[02:54] Saviq: happy birthday, btw [02:59] Saviq: yes...happy bday! hope it was a good one! [08:09] guys, does xvfbtestPreviewPayments succeed for you ? [08:09] @unity8: ↑ [08:09] @unity: ↑ [08:10] tsdgeos, LANG=C [08:10] still fails [08:10] very funnily though :D [08:10] ah [08:10] lang doesn't change numeric [08:11] let's see LC_ALL [08:11] right [08:11] yeah that works now [08:11] yeah, LC_ALL sorry [08:11] didn't we decide against this and make tests work with all locales at the end? [08:12] tsdgeos, well, we decided that tests need to make sure they have their env set up correctly... but it doesn't work with our QmlTest macros, /me needs to fix that [08:12] ok [08:12] it's a bug, no worries [08:12] :D [08:15] tsdgeos, got linkedin around? check the new job of this guy https://www.linkedin.com/profile/view?id=58421147 :D [08:16] lol [08:16] just couldn't stay away [08:16] * Saviq likes [08:28] weird, my dash_overview branch has no qmluitests failures [08:28] i thought it'd fail since i increased the number of dash scopes [08:29] let's see what autopilot says [08:41] gotta go taxes, biab [08:55] anyone knows why my EdgeDragArea behaves different in ./run.sh than in make tryDash ? [08:56] ah in try the timer is disabled i think [09:00] i think i still don't understand fully how to use the EdgeDragArea [09:00] too much automagic and too few documentation [09:01] tsdgeos: yeah... there's a special timing source for testing [09:02] iirc you can even control that from your test [09:02] but I don't think I did that [09:02] ever [09:39] mhr3: any ETA for when we can expect a backend for the dash overview thing? [09:40] tsdgeos, could you write down somewhere all the changes you need? [09:42] sure [09:45] mhr3: http://paste.ubuntu.com/7802652/ [09:45] not much [09:48] tsdgeos, what do you expect searches to return? [09:52] i hanven't really explored searches (as in my fake backend doesn't implement them), but i think what we discussed with saviq was just getting categories back and i guess if their names are not "favorites" and "all" i switch to results view? [09:53] tsdgeos, I think it doesn't matter what the categories' names are [09:54] tsdgeos, as long as search query != "", you switch to standard LVWPH [09:54] Saviq: right [09:54] mhr3: so yeah, just return categories as usual [09:54] tsdgeos, but you still want the special ResultsModel in there? [09:55] hmmm [09:55] or will the search be animation-free? [09:55] yes, special resultsmodel there would be great if possible [09:56] Saviq: there's no "Done" on search, right? [09:56] tsdgeos, nope [09:56] tsdgeos, there's a "back" [09:56] ok, one weird transition to worry about :D [09:56] Saviq: the header one [09:56] +less [09:56] no? [09:56] tsdgeos, yes [09:56] ok [09:57] tsdgeos, https://f966f709-a-c881af26-s-sites.googlegroups.com/a/canonical.com/unity8dash/scopes/dash-overview/2a.searchbox%20focus.png [09:57] https://f966f709-a-c881af26-s-sites.googlegroups.com/a/canonical.com/unity8dash/scopes/dash-overview/5.results.png [09:57] tsdgeos, oh, right you don't need that scopeIndex method if the user just taps on the result directly, right [09:57] mhr3, nope [09:57] mhr3: correct [09:57] hmmm [09:57] well [09:57] so no need to mangle search resultsmodels [09:57] hmmm [09:58] let me think :D [09:58] beep [09:58] the end, you can't change that decision now [09:59] i don't need the scopeIndex [09:59] but i need the scopeId role [10:00] so i can know if that search result is a favorite or not [10:00] so if you're going to give me one [10:00] give me both ? [10:00] i could just add that role everywhere [10:01] and waht would it return? [10:01] just "" ? [10:01] same thing it will return for non-scope results [10:01] you don't expect the nike result to actually have a scopeId, do you? [10:02] i expect it to be undefined [10:02] not "" [10:02] but nevermind [10:02] if you want too much, it can be undefined [10:02] it's ok [10:03] although... hmm, not sure if QVariant() is undefined or null [10:03] small details noone really cares about :D [10:04] mhr3, undefined [10:04] good [10:05] mhr3, oh can you show me designs for header links? [10:06] sure, let me find them [10:06] Saviq, page 11 @ https://docs.google.com/a/canonical.com/document/d/1Y9H3SY2uwbeLZSta9bGdLcyGr-tP0LwHyIUigTnieww/edit#heading=h.ihcuz26wprxs [10:07] or 12 [10:07] depending on your resolution :) [10:07] yey good gdocs [10:08] mhr3, ok, that helps, I must've been looking at something old [10:09] mhr3, btw, should we talk about results-in-previews or not yet? [10:10] Saviq, had a quick chat with thomas about it, we're starting to find issues with it :/ [10:11] mhr3, uh oh, so not yet [10:11] Saviq, but the simplest version of it could just could just be {"widget-type": "categories", "query": "foo"} [10:11] mhr3, yeah that's what I thought [10:12] i don't like it cause it does multiple queries, but well, good enough for v1 [10:12] mhr3, except I'd rather only have one category [10:12] I don't think we should ever expect more in there [10:12] hmmm :/ [10:12] mhr3, but ultimately we can just take the first one and ignore the rest [10:13] Saviq, can't you just pretend that you have multiple categories widgets if there are multiple categories? [10:13] mhr3, pretend as in not display them? [10:14] pretend as in expand one widget into multiple [10:14] s/expand/duplicate [10:14] mhr3, we'll get into see more / see less, nested delegate management... [10:14] oh dear, non-terminated s///, that's gonna bug me now [10:17] oh lol [10:17] testDash segfaults [10:17] it does indeed [10:17] i was checking for "Fail!" and that's why i thought all my qmluitests worked ^_^ [10:17] ;) [10:17] good guy CO [10:17] -o+i [10:18] Saviq, my problem is that if it doesn't support multiple categories, you not only have to do 2 queries, but (preview + number of cat widgets) [10:18] Saviq: so the upstart thing seems confirmed only happens in restart and not stop+start, right? [10:18] tsdgeos, yes [10:18] Saviq: can we change our code not to do restart ever? :D [10:19] Saviq, and iirc, every design i saw using that had the artist info + at least 2 other cats [10:19] tsdgeos, it's almost fixed [10:19] ok :) [10:19] tsdgeos, hmm hmmm [10:20] tsdgeos, we don't do restart [10:20] tsdgeos, but maybe if you do stop/start in quick succession, it's triggered as well.. [10:20] mhr3, show me? [10:21] Saviq: i guess you should comment in the bug saying we don't do restart :D [10:21] tsdgeos, is the only way I could reproduce... [10:21] ah [10:21] ok [10:21] tsdgeos, it doesn't die for me during autopilot [10:21] tsdgeos, if it does for you, maybe try and collect the same data I did [10:22] yeah [10:22] let me make sure qmluitests pass [10:22] Saviq, 12 @ https://docs.google.com/a/canonical.com/document/d/16-RX9drsBRWxQq7upD-nlin-D1NodBp8ICiLQwy262c/edit [10:22] and then i'll run autopilot [10:22] and see what's up [10:22] tsdgeos, wanna me build a package for you? [10:22] tsdgeos, with the bugfix? [10:22] Saviq: sure :) [10:25] mhr3, that's a search, not a preview [10:25] Saviq, and the reason why we're trying to turn in into a preview [10:27] mhr3, TBH it starts looking to me like we shouldn't have anything special for previews at all, then, but just allow more card types in standard search [10:28] mhr3, because suddenly they will want see all in previews, category headers behaving like in search etc. [10:29] Saviq, fundamentally the second screen is a preview of the artist, it being search is a design hack [10:30] Saviq, but yes, i don't think anyone thought about the behavior of the results widget in the preview [10:31] Saviq, so leave for next week? [10:31] mhr3, so from my PoV if we want multiple categories in previews (that are in a single widget), that's not happening [10:31] not for RTM [10:32] then some result cards will be interactive:false [10:32] and again without any indication :/ [10:32] mhr3, I was only expecting a single category per widget, no expansion or such [10:33] mhr3, not sure why your last statement? [10:34] Saviq, cause they will still want that screen, so the big artist card will be search result, and it's just an info card, no tap action, no preview [10:34] mhr3, not sure what you want me to say [10:34] "this is crazy"? :) [10:34] mhr3, which part? [10:35] i want you to say that [10:35] but which part? [10:35] the interactive: false? [10:35] all of it really [10:35] misusing searches to display info [10:36] mhr3, that's fine [10:36] I mean sure, that's wrong [10:36] but doesn't weigh into the fact that [10:36] reality is what I said, we can try and squeeze single-category-per-widget, no expansion or anything like that [10:37] otherwise we start nesting [10:37] Saviq, maybe it was supposed to behave like the header link categories [10:37] Saviq, oh right... yea those shouldn't be expandable [10:37] even if they do have loads of results [10:37] mhr3, also no sticky headers [10:38] Saviq, ok, my takeaway from this is that we need to talk to design how they wanted it [10:38] mhr3, inded [10:38] e [10:39] wanna do that asap, or on the sprint? [10:41] sprint [10:41] brrr [10:42] running autopilot locally got 38 failures :S [10:43] but the 2 that failed on CI actually pass :D [10:43] tsdgeos, http://people.canonical.com/~msawicz/upstart+patch.tar [10:44] tx [10:44] someone i misread as rar [10:44] and thught [10:44] what?¿ [10:46] ;) [10:46] Saviq: do local upstart configs (in ~/.config/upstart) override the ones in /usr/share/upstart ? [10:46] Saviq, omg omg, did you fix upstart??? [10:46] dednick, the ones in /usr/share/upstart/sessions/ [10:46] dednick, yes [10:46] mhr3, jodh did (maybe) [10:47] Saviq: hm. so a user could change the unity8 startup environment without root access? hmm.. [10:48] dednick, well, yeah, they could echo "FOO=blah" > ~/.profile, too [10:48] dednick, why? [10:48] dednick, users generally have power over what's happening in their session... [10:50] Saviq: user can change trusted socket to a local file through an environment variable = bad (as far as i can see it) [10:51] dednick, they can also launch apps without confinement [10:51] dednick, we just need to s/trusted/kinda trusted/ :) [10:52] dednick, as long as *apps* can't do it, it's fine [10:52] Saviq: but won't any app have access to ~/.config/upstart ? [10:52] dednick, nope [10:53] dednick, confined apps only have write access to their own folder [10:53] oh. well that's fine then. [10:53] dednick, and they only have read access to the minimal set of devices / files that's required for them to work [10:56] dednick, FTR https://wiki.ubuntu.com/SecurityTeam/Specifications/ApplicationConfinement [11:04] Holas [11:10] Saviq: how do we define "Search screen" for the purpose of disabling dsah overview, current scope has a searchquery != "" ? [11:10] tsdgeos, yeah [11:10] tsdgeos, as everywhere else we do [11:10] mhr3, we should probably have a think about this some time ↑↑ [11:11] mhr3, like do changed departments, filters constitute a search? [11:21] Saviq, yes [11:22] Saviq, for example design specifies that departments are done only when not doing actual search === MacSlow is now known as MacSlow|lunch [11:22] eh, as in query == "" [11:22] but dep_id can be changing [11:23] same for filters [11:25] mhr3, right, I think you might need to expose "isSurfacing" or something... [11:26] tsdgeos, ↑ [11:26] tsdgeos, but TBH I really don't see why the limitation to not go into overview when searching / previewing... [11:26] Saviq: don't look at me :d [11:26] +1, feels very artificial [11:31] greyback, good news [11:31] greyback, https://code.launchpad.net/~saviq/qtmir/gles-sync-20140716/+merge/227000 is what we'll need for qtmir-gles packages whenever releasing them [11:31] s/them/qtmir/ [11:31] greyback, it will pick up the source from which qtmir itself got built [11:32] packaging updates need to be cloned, too, though [11:32] but at least no carrying source around [11:32] Saviq: nice! Has qtubuntu similar? [11:33] greyback, https://code.launchpad.net/~unity-team/qtubuntu/gles-qtcomp-refactor/+merge/226870 [11:34] Saviq: bootiful [11:34] greyback, indeed [11:35] Saviq: so walk me through a future qtubuntu landing. I've a MR for qtubuntu. I create same MR for qtubuntu/gles. I propose both for a silo. Everything just works? [11:35] greyback, you propose a MR for qtubuntu, put it in a silo and build [11:35] greyback, then you get the version qtubuntu got and prep a MR for qtubuntu/gles [11:35] greyback, add to silo, reconfigure, build [11:36] greyback, if you know the version you'll get, you can add them together, but need to build one after the other [11:36] hence the "X-Auto-Uploader: no-rewrite-version" [11:36] greyback, yes, we want the exact same version qtubuntu got [11:36] got it [11:36] greyback, only -0ubuntu1 we might need to bump for no-change rebuilds and such === alan_g is now known as alan_g|lunch [12:19] Saviq: hey. did you know that qt apps segfault on clean exit? [12:19] on phone. [12:19] dednick, they were doing that from time to time, but no, I didn't know recently [12:21] Saviq: if started by commandline. dont know of any apps which actually perform a clean exit though... most are stopped through u8 [12:21] dednick, yeah, confirmed [12:22] dednick, reporting now [12:23] dednick, crash in mirclient [12:23] Saviq: ah, how did you get a trace? mine was useless [12:23] can you post the stacktrace? [12:24] dednick, waited for apport to collect [12:24] dednick, and got it from the .crash === _salem is now known as salem_ [12:24] probably dont have symbols [12:24] dednick, I don't have symbols either... [12:25] Saviq, dednick o/ [12:27] dednick, https://bugs.launchpad.net/ubuntu/+source/qtdeclarative-opensource-src/+bug/1342694 [12:27] Error: ubuntu bug 1342694 not found [12:28] Saviq: thanks. tvoss ^ [12:29] oh wait, gotta subscribe 'im [12:29] done [12:33] Saviq: plz subscribe me too [12:33] greyback, done [12:33] will make public once it retraces anyway [12:34] ta [12:35] /food [12:36] sometimes I get scopes into a state where I can't drag them left/right any more. only up/down === alan_g|lunch is now known as alan_g === greyback is now known as greyback|lunch [13:09] mzanetti, are you getting a crash when you tap on an app in qtcomp? [13:10] dandrader: no [13:10] dandrader: define "app" [13:10] mzanetti, any app [13:10] mzanetti, I just flashed and added the ppa [13:10] on the app surface? [13:10] mzanetti, yes [13:10] dandrader: nope... works fine here [13:11] dandrader: I'm using the silo [13:11] dandrader: you're using the phone-right-edge ppa? [13:11] mzanetti, yes [13:11] hmm... not sure when that was rebuilt the last time [13:11] mzanetti, so I should not use the ppa anymore? [13:12] dandrader: hmm... I switched to the silo lately as kgunn keeps that up-to-date for us [13:12] dandrader: but both should work, you just might need to trigger rebuilds on the packages in the right-edge ppa === dandrader is now known as dandrader|afk [13:16] mzanetti: dandrader|afk ....i saw a lot of choo-choo silo6 build pings....i assume we've rebuilt all at this point....i think [13:17] we need upload lp:~unity-team/qtubuntu/qtubuntu-gles & lp:~unity-team/qtmir/qtmir-gles [13:17] per Saviq's mail ^ [13:18] Saviq: is "ignore twins" ok ? or still we need to upload those packages ? === dandrader|afk is now known as dandrader [13:25] kgunn, where's that silo and how do I use it? [13:25] and what's "choo-choo"? [13:26] dandrader: you naughty boy... https://wiki.ubuntu.com/Unity8/QtComp [13:27] silo 6....all the instructions are there ^ [13:27] kgunn, ok, thanks! [13:27] choo-choo is freenoded irc bot on #ubuntu-ci-choo-choo [13:27] that tells you when stuff is built, reconfig'd, failed, merged, uploaded etc in your silo [13:28] dandrader: you can join it, and it'll tell you when silo6 is being built (i already had added your irc nick to be pung) [13:33] dednick: so does the ~mirconnection fail only for command line launched apps? or all apps ? [13:33] fail....crash rathe [13:33] r [13:35] kgunn: not sure. i'll try [13:38] kgunn: happens with app-launch as well [13:38] kgunn, you shouldn't need to ignore twins any more [13:38] Saviq: ta [13:38] kgunn, I added the MPs that create the twins now [13:39] kgunn, but any time you rebuild qtmir or qtubuntu, we'll need to bump the respective -gles MPs and rebuild those afterwards, too [13:40] Saviq: yep...got it (bump = empty commit? +reconfig) [13:40] kgunn, no, just bump https://code.launchpad.net/~saviq/qtmir/gles-sync-20140716/+merge/227000 or https://code.launchpad.net/~unity-team/qtubuntu/gles-qtcomp-refactor/+merge/226870 respectively [13:40] kgunn, and rebuild [13:40] kgunn, bump == sync the changelog version with what ends up in the silo [13:41] kgunn, basically `dch -v $the-non-gles-version` [13:41] Saviq: oh i see....this is the magic i didn't understand.... [13:41] kgunn, yeah, we just magicked it this morning [13:41] lol [13:41] kgunn, basically what happens: [13:41] kgunn, you build qtmir in silo [13:42] kgunn, then you prepare an MP for qtmir-gles that adds a changelog entry for the same version the silo gave up for qtmir, and update the silo number in debian/watch [13:42] kgunn, then you add that MP to the same silo and reconfigure [13:42] and build -gles [13:43] kgunn, in case you rebuild qtmir, you need to update the -gles changelog bump and rebuild it as well [13:43] yep, that all makes sense....so upon a rebuild it'll actually fail for the gles's [13:43] until you bump/rebuild [13:43] yeah [13:44] you always need to build non-gles first [13:44] kgunn, all this because now -gles packages get the qtmir tarball directly from the PPA [13:44] kgunn, the -gles bzr branch carries no source, just the debian/ folder [13:44] yep...i see it in those mp's you linked [13:45] it's a bit convoluted, but feels sanest and safest, and easiest to follow (because there's an actual branch, with only relevant changes) [13:45] and no need for manual uploads to the silo and such [14:01] dednick: silo18 will work for purpose of progress right? i mean its annoying, but guys doing payment service can progress [14:02] kgunn: we ned to update unity8 to set an environment variable. just testing now. [14:02] tvoss: we're going with the env var rather than the hack in unity8 right? [14:02] dednick, yup [14:03] dednick, please file a bug to remind us of the todo to clean up [14:03] Saviq, what's the story with those -gles packages? What do they do? I'm lost... [14:03] thostr_: ^ one sec, looks like we'll need to add one more mp to that silo [14:03] dandrader, we need them for the i386 emulator [14:04] hmmm.... [14:04] dednick: actually, cemil marked mir silo as to be released (which i suppose is actually ok)....should we just get a seperate silo ? [14:04] dandrader, so they basically are built from the same source that normal packages, but build against GL ES instead of GL (until Qt supports live switching / detection) [14:05] kgunn: yeah, it can be done separately [14:05] tedg: ^ [14:06] just needs a mod to the upstart conf file to use trust sessions [14:06] kgunn, is the qtcomp silo ready with latest changes ? [14:07] mterry, re "isn't that what -f is for"... sure, it is, but it will also load fake unity, fake foo, fake bar... and it will change the current behaviour [14:07] mterry, I basically think, if possible, the default behaviour of ./run.sh at least should be that it doesn't require unlock [14:07] dednick, Can you ping me when that gets a silo. [14:07] ? [14:07] tedg: yup [14:08] om26er: lemme check [14:08] om26er: yes [14:10] Saviq, I'm not thrilled with making "the real run.sh" be a half-mock monstrosity, but sure. easy to do [14:10] mterry, I'm not thrilled about having to enter my password everytime I run it.... [14:11] mterry, and sure, once we get split greeter, we'd just launch unlocked, but until then it would get really annoying to have to unlock every time [14:11] Saviq, I don't see anything in lp:qtmir/gles that makes it build qtmir sources differently (like a compile or config argument or something) [14:11] Saviq, how does that magic happens? [14:12] dandrader, it's greyback's doing [14:12] Saviq, hrm... I don't suppose there's a way to do QML2_IMPORT_PATH but for only one module instead of a directory of modules...? [14:12] dandrader, but the least of it is that it depends on the gles -dev packages http://bazaar.launchpad.net/~mir-team/qtmir/gles/view/head:/debian/control [14:12] mterry, unfortunately not [14:13] mterry, but maybe you can abuse nonmirplugins [14:13] mterry, same as we have the mock Unity.Application [14:13] mterry, which would kind-of make sense [14:13] kind of... :) [14:13] mterry, adverse effect would be an X11 unity session... but no one does that anyway [14:14] Saviq, ahhh... right! [14:14] Saviq, and if they do, they have a greeter right? [14:14] mterry, sure, they would (but they don't) [14:15] mterry, I know what you mean, I understand it's not ideal, but it'll get really really annoying real soon [14:16] Saviq, I always tend to run with -f so I don't know your pain, but I can imagine [14:16] mterry, yeah, I rarely run with -f [14:16] and I think most of us do [14:17] dednick, looks like Mir ' [14:17] 'innit [14:17] https://bugs.launchpad.net/ubuntu/+source/mir/+bug/1342694/comments/2 [14:17] Ubuntu bug 1342694 in qtdeclarative-opensource-src (Ubuntu) "qmlscene crashed with SIGSEGV in _M_release()" [Medium,New] [14:17] Saviq: yup, that it does [14:22] Saviq, run.sh takes a while on my machine to load up. It didn't used to. does that happen for you? [14:22] mterry, bug #1340086 [14:22] bug 1340086 in telephony-service (Ubuntu) "Importing Ubuntu.Telephony 0.1 causes hang when there's no telephony accounts" [High,Fix released] https://launchpad.net/bugs/1340086 [14:22] mterry, update / restart telephony-service [14:24] ah great [14:24] tsdgeos, is your overview branch pushed anywhere? [14:24] mhr3: https://code.launchpad.net/~aacid/unity8/dash_overview/+merge/226449 [14:24] thx [14:24] i'll try to hack the plugin to do what you need === greyback|lunch is now known as greyback [14:32] Saviq, I'm not thrilled with installing it in nonmirplugins for actual installation, but how do you feel about creating a symlink in builddir/nonmirplugins when using ./run.sh? It's hacky, but it's a development/testing thing anyway [14:32] mterry, could work [14:37] Saviq, OK done, let me know what you think [14:38] what happened to the meeting? [14:38] did we do mumble today after all? [14:39] mterry, hangout! [14:39] dednick, ↑ [14:39] I was there since 20 minutes ago, link must have changed [14:39] oh you're here ;) [14:39] :) [14:40] mterry, sometimes it happens that even though you have the same link, you're on a different hangout ;) [14:43] mhr3: we talked briefly the other day about what it would take to support running scopes using the new APIs against the scopes test tool on Trusty, I need to get a little more information about how much work would actually be involved in that [14:43] my take-away from that conversation was that it's the test tool itself that needs the most backporting work [14:44] because it uses some new Unity 8 components that aren't in 14.04 [14:44] mhall119, right, which ultimately means you'd be backporting u8 itself [14:44] all of it, or just the components that are shared with the test tool? [14:45]