days, in fact [07:25] well [07:25] we produce a lot of images where jenkins thinks they are good [07:25] if they actually _were_ good, the release team wouldn't need to work three times and 10 respins to actually _make_ them work [07:26] s/times/days/ [07:26] yeah, but we know we need the respins because of the manual testing [07:26] well [07:26] so, those respins are usually about fixing installer bugs and such [07:26] we need the respins because of the bugs only found by manual testing [07:26] correct [07:26] yeah, so those bugs don't really block testing, do they? [07:27] if the daily image on milestone date minus two days were good, we would just use it [07:27] sure they do [07:27] we had a lot of cases where the installs completely fell over [07:27] or the result didn't actually work [07:27] yeah, but those installs falling over are test results, no? [07:27] so we fix the bugs [07:27] correct [07:28] and then need to re-test again until we have an image which really works [07:28] right [07:28] once we have that, we unfreeze and publish that [07:28] so [07:28] which is the really easy part [07:28] a. it's not really frozen, is it, as we keep changing the installer [07:28] (well, there's documentation and announcements to write, and so on) [07:29] b. we should be doing that at will, not only a few times per cycle [07:29] that's what we call a freeze - only upload minimal changes that fix the breakage, and not upload other stuff which introduces breakage again [07:29] so, I think making sure the installer works is very very good [07:30] but I think having the whole milestone process is a lot of effort for that one focused area [07:30] we could just have people test the installer weekly, etc... [07:30] and fix the bugs soon after they are introduced [07:30] we could, but I really doubt it would go as well [07:30] infinity hit the nail on the head [07:30] although he expressed it differently [07:30] oh? [07:30] you need to _make_ time for this [07:31] indeed [07:31] otherwise it will not happen [07:31] but we need to make time for it more than a few times per cycle [07:31] we don't have a lot of devs who say "gosh, I'm bored today, let's test some images" [07:31] rickspencer3: I agree [07:31] it takes a lot more time to fix bugs weeks or months after they are introduced [07:31] so if anything, we need more milestones, not fewer [07:32] lol [07:32] that's what I meant with "we can certainly discuss about the frequency" [07:32] well, we need more frequent testing, and higher standards for "daily quality" [07:32] but I don't see how we can drop the entire concept and process [07:32] pitti, well, we drop it when it is no longer useful to us [07:33] and it is no longer useful to us when our daily quality exceeds what we get from milestones today [07:33] when our automated daily testing is good enough one day, we won't need it, yes [07:33] to be clear, I'm not arguing to drop milestones, I didn't actually bring that up [07:33] oh, I think some manual testing is always needed [07:33] right, it won't happen in the next two years [07:33] I think automated testing tells you how worth it is to do manual testing [07:33] actually, I think it will happen soon [07:33] (automatic testing being sufficiently good, I mean) [07:33] right [07:34] e. g. for the installer we do not test the UI at all [07:34] but I think we can rally the testing community to give us more actionable and more frequent feedback [07:34] the preseeding is by and large a special-case code path [07:34] pitti, it seems like the installer is such a special case for us, we should be much more focused on it [07:34] and we don't test GL, graphics, unity, sound, etc. [07:35] like we have all of this release machinery and effort built around the whole distro, when it really comes down to ensuring the installer works [07:35] I'd actually like to drop the preseeding and do real interaction with the installer UI/widgets [07:35] well, it's installer + kernel + X.org + unity + sound + different hardware platforms [07:37] interestingly, xorg and unity do a lot of testing outside of the distro testing cadence [07:37] right; the more dev release daily users we have, the better (and they are growing, I think) [07:38] yeah, and the better our daily quality, the more we will get [07:38] but I think the community team can help by bringing rigor more rigor to the dev release testers [07:38] I bet there are a lot of people who would love to contribute in that manner [07:39] * micahg would love a way to automate UI testing :) [07:40] micahg, sounds like you find a nice weekend project ;) [07:41] yeah, it's also one of the things I'd like to work on [07:41] I already test apport that way (GTK and KDE) [07:41] by emitting clicks to the buttons and other UI parts, and checkign the state of the program and widgets afterwards [07:42] it's still a lot of overhead, we need to think about how to make this radically simpler [07:42] but it does work in principle [07:42] didn't someone make a desktop recorder kind of app at some point? [07:43] like you click around and it generates Python code for your clicks? [07:43] dogtail [07:43] it requires a11y enabled, and only works for blackbox testing [07:43] but for this kind of installer test, blackbox might actually suffice [07:46] in the meantime, it sounds like we should see if we can arrange some people to test the installer on more like a weekly basis [07:46] (by manual testing, I mean) [07:49] micahg, ldtp is good for UI testing of GTK apps but not all components are exposed in Ubiquity. For Qt, the testability driver is very good. [07:49] jibel: I need non-GTK :), I'd ideally like to automate distro firefox testing :) [07:50] I thought firefox was gtk2? [07:50] yes, for some things [07:50] ldtp works quite fine with firefox, but identifying links in pages is a real pain [07:50] jibel: links can be done other ways [07:51] well, you can mix technologies [07:52] right [07:52] jibel: ok, I'll have to look at ldtp at some point then, thanks [07:52] * micahg loves 500 errors when searching for stuff [07:56] good morning [07:57] hey didrocks [07:57] * pitti hugs didrocks for the lost game, but at least you are in the quarter finals [07:57] hey pitti [07:58] pitti: are we? I'm totally disconnected from football world :) [07:58] * didrocks hugs pitti nevertheless [08:07] hey [08:07] good mooooorning desktoppers! [08:08] * Sweetshark survived a good flamefest yesterday: http://nabble.documentfoundation.org/ANN-Please-use-Gerrit-from-now-on-for-Patch-Review-td3990754.html [08:10] seb128: whats the current state with libreoffice SRU/MRE, btw? i somehow lost track there ... [08:10] bonjour seb128 [08:10] hey Sweetshark [08:10] hey Sweetshark, pitti [08:10] seb128: hope you aren't too sad about the game [08:10] * Sweetshark waves at pitti [08:10] Sweetshark, it has been acked in principle, it just needs a courageous SRU team member to press the button [08:11] seb128: lets make that a big red botton with the text "nuke" on it ... [08:11] pitti, if we had to play bad once it was better to be this game, I hope we do better against Spain! [08:12] Sweetshark, yeah, RAOF said he was looking at libreoffice yesterday, not sure how far he went, otherwise we will need to chase down slangasek to ack it, I don't think the other SRU team members will do it [08:12] seb128: spain-croatia wasnt really showing any spanish dominance, so you should stand your chances ... [08:13] Sweetshark, yeah, we both had a day off, let's see how that plays out for the next game... [08:19] seb128: I finally managed a backtrace for my weird bug: Bug #1015443 [08:19] Launchpad bug 1015443 in thunderbird "Thunderbird hangs the desktop when escaping from password prompt" [High,Triaged] https://launchpad.net/bugs/1015443 [08:27] seb128:the thing is, I can't seem to reproduce on my metal precise system in the same env [08:29] micahg, is that an unity-2d issue again? ;-) [08:29] seb128: no, I can reproduce in lucid and natty w/out unity [08:29] I am thinking it might be an X/driver issue of some sort though [08:30] since I can't reproduce in the host env [08:30] micahg, is that blocking the tb13 update? [08:30] micahg, what happened to firefox 13.0.1? [08:30] seb128: yes [08:30] seb128: well, I wanted to try to get TB out, but ran into this again, I'll go back to finishing FF now [08:31] micahg, that doesn't seem worth blocking the update honestly, especially if that happens only in vms [08:31] seb128: well, I don't know what video driver combinations would trigger it [08:31] if that's indeed the issue, if it's only VMs, I agree :) [08:32] micahg, are you sure it's a new issue in 13? [08:32] seb128: yes, I reverted to 12 and can't reproduce [08:34] micahg, did you open an upstream bug? [08:34] seb128: no, as I don't know if it affects upstream or not (depends if it's a thunderbird bug or a bug elsewhere in the stack) [08:35] micahg, we really have an issue there in the velocity of how we deal with problems and we need to sort that [08:36] micahg, we are over a week late for that update already, and that bug seems nowhere close from being resolved or debugged :-( [08:36] seb128: yes, indeed, we'll be testing 14 in advance [08:36] micahg, it usually doesn't hurt to take upstream input on issues [08:36] seb128: unfortunately, when I tried to reproduce last week, I couldn't, so I figured that maybe it was fixed by the unity update [08:37] seb128: but, we are slated to do beta testing this time around, so that should give us more time to fix issues as they pop up [08:38] micahg, ok, at least that's a move in the right direction ;-) [08:38] micahg, can you go finish the firefox update, I will try to see with chrisccoulson if we can help on the tb hang you are having [08:38] seb128: yep, doing so now [08:39] thanks [08:40] seb128: FWIW, the protocol testing that I managed to do before it froze each time seemed to be fine [08:41] micahg, does going to a vt and killing tb release the lock, i.e do you get your desktop back if you do that? [08:42] seb128: let me see, ISTR no, but I'll try again [08:42] micahg, the "hang" description is weird, keyboard working and no mouse? [08:43] ooh, yeah, it does release it [08:45] more info, mouse moves (menubar appears, but clicks aren't recognized) [08:45] * micahg adds to the report [08:45] pitti: do you remember the set_data/get_data in gobject we talked about some time ago? I use it in the tests in software-center quite a bit to store various bits and piece of information, is it coming back as a compat function for gobject or do I need to rip it all out and replace ? [08:47] micahg, chrisccoulson: if stopping tb releases it then it seems that tb keeps an active grab [08:54] mvo: it's meant to be put back, but as deprecated API; so at some point it will disappear [09:05] pitti, did you see that davidz commented on the gvfs,loop mounting bug saying there were perhaps 2 bugs in there? [09:06] seb128: I saw that, yes; I was hoping to get a confirmation on the Ubuntu bug that we are really talking about the same bug [09:07] pitti, ah right, I did a ping on launchpad then ;-) [09:07] pitti, thanks [09:14] pitti: ok, thanks, I have a look at this then [09:15] mvo: I had an intial attempt for a backwards compat patch, but it doesn't work for many kinds of classes [09:15] since then I didn't find time to get back to this [09:16] presumably I'll revert the removal of it and add a g_warning() to it [09:16] and we'll remove it in GNOME 3.8 [09:20] pitti: ok, thanks [09:24] mvo: done now [09:33] seb128, was there an SRU for nautilus lately? [09:34] rickspencer3, we got one to -updates a week ago and a new SRU candidate in proposed the same day [09:34] rickspencer3, any issue? [09:34] seb128, well, nautilus just crashed on me on 12.04 [09:35] but if there was not update ... [09:35] nothing really to see here [09:35] :) [09:35] whoopsie-daisey will let you know if there is a problem ;) [09:35] rickspencer3, yeah, I doubt it's due to an update, you just likely hit one not-so-new bug [09:36] rickspencer3, right [09:36] rickspencer3, what were you doing when the issue happened? [09:36] renaming a newly created directory [09:39] rickspencer3, I've seen report suggesting issues around that in the past :-( [09:39] I'm sure it will be fixed in the next SRU ;) [09:39] rickspencer3, could be bug #985848 [09:39] Launchpad bug 985848 in nautilus "nautilus crashed with SIGSEGV in icon_rename_ended_cb()" [Medium,New] https://launchpad.net/bugs/985848 [09:40] exactly [09:40] rickspencer3, that one is ranked 11 on the nautilus bugs on errors.ubuntu.com with a frequency of 73 [09:40] there you go [09:40] I hope we get to it at some point [09:41] rickspencer3, thanks for point it ;-) [09:42] after the SRU, we should see if the crash stops getting reported [09:42] will be a good test of whoopsie-daisy [09:45] hi everybody, [09:45] is there any reason for still having revert_git_* patches in g-c-c and g-s-d? [09:58] pitti, ping - does apport only have a CLI interface (ubuntu-bug) and not a (trivial) GUI interface (dash -> File A Bug -> enter package name in a zenity popup) from deliberate policy (filing bugs is technical, thus terminal-only) or because no-one's had a chance to make a simple GUI handler? [09:58] aquarius: actually we have had a GUI for this until precise [09:58] Help -> Report a bug.. [09:59] pitti, yep, that's perfect for GUI programs (and I know it's gone away now), but it's harder for "ubuntu-bug linux", for example :) [10:00] achiang: we also used to have that, but quickly tore it down [10:00] we got a massive increase in absolutely useless and misassigned bugs [10:01] mitya57, the same reason we added them to start? [10:03] pitti, no problem; it's policy to not have such a tool, which is fine with me. I was debating writing a quick one, but figured that you might not have it by decision rather than by lack of resources, which is in fact the case :) [10:03] seb128: we added them because we wanted the new version but were in UIF, didn't we? [10:03] aquarius: right [10:03] s/UIF/FF/ [10:03] aquarius: the compromise that we found was bearable was to call ubuntu-bug without arguments [10:04] mitya57, no, we added them because compiz still uses gconf and not gsettings and because we don't use systemd and don't have their datetimed service [10:04] mitya57, neither of those assumptions changed [10:05] mitya57, though the compiz on gsettings update is being worked on this week [10:05] pitti, cool, I didn't know about that :) [10:06] seb128: ok, thanks [10:06] aquarius, we don't lack bugs reports and we don't need to make bugs easier to file ;-) [10:06] * aquarius grins [10:07] mitya57, did you ask for a reason? i.e are they creating any issue? [10:07] seb128, I did wonder if that was the case, which is why I asked rather than saying "here is a gui thing!" :-) [10:07] seb128: no, just wondering [10:07] mitya57, btw https://launchpad.net/~ubuntu-desktop/+archive/ppa has g-s-d and g-c-c with the gconf->gsettings reverts dropped if you want to test it [10:09] seb128: I'll test that when I'll update my quantal WM to the latest quantal :) [10:11] will Quantal have EDS 3.5 with the new storage format? [10:12] xclaesse, not decided yet but it's likely yes [10:12] xclaesse, do you need it or do you recommend against it? ;-) [10:13] seb128, we definitely need it for Folks [10:13] master is already ported to the new API [10:13] ok [10:13] let's see how that goes [10:15] seb128, cool :)à [10:15] seb128, the problem is that once the format is migrated, you can't go back [10:15] which is problematic for testing :p [10:15] right, well going backward in versions is an issue for distros in any case [10:16] but we tend to not jump on such transitions, especially for evolution which has a record of making those bumpy for a while [10:16] so it might take some weeks before we get the new version of e-d-s [10:16] seb128, I was planning to upgrade to quantal once it has the newer EDS, so I can use latest folks again :p [10:17] xclaesse, we will let you know when that happens ;-) [10:17] even when building newer EDS in jhbuild, it will migrate your DB and you're screwed [10:18] seb128, thanks :) === MacSlow is now known as MacSlow|lunch === ronoc is now known as ronoc-lunch [12:33] * achiang waves to pitti :) [12:34] hey achiang [12:34] achiang: argh, tab failure, sorry :) [12:34] pitti: np. my error checker fixed it. but it's been a while since i said "hi" anyway. :) [12:35] achiang: indeed, how are you these days? [12:36] pitti: a bit tired. in korea for work, the taxi drivers decided to have a national strike today. so i got to explore the korean bus and metro system to get where i needed to go. 2 hours later, i am back from the customer site back to hotel [12:37] achiang: argh; consider it a real-life sightseeing tour? [12:38] pitti: "forced" tourism... sounds kinda funny. :) wie gehts? [12:38] achiang: quite well, thanks; I'm getting into my new QA role, and making some progress [12:39] pitti: good to hear! you sounded quite excited in your blog post so i think we're all excited by proxy. :) [12:39] lol [12:40] I'm keeping a record on G+ (quite nice, and QA team policy anyway0 [12:42] :) [12:42] ok, time to write a few status emails and then perhaps... bed [12:42] cheers, pitti === MacSlow|lunch is now known as MacSlow [12:48] chrisccoulson, did you know enigmail is broken in quantal? [12:50] popey, yeah :) [13:16] micahg, how is the tb debugging going? [13:16] seb128: hrm, I thought chrisccoulson was going to do that :( [13:16] micahg, he said he can't reproduce the issue [13:16] micahg, I fear you will have to do it [13:17] micahg, he has also other stuff to work on for quantal... [13:17] micahg, let we know if we can help on specific topics, but it doesn't seem we are in a position to just do the debugging and resolve the issue for you there [13:18] seb128: awesome, ok, I guess I'll look into this further when I start later today [13:18] micahg, ok, thanks, let us know how it goes [13:18] * micahg wonders if he missed scrollback somewhere [13:20] micahg, well, scrollback never said chrisccoulson was going to work on,fix that issue either ;-) [13:20] micahg, but I did chat with him out of the channel and we can't reproduce the issue which makes it hard to work on it [13:40] mvo, darn, I was hoping you had been hiding some amazing UI driving framework that you had been too lazy to use. :-/ [13:40] mvo, the best I've used is ldtp, but it's very rickety [13:44] mterry: yeah, excatly, I wasn't overly impressed with that one :/ [13:48] mvo, well, I'm happy to write various tests, but since doing so might involve various code changes (for instrumentation) and since code is moving around a lot in my branches, would you mind if I did a big test branch after all these intermediate branches land? (ala the indentation one) [13:56] mterry: that is fine with me (for the given reasons) [13:56] mvo, cool, I'll comment on the merge request page saying that's my plan [13:58] ta === zyga is now known as zyga-food [14:23] seb128, I just released telepathy-gabble 0.16.1 stable release, I would appreciate if that can make its way into ubuntu precise [14:23] seb128, it fix issue connecting to MSN's xmpp server [14:24] that just regressed last week because of a change on their server [14:24] details: http://blogs.gnome.org/xclaesse/2012/06/20/problems-with-windows-lives-xmpp-server/ [14:28] xclaesse, thanks for the notice, we will SRU that [14:28] kenvandine, ^ did you see bug reports about that? [14:28] kenvandine, did you want me to have a look to that SRU? [14:29] i hadn't seen a bug about that... but that doesn't mean we didn't get a new one [14:29] seb128, if you could please :) [14:29] i couldn't do it today [14:29] thanks ç [14:29] ! [14:30] we got upstream reports [14:30] dunno about lp [14:31] there are over 10k users of that, so surely a lot of people noticed it [14:31] btw, it went from ~1000 to ~9000 just in a few days after LTS release :D [14:31] xclaesse, ;-) [14:32] compared to fedora's release almost unoticed in the chart :p [14:32] xclaesse, I will check our bug reports, I just want one to attach to the SRU (we need a bug for tracking purposes) [14:32] kenvandine, no worry, I will do it [14:33] seb128, you rock! === zyga-food is now known as zyga === skaet_ is now known as skaet === zyga is now known as zyga-afk [15:54] Laney, ping https://bugs.launchpad.net/ubuntu/+bug/1011361 ;-) [15:54] Ubuntu bug 1011361 in ubuntu "[needs-packaging] libpwquality" [Wishlist,Triaged] [15:54] Laney, I saw robert_ancell uploaded it, it's in quantal NEW queue [15:55] heh [15:56] I think it would be good to put it in exp, but I'm not sure I want to be uploader ;-) [15:57] Laney, start by putting it in the pkg-gnome svn? ;;-) [15:57] * Laney checks if Robert is a member there [15:58] using SVN scares me these days, all of the actions being pushed out immediately [16:00] Laney, he is a member of pkg-gnome, I dropped him an email about the topic but he didn't seem interested [16:01] Laney, he mentioned "the ridiculous amount of Debian paperwork for new packages" [16:01] one bug? [16:01] * Laney will look at it then [16:01] Laney, though he discussed with mbiebl apparently who said the package seemed to be fine [16:01] just commit it to the svn and as a base for whoever will be wanting to pick that up ;-) [16:01] sure [16:02] Laney, one bug, well apparently he failed to open this one [16:02] I would have to review it for ubuntuisms though [16:02] thanks for the ridiculous bts email system :p [16:02] which is why I'd rather he did it, as he'd know [16:02] Laney, I doubt there is any, I reviewed it quickly, it seems fine [16:02] ok, will do === fenris is now known as Guest34289 [16:20] Laney, seb128: didn't have time really to look at libpwquality as I had other stuff on the todo list and [16:21] libpwquality being 3.5 was a low prio [16:21] it's ok, not a high priority indeed [16:21] I'll maybe put it in exp [16:21] I think I made a few quick remarks regardign the location of the pam modules [16:22] which was wrong /usr/lib/security iirc [16:29] mbiebl: you mean it should be in a multiarch path? [16:31] Laney: they usually go to /lib/security or /lib/$(MA)/security [16:31] yeah, looks (from the .install file) like this is going to /lib/security [16:31] laney@raleigh> head libpam-pwquality.install ~/temp/libpwquality-1.1.0/debian [16:31] ok, then robert has fixed this already [16:31] lib/security/*.so [16:31] cool === Guest34289 is now known as ejat === larsu_ is now known as larsu === larsu is now known as Guest84126 === jbicha is now known as Guest6354 === Guest6354 is now known as jbicha_ === jbicha_ is now known as jbicha [23:33] RAOF, TheMuso, robert_ancell: we missed our meeting time yesterday, but I'm assuming no one had agenda items? [23:33] Correct. [23:34] bryceh, yep [23:35] Correct. [23:38] great, thanks [23:48] /c/c