[03:07] charles, how do you do SRUs for the indicators? [03:07] just chuck stuff in debian/patches or make a new release? [03:12] Good morning [06:25] morning folks [06:25] hey willcooke! [06:25] ready for day #2? ;) [06:25] hehe [06:25] kinda [06:26] I'm off in to London in a little while, so I wanted to get the ball rolling before I sit on the train [06:26] actually, I have a question for you didrocks [06:26] sure [06:26] I'm trying to work out who is in our team (pretty fundamental, right) [06:26] so let me send you a link [06:26] well, it's a good start :) [06:30] hey seb128, can you make me an admin on https://launchpad.net/~canonical-desktop-team [06:31] good morning desktopers [06:31] willcooke, sure, sorry I didn't respond because I'm not fully in work mode yet, but can do that ;-) [06:32] no hurry seb128 [06:32] I'm not either - no coffee [07:33] Hey willcooke. [07:34] howdy TheMuso - good to meet you :) [07:34] You too. [08:01] right - off to London. Be back online in about an hour (depending on the trains) [08:02] yo [08:05] hey Laney, how are you? [08:08] hey seb128, doing good thank you [08:09] just sat around last night and napped/watched some tv, didn't end up going climbing ;-) [08:09] pent up energy, raaargh [08:09] you? [08:10] aloha [08:13] hey czajkowski [08:13] wie gehts? [08:14] not just yet :) 3 hrs to flight :) time to print of name badges and check in yet [08:15] it's a glamorous life you've got there [08:15] Laney, I'm fine thanks, not fully awake yet ;-) [08:15] morning Laney, czajkowski [08:15] czajkowski, hey, unity working today? ;-) [08:15] seb128: still keeping up this early morning IRCing though :P [08:16] Laney, ye [08:16] Laney: see, going in Lyon can change a man! :) [08:16] Laney, yeah, there is no budging there! [08:16] lol [08:16] well, let's see and wait how it goes in winter ;-) [08:16] seb128: eh no :( but I now have all the commands to restart various things so I can get back onto my main account and not live on guest [08:17] thanks for asking :) [08:17] but it's actually alright for summer [08:17] seb128: same, but with hot wine :p [08:17] czajkowski, :-( [08:17] didrocks, lol [08:17] didrocks, then I'm the one with the drinking problem, right? [08:17] you obviously are :) [08:19] tsss [08:35] * didrocks retries to build a desktop docker container [08:35] first automated trial was a fail :) [08:40] didrocks, enjoying docker then? :-) [08:44] seb128: yeah, it's quite nice :) [08:44] just needs to find a way to install ubuntu-desktop without any priviledge [08:45] ensuring that I create a passwordless sudo user and so on… [08:45] (to have that container running what we need) [08:45] and ensuring that coverage works as expected, even within the container [08:59] willcooke: come on, The IT Crowd reference on day 2! You set the bar high :) [09:01] didrocks, I'm going to Gimp up an A team picture with your faces on it later and put it on my desk. [09:03] :) [09:03] nearly in to London now - on the train atm. 4G is working quite well. I get 20 Mb down, which is more than I get on ADSL at home [09:03] :) [09:03] shutting down - bbiab [09:49] * didrocks goes for exercising [09:49] seems like no rain this time :) [09:50] didrocks, enjoy! === mpesari is now known as pesari [10:08] seb128, do you have a calendar entry for the weekly meeting you could invite me to? If not I'll create one and send it round [10:09] willcooke: It's on the Ubuntu fridge calendar [10:09] * Laney laughs at the optimistic Agenda link there [10:10] willcooke, there is one in the UE calendar owned by Michelle as well [10:10] which I copied in my calendar, personally [10:10] hey willcooke, welcome back! [10:10] but if you prefer it handled some other way please do [10:10] so that's the man who took my job! [10:10] lol [10:10] haha [10:11] willcooke: in case that's confusing to you, ask any desktop team member over a beer :) === willcooke is now known as willcooke|meetin [10:11] :) [10:11] pitti for VP of Ubuntu Engineering! === pitti is now known as somedude [10:12] Laney: pitti who? === somedude is now known as pitti [10:13] Laney, forget about that, let's just appoint some dude [10:13] * Laney looks at a photo and sees pitti in the middle of fading out === Zdra is now known as xclaesse === shiznix_ is now known as shiznix [10:26] seb128: some dude isn't he like that pitti chap a French guy who thinks he's German? [10:27] davmor2: mec, tu dois parler français ici ! [10:28] pitti: does that make you the anti seb128 ? [10:28] davmor2: koennte sein :) === larsu_ is now known as larsu [11:09] desrt: hey Ryan, how are you? [11:09] pitti: hi! pretty good [11:10] brain recovering nicely after a week of drowning in another language [11:10] desrt: do you see any reason to not commit the cgmanager bits to systemd-shim trunk, and making a new release? we're shipping that now in Ubuntu, and would like to ship it in Debian too [11:10] desrt: Esperanto? :-) [11:10] yup [11:10] just went to a week-long event.... no english spoken there.... [11:10] desrt: très cool ! [11:10] so ya [11:10] we can merge that stuff now [11:11] i actually told hallyn that i'd be OK if you pushed it (since you have commit access) [11:11] desrt: it's not perfect yet (old sessions don't get cleaned up sometimes), but it's workign well enough [11:11] ya... the stop thing is hard [11:11] desrt: ah ok, that got lost in the Chinese whispers then [11:11] because you stop the session unit, and we don't know the cgroup path (ie: under which user) [11:11] so we can't properly kill it [11:18] * Sweetsha1k smirks rolling through the backlog. === MacSlow is now known as MacSlow|lunch === willcooke|meetin is now known as willcooke [12:18] didrocks, ping [12:18] darkxst: please don't use contextless ping, just ask ;) [12:20] (especially if you are not around to ask the next question) [12:20] didrocks, so have been chatting with tracker upstream, unit tests are 100% guaranteed to pass (they won't do a release with these broken) [12:20] darkxst: ok, so we can run them in our infra? [12:20] functional tests are a bit iffy, at best its like 90% that will pass [12:21] do we know which ones are flaky? can they be filtered out? [12:21] didrocks, I already hooked up unit tests, but under autopkgtest since they require install tracker to run [12:22] they are not really unit tests I guess then, but sounds good to me :) [12:22] didrocks, I know which ones are flaky right now, but seems like not properly maintained upstream [12:22] darkxst: they don't plan to maintain them? not sure it's a really good sign of quality for main then [12:23] didrocks, seems to be a case of too many tests [12:24] darkxst: would be a good contribution to upstream as well to try to separate reliable ones to flaky ones [12:24] unit tests are maintained, functional tests can lag a bit [12:24] and we can hook them up in autopkgtests [12:25] yeh though I don't quite get the distinction here, from my quick look through test cases seems to be alot of duplication between unit tests and functional tests [12:26] darkxst: so, if they don't want to maintain them, why not proposing to remove them upstream? [12:26] bad tests or unlooked tests are worse than not having tests [12:28] didrocks, unmaintained is probably not the right word, but a functional test failure doesn't seem to be a release blocker [12:29] darkxst: do we know how many tests are failing right now? (like the % on the total number?) [12:29] we can have a threshold in that case [12:30] probably 6-7 fail, 1-2 just hang and timeout [12:31] not too sure how many tests there are, but probably over 100 [12:31] darkxst: ok, maybe in a first approach, let's just blacklist the inconsistant one [12:31] at least, if there is a big breakage, we can then detect them [12:33] didrocks, unit tests should catch that, in fact they already caught there first breakage on build #2 with the glib update [12:34] darkxst: well, it's to protect you, then, if you are fine in having something that can be broken without detection and think the unit tests are enough, good for me if you wired them up [12:37] didrocks, the unit tests, go well beyond unit testing from what I can see. [12:37] darkxst: from your descriptions, yeah, I just gave a quick look and they are clealy not unit tests :p [12:37] so yeah, let's ensure that's wired up and that should be enough [12:39] didrocks, they are https://jenkins.qa.ubuntu.com/view/Utopic/view/AutoPkgTest/job/utopic-adt-tracker/ [12:40] darkxst: are they? I keep seeing: # TOTAL: 0 [12:41] but yeah, some logs shows them passing [12:41] however, the counter doesn't work :/ [12:43] right, the counter on stdout is broken [12:44] * didrocks sighs [12:44] and upstream would say it's not important as well? === MacSlow|lunch is now known as MacSlow [12:53] didrocks, if a "unit" tests fails that is important, and it should trigger a failure regardless of the counters [12:54] darkxst: yeah, but it's still a sign that details don't seem to matter to upstream, which is reflecting on the global quality of the product [12:54] that's not a very good sign for something we have to commit to main [12:58] didrocks, it could also be fallout from switching to the newever glib.mk testing framework [13:02] tracker 1.0 branch is still actively maintained, and should be for the foreseeable future, and api/abi should be pretty stable from here on in [13:02] yeah, I'm more concern about other moving pieces breaking them and not being catchup [13:03] and the fact that the tests are in a seemly weird tests don't give the right signal [13:03] but again, as long as ubuntu desktop doesn't use it and you are happy in Ubuntu GNOME to take the charge of maintaining, I'm fine with it being in main [13:04] desrt: pushed to master FYI; want to cut a release? [13:05] desrt: or perhaps let's wait a bit until cgmanager is in Debian, perhaps we'll get some more fixes by then [13:08] didrocks, tests aside, it really is way ahead of where it was a year or two ago, [13:08] (stability wise) [13:11] darkxst: yeah, just fix the other issues then, and I'll +1 [13:12] pitti: i wouldn't mind getting the Stop thing working either... [13:12] pitti: could do it by walking the cgroup tree to find the group with the matching name.... [13:13] desrt: I think this might not even be an issue of removing the cgroup, but just to invoke the cgroups' cleanup notification? [13:13] didrocks, sorting out the COPYING.LGPL that should be in next weeks 1.0.3 [13:13] great :) [13:13] pitti: from what i understand, the entire point of this cgroup exercise is to have a reliable way to kill off a group of processes when the session exits [13:13] so ya.... implementing Stop is kinda _the_ point [13:13] desrt: if I log out of a session, the cgroup (like session-c4.scope) goes away, but it's still in loginctl as "State=closing" [13:14] desrt: yes, of course; but that's not even what I meant [13:14] hallyn wanted it for another reason to do with users being able to create containers.... so he's happy now, sure.... but in theory, i think Stop is probably pretty important [13:14] shlibs.local looks all correct to me, I don't see anything obsolete in there, and ricotz says the rpath thing is a non-issue [13:14] hmmm [13:21] darkxst: why is it a non issue? We don't accept rpath in ubuntu apart if there is some real good reasons to have it [13:21] ricotz, ^? [13:23] didrocks, i commented on the bug, which was not completely right [13:24] ricotz: ah fine, I'll have a look once you reassign (and the rest is ready) then [13:24] although there are no rpath references in the source [13:24] and config.rpath is created due some outdated autotools mojo [13:24] git master doesnt pick it up anymore [13:24] great, so let's wait for next release :) [13:25] hmm, i guess we are going with 1.0.x here [13:25] git master is 1.1.x though [13:25] mterry: welcome back! had some nice vacations? [13:25] pitti, yeah I was just camping with friends and family. relaxing :) [13:25] pitti, hopefully that deja-dup upload clears the blockage [13:26] mterry: and thanks for the deja-dup fix [13:26] (thanks to Laney) [13:26] mterry: oh, I was actually concerned the other way around -- glib breaking deja-dup :) [13:26] mterry, hey ;-) [13:26] that glib is removed anyway [13:26] pitti, :) no thankfully just a silly warning message [13:26] I kind of like having a working backup [13:26] Laney, oh ok [13:26] mterry, I +1ed your settings changes [13:26] but yeah, it has a new warning which makes some tests sad [13:26] so still good [13:27] mterry: btw, I switched from rsnapshot (which I'd used for like 10 years) to duplicity a few months ago; love it! [13:27] seb128, awesome thanks! Now the blocker for that whole silo is the underlying config changes to livecd-rootfs [13:27] pitti, oh really! Neat [13:27] mterry: camping sounds great, had nice weather? [13:27] pitti, yeah it was actually nicely cloud-covered -- no rain, but not too much sun [13:28] mterry: yeah, I configured deja-dup for my wife's computer a year or so ago; not quite flexible enough for me as a UI, but the duplicity backend is great [13:28] pitti, yeah you aren't DD's target audience ;) [13:28] mterry: (not a critique -- if the UI was so flexible it would be too hard for most users :) ) [13:29] mterry: well, actually, the only feature I'm missing is specifying an --exclude-globbing-filelist [13:30] pitti, mmm yes, I believe there is a feature request bug for that that I've been ignoring for the past many years [13:30] mterry: yeah, no biggie; I like cron and CLI [13:31] mterry: but I was quite amazed how easy and fire&forget it is, on my wife's PC [13:31] mterry: she doesn't have quite my discipline with keeping $HOME neatly organized, and I don't want to carefully go and blacklist stuff, so DD is just great [13:31] :) === m_conley_away is now known as m_conley [14:04] since we got half of gstreamer 1.4 already, I'm going to look at the rest [14:04] happy days [14:06] Laney, thanks [14:12] Laney, shrug though on the people complaining about that sync [14:12] yeah... [14:13] it's not even a transition, just a bug [14:13] not that anyone even investigated it yet [14:13] no [14:13] they claim they are too busy for that [14:13] how constructive! [14:14] back in the day we'd have taken all of the point releases, then we'd have found the issue sooner ;-) [14:17] hehe === alan_g is now known as alan_g|tea === alan_g|tea is now known as alan_g === Sweetsha1k is now known as Sweetshark [15:19] seb128: would https://launchpadlibrarian.net/180511188/gst-plugins-base1.0_1.4.0-1_1.4.0-1ubuntu1.diff.gz and https://launchpadlibrarian.net/180510137/gstreamer1.0_1.4.0-1_1.4.0-1ubuntu1.diff.gz be ok with you? I'm working toward makeing apt-get install ubuntu-sdk-libs-dev:$target work and the git stuff is not multiarch enabled yet [15:20] mvo_, what do you mean by git stuff there? [15:20] seb128: gir stuff [15:20] seb128: sorry, autotyping :) [15:20] oh [15:20] how does it work for other libs? I can't believe none of our gobject libs are multiarched [15:20] but having -dev depends on the gir is standard rule in Debian [15:21] seb128: the libs are multi-arch, this is about being able to install a cross build environment, so the -dev lib needs to be multi-arch friendly [15:21] seb128: sorry, I should have mentioned that in the changelog [15:22] mvo_, well, as said you are going to have that issue with gobject based lib [15:22] dev depends on gir is standard [15:22] also we are in sync with debian for those packages [15:22] those changes look lik a local hack rather than a proper solution? [15:23] seb128: the proper solution is to make gir1.2-* package multiarch capable [15:23] seb128: but I'm taking one step at a time [15:24] seb128: the problem is that gir has arch dependant typelibs but a single path [15:24] seb128: we need triplet paths [15:24] mvo_, can you talk to slomo about getting that one step in Debian so we can stay in sync? [15:24] seb128: but that probably(?) means that all the loads need adjustment? i.e. perl/python etc. or is there a single loader codebase :) ? if so, I'm happy to start hacking on that [15:25] not sure if anyone has looked at multiarching gir, did you try asking pitti? [15:25] he knows that stack quite well since he's one of the maintainers [15:25] I haven't looked into this [15:25] k [15:25] mvo_, still, changing that depends to a recommends look wrong, so -1 from me [15:26] pitti: do you know if there is a single loading code or if each binding implements this? the path lookup [15:26] as said you are going to hit that issue with any lib shipping a gir [15:26] seb128: well, we are not shipping any other gir libs in ubuntu-sdk-libs [15:26] mvo_: no, (hopefull) all bindings will use libgirepository, which is built by gobject-introspection [15:26] seb128: those are the only offenders [15:26] and that seems a workaround, likely to create bugs and add divergence on debian [15:26] mvo_: so only that library does the .typelib lookup and parsing [15:28] seb128: well, patching libgirepository and multi-arching the gir packages is also a diversion and may trigger bugs (and a even bigger one). the alternative might be to simply drop the libs from ubuntu-sdk-libs-dev [15:28] pitti: thanks, I check that out then [15:29] mvo_, well, multiarching should be done in debian for those [15:29] but otherwise +1 from dropping the depends then [15:29] seb128: right, I think we agree that the proper fix is gir multiarch but that is some way away [15:30] mvo_, sorry, it's our weekly meeting time [15:30] hey desktopers [15:30] ok [15:30] it's meeting time ;-) [15:31] qengho, Sweetshark, Laney, tkamppeter, desrt, attente, larsu, didrocks, FJKong, willcooke, hey [15:31] \o/ [15:31] let's start by welcoming our new manager ;-) [15:31] willcooke, welcome on board! [15:31] willcooke: what's up? [15:31] thank you very much [15:31] welcome willcooke! [15:32] Hi guys, sorry I still havent got round to meeting you all 1 to 1 yet. [15:32] I will get that sorted as a matter of priority [15:32] hey! [15:32] o/ [15:32] welcome willcooke! [15:33] thanks Sweetshark [15:33] willcooke, did you want to say anything else before we start our usual team roundtable? [15:34] we have a slot after the roundtable for extra discussions ;-) [15:34] I don't have much to say right now, other than what I've already said about 1:1 meetings with everyone. [15:34] k, makes sense [15:34] willcooke, thanks [15:34] ok, let's get started then [15:34] qengho, hey, you can start ;-) [15:36] no qengho? [15:36] Sweetshark, your turn then! [15:36] - ongoing: create a docker/saltstacked ubuntu tinderbox for easy deployment upstream, using http://tinderbox.libreoffice.org/MASTER/status.html and http://skyfromme.wordpress.com/2013/08/03/tb3-more-efficient-tinderboxing/ [15:36] - ongoing: reenabling tests on 4.3 -- might be fixed by --disable-mergelibs [15:36] - travel coord. for Bern (LibreOffice conference) and Montevideo (CISL Avanza) [15:36] - cleaned up/merged changelogs for 4.2.4 security SRU [15:36] - waiting for 4.2.6.2 tag (today) for 14.04 SRU [15:36] - looked into bug 1170035: unfortunately a mess to solve with -l10n split out and *.desktop files having all translations in one file (which is not in an -l10n package) :/ [15:36] - various bugzilla and launchpad prowling [15:36] - leading the ESC call, provided some LibreOffice 4.2 vs. 4.3 bug stats: http://nabble.documentfoundation.org/bjoern-michaelsen-canonical-com-minutes-of-ESC-call-2014-07-17-td4115852.html -- likely will blogify those ~around 4.3.0 upstream release [15:36] - trying (and mostly failing) to ignore the heat [15:36] EOF [15:37] Sweetshark, let me know when you get some SRU ready/if you need some sponsoring [15:37] didrocks, seems like you are not alone playing with docker ;-) [15:37] looking so :) [15:37] seb128: sure, will need some time riping in the PPA still [15:37] Sweetshark, thanks [15:38] k [15:38] Laney, hey, your turn [15:38] hai [15:38] • Sponsor xdg-utils for ochosi, provide advice on the sru process [15:38] • Update glib, find bugs (pcre3 in Debian breaks tests, gmutex native impl breaks shotwell) [15:38] ∘ Merge tracker for test fix due to new warning [15:38] ∘ Patch deja-dup likewise [15:38] didrocks: kepp you hand of my docker *sternlook* [15:38] • Find patch for harfbuzz/myanmar crash (#1303516), upload + SRU [15:38] • u-s-s [15:38] ∘ Work a lot on test fixes, seems green now [15:38] ∘ Fix, not optimally due to SDK bugs, bug with scrolling in sound panel [15:38] ∘ On jd_strand's advice, revert apparmor profile to fix sound playback [15:38] ∘ Increase spacing in the main grid [15:38] ∘ Reviews/pushback on scope creep in the schemas [15:38] • DMB: Finish zequence's mail application, flush a large set of updates to all flavour packagesets - the script hadn't been run for several months if not a year. [15:38] • Starting to look at gst 1.4 now that we got the base packages in Utopic by syncing (good done, bad not yet) [15:38] • Off Thursday + Friday this week, then GUADEC, then a trip to China(!) [15:38] ⚘ [15:39] Laney, enjoy the days off before the travelling ;-) [15:40] well done on the harfbuzz patch, installer bugs are not nice [15:40] we've already consulted the Good Beer Guide :P [15:40] Laney, thanks! [15:40] lol [15:40] good thinking ;-) [15:40] Ah, I missed the beep. I'm here. [15:40] qengho, ok, your turn then! [15:40] - chromium-browser 36.0.1985.125 preparation, [15:40] -- fixing webapps patches for package, [15:40] -- and fixing mir patches for testing. [15:40] EOF [15:41] qengho, what's the status for chromium 36? it has been some week where you say it's ready for upload [15:41] is that blocked on the security team? [15:41] That was 35. Security has it. [15:41] not utopic though?! [15:42] oh, well, not other series either [15:42] https://launchpad.net/ubuntu/+source/chromium-browser [15:42] so blocked on security? did you try to ping them? what is blocking? can we help to unblock? [15:42] That's true, I could have found someone else to upload to U. :( [15:43] I hear from CCC on #security that it's on his list, but I haven't harassed him. [15:43] ok [15:43] let's wait a ping and ping jdstrand or chrisccoulson then [15:43] thanks [15:44] qengho, seb128 - let me know if you want me to take care of that [15:44] willcooke, well, I don't know for how long it's blocked on them but it seems to be a while, that upload was ready to go a week ago already [15:44] willcooke, if you want to check with jdstrand was is blocking that would be nice [15:44] seb128, ack [15:44] Thanks. [15:45] qengho, thanks [15:45] willcooke, thanks as well ;-) [15:45] np [15:45] tkamppeter, hey, your turn [15:46] no tkamppeter? [15:46] desrt, ok, your turn then [15:47] hey [15:47] was on vacation all of last week, and i don't remember _too_ much about the week before that [15:47] but git log says that i landed the gmutex-natively-on-linux stuff along with the big performance improvements to GDBus for fixed-sized arrays [15:47] i'd say that this sounds nice, but Laney already tells me otherwise ;) [15:48] did you expect it to make working apps into non working ones? :) [15:48] even if it turns out to be shotwell being wrong [15:48] :) [15:48] after that i started taking larsu's comments about my markup parser stuff into account and did a fair bit of work for fleshing out the interface to something that we could think about landing... found a few thorns on the API meanwhile, but we talked about that today and came up with a good plan [15:48] Laney: who knows.... maybe out there somewhere is a non-working app that now works properly, too :) [15:49] anyways, the comments on the bug said that the speed improvements can be quite nice [15:49] so it'll be a win in the end [15:49] Laney: thanks for the support :) [15:50] eof for me. [15:50] desrt, thanks [15:50] attente, hey, your turn [15:50] - system-config-printer: Work on driver download hanging when initiated out of the D-Bus service (used by other printer setup tools). [15:50] - Mentoring of GSoC students, midterm evaluations [15:50] - Some organizational stuff for OpenPrinting Summit. [15:50] - Bugs. [15:50] :) [15:51] lol [15:51] tkamppeter, thanks [15:51] attente, your turn now then I guess ;-) [15:51] hey hey [15:51] trying to do dynamic language switching again, got it somewhat working after using an insecure apparmor policy [15:51] so working on a separate system settings service that will live on the session bus and provide read-only access to the current accountsservice user's properties [15:52] attente: why not in something like g-s-d or so? [15:53] on the device? [15:53] hmm. don't we have something like this there? [15:53] seems weird that we would have a separate daemon only for this one thing. [15:54] let's talk later... [15:54] yeah, we don't have a g-s-d equivalent in there [15:54] we have indicator-* there though [15:54] but might be a good topic for when we are together at GUADEC [15:54] one more won't hurt! [15:54] seb128: good call [15:54] [15:54] larsu just got an elbow in the side for that comment.... before i saw the ... [15:54] attente, I did a landing of u-g-m to utopic with the 3 branches that were ready [15:54] desrt: I forgive you [15:55] seb128: great, thanks! [15:55] attente, the SRU are still on my list, trying to get to those this week [15:55] yw [15:55] attente, anything else? [15:55] seb128, that's it from me [15:55] ok [15:55] attente, when do you travel? [15:56] * Sweetshark hands out a new set of elbows to spend to everyone. [15:56] not a topic meeting so moving on ;-) [15:56] larsu, hey, your turn! [15:56] hey [15:57] I wasn't very productive last week because I didn't feel very well [15:57] seb128: friday evening, arriving sat morning [15:57] I did manage to look over attente's branch and test it on my machine [15:57] thanks again attente [15:57] attente, k, safe travel! [15:57] also talked to Company how we could have a better situation with o-s, but he says it'll be hard [15:57] larsu, oh, I hope you are feeling better [15:57]