[05:10] hello [05:40] Oh hello hikiko === faenil is now known as faenil_ [07:27] good morning desktopers [07:28] xnox, seems like those firefox ppa builds are still unhappy :-( [07:29] yeah, i know =( [07:29] oh, and hey xnox ;-) [07:29] i'm not sure what i'm doing right or wrong. will be doing "interactive" build on s390x to see what's happening, becuase it looks like it's not doing what i thought it was going to =) [07:30] chrisccoulson might be able to help you maybe? he knows the build system at least [07:31] The build system has changed a lot since I last worked on firefox [07:31] You can ask me anything about the chromium one though ;) [07:31] chrisccoulson, which one of the two is a more complete linux distribution? [07:33] heh [07:33] The chromium one is a lot simpler [07:35] morning, currently upgrade to Yakkety fails with bug 1611256, could someone have a look? [07:35] bug 1611256 in ubuntu-release-upgrader (Ubuntu) "X to Y upgrade fails with gconf2 depends on python3:any; however: Package python3 is not configured yet." [Critical,Confirmed] https://launchpad.net/bugs/1611256 [07:37] cyphermox, doko, ^ [07:37] hey jibel [07:37] Salut seb128 [07:43] morning guys [07:43] chrisccoulson, so even though i have patched icu.m4, the *b.dat is not getting picked up [07:43] so i'm guessing that moz.build is not actually using that.... [07:48] hey Trevinho [07:52] hey seb128 ... [07:52] wow, I'm at my parents' these days... And... no electricity for three hours today --- [07:52] :-/ [07:52] urg [07:52] how come?! [07:53] works... [07:53] or doesn't [07:53] thank god I've a mobile 4g modem, so I can have high(er) speed connection anyway [07:53] I just hope batteries would last, but they should [07:54] or... I'd just move to somewhere else :) [07:56] chrisccoulson, or maybe i should not use autoconf2.69 manual when building with autoconf2.13.... [07:57] o hai everyone! [07:57] forgot to mention in the meeting yesterday: GUADEC starts tomorrow! [08:00] hey desrt, I guess you are going? is anybody else from the team going? [08:00] uh... larsu? :) [08:01] haha [08:01] * larsu smiles [08:01] we're kickin' it oldschool. sharing a hotel room and everything. [08:01] oh... guadec! [08:03] nice [08:03] how big is GUADEC this year? [08:03] no idea. [08:03] "not as big as the universala kongreso" [08:03] well at least 2 people going we can say ;-) [08:04] guten morgen [08:04] reminds me i need to do my duolingo1 [08:04] morning, laney :) [08:04] you're not on the guadec train, are you? [08:04] nein [08:04] i'm going to be at the seaside [08:04] too bad. [08:04] hey Laney [08:05] hey seb128 [08:05] how's it hanging? [08:05] probably much smaller even than in recent years :( [08:05] seb128: 4 people from Lyon going! :) [08:05] hi Laney and seb128 and Trevinho and desrt and didrocks [08:05] hey larsu! [08:05] hey larsk [08:06] didrocks, #gnomefr is Lyon based indeed [08:06] jbicha, can you email ubuntu-desktop@ asking for your ubuntu-desktop membership to be reactivated? ;-) [08:07] Laney: oh, who is that? The person you hired to replace me? [08:08] the evolved form [08:08] (too much pokemon) [08:09] btw, I stumbled again upon this yesterday: http://aruiz.synaptia.net/siliconisland/2012/08/why-cant-we-be-friends.html [08:09] see how larsu looks young! :) [08:09] or rather "looked" [08:10] haha [08:10] I don't see any difference [08:10] haha awesome! [08:10] we all look young on there [08:11] Laney hadn't been born at that time [08:11] true facts! :) [08:11] maybe I was conceived there [08:11] * Laney shudders [08:12] oh and where is desrt? [08:12] lol [08:12] :( [08:13] yeah, unsure… [08:13] i think i was dreaming up the dbus protocol for GMenuModel or something [08:13] actually, more likely thinking about action muxers [08:13] rock and roll [08:16] wtf [08:16] someone just retried the failed build of ffmpeg so fast that by the time I clicked the link from the email it was already gone [08:17] that was the second failure too [08:17] third time's the charm [08:17] somebody who just joined IRC? ;-) [08:17] * Laney has joins off [08:18] but I uploaded it, so only I would have had the email [08:18] * Laney confuse [08:21] Laney: don't worry, it's Wednesday [08:21] * larsu hopes to add to the confusion [08:24] * Laney goes back into the cave [08:24] so.... it's fivethirtyeight season again [08:25] i'm really going to enjoy opening this page every morning for the next couple of months [08:40] chrisccoulson, any pointers as to how to pass --disable-ion on powerpc & s390x as a.... configure flag? [08:41] ./debian/rules is madness in firefox package [08:41] xnox, debian/config/mozconfig.in [08:41] if we're passing things like that, we should really be turning off the official branding too [08:42] well. [08:42] upstream has no ion support for powerpc & s390x, so they should be doing the right thing by default. [08:43] i might submit upstream patch to their mozconfig too. [08:43] i'm not uploading this, just trying to make it build. [08:43] however, imho ion jit should just use gcc atomics and stop pretending they don't work on powerpc & s390x [08:57] chrisccoulson, most of our ports arches are not supported upstream and are tier-3 [08:57] the littleendian ones have atomics support implemented upstream, e.g. for ion jit support [08:57] however there are no atomics support upstream committed for powerpc (32bit) nor for s390x. [08:58] however fedora do build s390x firefox.... http://s390.koji.fedoraproject.org/kojifiles/packages/firefox/48.0/5.fc26/data/logs/s390x/build.log [08:59] with-system-icu [09:00] yet with official branding [09:03] doookoooo [09:03] grrr, just got an update + sru rejected because of an upload not commited to the package vcs [09:30] morning from London [09:30] urgh, lag [09:31] hey willcooke! [09:31] new wifi card or still experiencing grumpy nm at the office? ;-) [09:31] made good time this morning 1hr 15mins [09:31] I've got an ethernet cable :) [09:31] it's the future [09:35] :-) [09:38] errands and lunch, bbl [09:38] seb128, waaaiit [09:39] yes? [09:40] well, seb127 could better tell what exactly is wrong [09:41] doko, https://launchpad.net/ubuntu/+source/vino/3.8.1-0ubuntu9 was not commited to lp:~ubuntu-desktop/vino/ubuntu (which is the vcs referenced in debian/control), so I did bzr pull/change/commit/push/dput -> got rejected email because the version was existing in the archive with different content [09:42] yes I could have checked the archive first [09:42] but ideally the vcs is uptodate and we avoid those issues [09:42] I was told to leave vino alone, and this one just is a no-change upload ... [09:43] unsure what you mean leave it alone [09:43] but if you do an upload please update the vcs [09:43] even in no change uploads [09:43] because otherwise we hit cases like I just described [09:43] I had to uncommit, merge your work, do mine back and push overwrite [09:43] anyway lunch time [09:43] bbl [09:44] no way, if this can't be automated [09:48] chrisccoulson, seb128 - i give up on s390x/powerpc for now. It builds in fedora, doesn't build in debian. [09:48] need to dig more into what is borked. [10:16] hey desktoper [10:16] s [10:16] ouch, no electrictity till 14:00... What bastards! :/ [10:17] building against cups 2.2~rc1-1 seems broken "/usr/include/cups/ppd.h:40:22: fatal error: raster.h: No such file or directory" === davidcalle is now known as davidcalle|afk [10:50] xnox, thanks === hikiko is now known as hikiko|ln [11:18] main laptop power is over... Well, lunch time anyway... Hope to get electricity back in a bit [11:18] good luck! [11:33] Trevinho, enjoy lunch! [11:36] ricotz, hey, that's worth mentioning to tkamppeter when he gets online [11:38] ricotz, do you have a build log showing the issue or a bug report? [11:40] seb128, I see, not anymore since I retried the builds without -proposed enabled [11:40] k, what source was it? [11:40] gtk+3.0 [11:41] let me try in a pbuilder [11:41] it was 3.21.4 but I assume 3.20.6 will fail as well [11:42] likely if it's a cups bug [11:53] willcooke: http://www8.hp.com/us/en/products/proliant-servers/product-detail.html?oid=7152904#!tab=features <- requesting this as next hardware replacement [11:53] willcooke: (including the private power plant needed for it) [11:55] Sweet5hark, :DD === hikiko|ln is now known as hikiko [12:07] * Laney cries [12:16] seb128: ricotz: see https://bugs.debian.org/833889 (I accidentally filed against the wrong pkg yesterday!) [12:16] Debian bug 833889 in libcups2-dev "libcups2-dev: Missing dependency on libcupsimage-2-dev" [Serious,Open] [12:18] jbicha, thx [12:26] jbicha, k, I had the same conclusion ... [12:26] Laney, what's the issue? :-( [12:28] jbicha, do we really need to make gnome-session depends on the new gnome-shell? [12:28] that blocks gnome-session gnome-keyring etc in proposed [12:28] can't fix ffmpeg [12:28] :-( [12:30] seb128: I haven't tried yet but https://tracker.debian.org/media/packages/g/gnome-session/changelog-3.20.0-1 [12:31] if it's only a .desktop name that's easy to patch [12:39] yes or we can try to fix gjs [12:40] .... and FUIT LUX! [12:40] last week, I triggered an autopkgtest.u.c from my PPA, I saw the test on running.html but I have no idea where PPA tests results go… [12:41] you have to explore swift yourself [12:41] AFAIK [12:41] https://wiki.ubuntu.com/ProposedMigration/AutopkgtestInfrastructure#Example_queries [12:41] * Laney is running gjs tests locally now [12:44] Laney: thanks, found my test result [12:45] gjs passes its tests on my computer :| [12:53] here too [12:54] although I rebuilt it [12:54] let's try with the binaries in the archive [12:54] * Laney goes off to mooch around the library while that grinds away [12:55] this new office is good [13:00] is that the one with pants wifi? ;-) [13:04] better today [13:04] huh [13:04] so gjs passed when I rebuilt the source, but it failed when I used the archive binaries === mterry_ is now known as mterry [13:06] Laney: ok, how about we retry rebuilding gjs then, since gjs was built we got gobject-introspection 1.48 [13:07] worth a try [13:07] you want to do that upload? [13:07] sure, I'll do it now [13:07] cool [13:08] would be nice to understand this interaction enough to know what's making it fail [13:08] ho hum [13:09] i'll be back in a bit [13:27] jbicha, your evolution merge request is in the sponsoring queue but you should have access to merge/upload that one (upload is blocked on eds but you can probably already merge in the vcs) [13:29] oops, I didn't intend to propose that for merging since I can upload myself [13:29] but no, I can't push to ~ubuntu-desktop branches yet [13:31] oh, right, vcs doesn't match upload acls [13:31] I overlooked that [13:31] you can upload but not commit [13:31] :-/ [13:53] first half of august, suposedly the hottest time of the year. [13:53] outside: hail [13:56] Sweet5hark, yeah, but hot hail ... [13:57] jbicha: seb128: It's probably a good idea to push the branch under ~motu if it's moved to universe, and fix Vcs-Bzr in the package [13:59] pzzt! eis und heiss [13:59] Laney: I don't think ~motu is a good place for it, see how empty https://code.launchpad.net/~motu is [13:59] I doubt anybody browses that to find branches [14:00] There's history to keep here, and you might as well make it so that the branch is writable by those who can upload [14:00] will merge proposals spam all of ~motu? [14:00] There's an email address set for that team, so nope [14:00] just like I never get emails for ubuntu-desktop [14:04] jbicha: sorry, can't +1 your ubuntu-desktop application because I can't type my passphrase [14:04] * Laney has failed 9 times now [14:04] 12th time is a charm [14:06] * Sweet5hark hums "Im going slightly mad". [14:15] Can unity7 skip the unity8 autopkg tests here? http://people.canonical.com/~ubuntu-archive/proposed-migration/yakkety/update_excuses.html#unity [14:16] Sweet5hark: That's because global warming is a lie, told by big hail to distract you. [14:18] tedg, Laney might be able to help you (or anyone who has upload right might but I never remember if that's the case and how) [14:19] tedg: I'd love to... But.... no. At least unless someone would push a button [14:19] Laney: can maybe? ^ === davmor2_ is now known as davmor2 [14:20] It's the release team, but I don't want to [14:21] Get the Unity 8 team to fix the test [14:21] Saviq, ^ [14:21] I said last time I did it that I wasn't going to do it again [14:21] Laney: I've been told they've been fixed in some silo for long time... but [14:21] Laney: the thing is that is blocking unity7 too.. :-( [14:21] It needs to be a priority [14:22] which... blocks other works.. [14:22] Trevinho: Sorry that you have to feel the pain, but it's better in the long run [14:22] so you want us to annoy u8 people? [14:22] the issue is that it put the annoyance on the wrong people [14:22] Laney: I mean, I agree... but it's just we're affected by something which they aren't... [14:22] u8 team doesn't care about u7 being blocked [14:22] They do now [14:22] that's thet thing.. how seb128 said. [14:23] how do they do? [14:23] go direct the pressure at them instead of the release team [14:23] then they will care [14:23] I pinged Saviq yesterday [14:23] yay, /me getting bashed [14:23] not that I couldn't try to fix them too, but maybe it would take more time [14:23] it's always easier to disable, skip or ignore tests [14:23] but it's not often the right thing to do [14:23] in an ideal world we sure would fix every issue [14:24] guys, if it was *easy*, we'd have fixed them long ago, but most of those failures only happen, intermittently, on britney, not even in our CI [14:24] any case, OK, I will put more pressure on getting those to be more stable [14:25] Saviq, thanks [14:26] We've also see tests that only fail intermittently on britney and no where else. It seems that environment is different somehow. Causing people to disable autopkgtests :-/ [14:26] Trevinho: You should do what you said yesterday and split the dependency [14:26] Trevinho, meanwhile what work does it block? [14:26] thanks [14:26] I'll retry the tests meanwhile, maybe they work this time [14:26] Laney, that's cheating as much as skipping the test :p [14:26] you are virtual skipping the test by another mean [14:26] virtually [14:27] Sort of [14:27] if they go green you know they aren't totally broken [14:27] right [14:27] I meant the package split [14:27] not the retry [14:27] well [14:28] that just let unity7 in an remove the insensitive to fix unity8 [14:28] the point of testing unity8 when unity7 gets uploaded is that there's a relationship there [14:28] a unity 7 upload could break unity 8 [14:28] but if you remove the relationship then that can't happen [14:28] right, though in the current case it's not the unity7 schemas breaking the test [14:29] I know! [14:29] your point was that not skipping is putting incensitive to fix unity8 [14:29] seb128: Is it possible to get someone from desktop/foundations to look into britney's test environment. I know it's a bit of a science experiment, but it seems like the reliability could be improved. [14:29] but splitting is removing that [14:29] anyway just pointing it out [14:29] the separate schemas thing is still good imho [14:29] It'll remove it from unity 7 [14:29] but other projects will still have it [14:29] which is not much different from forcing the current unity7 upload in [14:30] It's not an argument which says that random packages should block on random other ones [14:30] Just to spread the pain around [14:30] It's not pain for pain's sake [14:31] right [14:31] well I think we each understand each others position [14:31] so no point arguing here [14:31] Trevinho, why is the migration blocking more work? [14:31] you could force merge the silo and do another landing in proposed no? [14:32] seb128: I am trying to rebuild the systemd unit silo, which has Unity in it. [14:33] seb128: ah, if I can force-merge the silo is fine... it was more blocking the SRU maybe [14:33] tedg, I doubt you are going to get anyone working on infra before ff, especially with pitti on holidays [14:33] Trevinho, I told you yesterday that SRU team doesn't care about migration [14:33] having it in yakkety-proposed is good enough [14:33] the rule is about making sure things land [14:33] once they are in proposed they are in the pipe [14:34] there is no way they are not going on [14:34] hmmm, /me went from "it doesnt work and I dont know why" to "it works and I dont know why" #notsureifimprovement [14:34] (well except if we delete the proposed version, but we usually don't) [14:34] seb128: Yeah, I'm more interested in getting it on the TODO list long term. Not the first time it has been an issue. [14:34] seb128: yeah, i remember that... It was just that pitti asked why the fixes weren't released in yakekty yet... So... [14:34] tedg, yeah, maybe talk to Steve... [14:34] seb128: K, will do. [14:34] just for you guys, Albert just found out why at least one of our tests was flaky https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1610165 [14:34] Launchpad bug 1610165 in ubuntu-ui-toolkit (Ubuntu) "qmltestrunner::DashContent::test_mainNavigation() does not always pass" [Undecided,New] [14:34] Trevinho, I don't think he said that for things that were in yakkety-proposed but not migrated [14:35] Saviq, thanks! [14:35] Sweet5hark: as long as it keeps working, I don't see a problem there :) [14:35] seb128: ok, I can ping sru team again [14:35] Sweet5hark, working is always an improvement ;-) [14:37] seb128, that said, for a long time now our autopkgtests in yakkety because of migration issues and that they are built with proposed but tested without it [14:37] seb128: as for finalizing the silo, I'm getting error 401... Unauthorized [14:38] seb128: should i ask trainguards? [14:38] looking at http://autopkgtest.ubuntu.com/packages/u/unity8/yakkety/amd64/ we've had a ~good track record before 07.21 or so [14:39] Trevinho, yes please [14:40] I feel bad now [14:40] Maybe a force-skiptest for unity7 wouldn't be so bad [14:40] ... [14:41] sooo, how does gtk handle precedence if a style context has multiple classes added? [14:41] Laney, don't, you have a defendable position and you said previous time you wouldn't skip it again so fair enough [14:42] Sweet5hark: Add a provider with a higher priority if you want it to win [14:42] Laney, I don't think it's blocking the world, like if it had been blocking the gtk update to land I would have argued more in favor of unblocking [14:42] seb128: You made me doubt myself! [14:42] :-( [14:42] sorry man! [14:42] I did skip loads of stuff for the qt migration [14:42] Laney: whats a provider and what is it doing in my gtk? [14:42] kubuntu tests are a whole new world [14:42] I can image [14:43] that one is almost impossible to get in [14:43] and that's without being strict [14:43] it would have been impossible by sticking to the "no workaround/exception" [14:43] Sweet5hark: It's a way of giving style information to a context [14:44] for example there are CSS style providers [14:44] which you can create in code [14:45] speaking of qt [14:45] * Laney goes to look at ffmpeg again [14:47] Laney: well, I just wanted to give the toolbar context in libreoffice another class, so that I can override the theming for libreoffice to be a bit different than for others. So I found the place where the context sets the "primary-toolbar" and "toolbar" classes to the context, and wanted to add a "libreoffice-toolbar" and I wonder if order matters in this. [14:48] Sweet5hark: Oh right, I should think that the priority only matters if you have conflicting selectors and rules within them [14:49] You should be able to just add the class [14:50] seb128, Trevinho, Laney, ok so the plot thickens - I assume what you're fighting is http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#unity? in that case it won't pass, courtesy of gcc6 I believe - we need a newer unity8 to be tested there (it's trying to test one from almost a month ago) [14:52] we've fixed the build issue (yes, I know, we shouldn't be building anything in autopkgtests, we're getting there) [14:52] hah [14:52] I'm working on the migration :-/ [14:52] but it never got into release pocket, so britney's trying to test the old one [14:52] yeah that will improve things for sure [14:53] nothing flaky (that we can see) or broken there then [14:54] Laney: right. so I now have a context that has "primary-toolbar" as class (so that other themes work as expected) and added a "libreoffice-toolbar" class so I can fix it for Ambience. Now I wonder how I write a selector that triggers on the "libreoffice-toolbar" class in Ambience and how to make sure it takes precedence over any selector on the generic "primary-toolbar" class in ambience ... [14:55] and so far I havent figured that out yet ... [15:03] Sweet5hark: Make your rule more specific (such as .primary-toolbar.libreoffice-toolbar), then it will be used [15:04] But the other one will match too - that's how CSS works - so you need to set any properties you care about in the more specific one [15:05] Laney: so how does that work for e.g. when one sets "background-image:something;" and the other one sets "background-image:none;"? is something "more specific" than "none"? [15:06] "To determine the effective style for a widget, all the matching rule sets are merged. As in CSS, rules apply by specificity, so the rules whose selectors more closely match a node will take precedence over the others." [15:13] * Sweet5hark suspects he sees heisenbehavior here, and the changes in theming he sees are results of the RNG doing a different throw, not because of code changes [15:13] *sigh* [15:38] Laney, seb128, can you guys ACK https://code.launchpad.net/~vanvugt/gsettings-ubuntu-touch-schemas/fix-1607240/+merge/301605 please? thanks [15:39] sure [15:40] done! [15:52] seb128: Objection re bug #1607937! There is an upload proposal in a PPA. [15:52] bug 1607937 in gnome-font-viewer (Ubuntu Xenial) "font viewer starred from dash: Unknown option --gapplication-service" [High,In progress] https://launchpad.net/bugs/1607937 [15:53] GunnarHj, the bug could be a bit more explicit about that, also I doubt we want a 3.16->3.20 update as a SRU [15:54] seb128: Sorry if it was 'hidden' in the middle of the description. As regards 3.16 -> 3.20, I argued for it. Few and small changes. [15:55] seb128: With that said, I can try to use a patch instead, if you prefer that. [15:56] GunnarHj, the changelog is misleading in this case, the update has new features (support for TTC fonts) [15:56] e.g Support TTC fonts [15:56] ups [15:56] https://git.gnome.org/browse/gnome-font-viewer/commit/?id=e512b0bd53636ebad6ce2428c44bfb02816ed7f2 [15:56] so you at least need a bug with a testcase for that changeset if you want to do the update [15:57] otherwise I would suggest just backportinghttps://git.gnome.org/browse/gnome-font-viewer/commit/?id=acf919c43cf0e7700aaf05931f424843ce4b29dd [15:57] seb128: Then I'll backport that commit. Getting back on the topic. [15:59] GunnarHj, thanks [16:00] seb128: Np. Your response was a calculated risk. ;) [16:00] :-) [16:05] GunnarHj: it's not impossible to argue that that update is bug fixes but it takes more work to be able to persuade the sru team [16:06] well the bug being fixes is fine [16:06] it's just that the upload has more changes [16:06] so those should be documented/explained why they are useful and how to test them [16:08] jbicha, seb128: Let's focus on the main issue. Patch coming soon. [16:08] GunnarHj: yeah, that's a lot less time and effort :) [16:29] k, time for sport and dinner [16:29] have a nice evening desktopers [16:29] see you tomorrow [16:30] bye seb128! [16:31] cya seb128 [16:34] * Trevinho finally killed a tiger mosquito and it's probably the best way to close the day too [16:35] going for a run... have a nice evening folks. [17:07] jbicha: Wrote a patch and confirmed that it fixes the issue. [17:35] Laney: gjs autopkgtest apparently works since gjs and gnome-shell 3.20 migrated [17:57] nice [19:05] Trevinho, hey, have you seen the crazyness thats happening with yakkety + U7? [19:05] compiz is like freaking out (the main UI keeps moving around/flashing?) As if its attempting to move a workspace but doesnt === davmor2 is now known as davmor2_Hols [20:31] hm, has anything changed in compiz in 16.10? i just received a bugreport against my theme that it suddenly has square corners instead of rounded ones...