[07:23] * Saviq is peeling today ;D [07:41] Saviq: should we do a landing with the approved stuff we have? [07:41] tsdgeos, yup, was planning to today [07:41] tsdgeos, wanted to see if we can get departments in [07:41] don't think so :/ [07:41] mhr3 is on holiday and he has a few things marked to fix in his MR i think [07:44] oh ok [08:35] Saviq: did you try the touchdown thing on the phone? [08:37] tsdgeos, ah maybe you meant recent apps? [08:37] tsdgeos, I only tried the app scope [08:37] as in the installed / available [08:37] Saviq: no no, regular apps [08:37] hmm looked fine... [08:37] * Saviq tries [08:39] let me try again too [08:39] actually happens on desktop [08:39] click on system settings [08:40] it's the same problem with artshapesize not forcing artshape [08:40] maybe the patch i have for yesterdays two line title vs art [08:40] fixes it too [08:40] * tsdgeos checks [08:40] tsdgeos, I don't have that patch ;) [08:40] i know [08:41] i'm just saying that maybe it fixes it [08:41] ah ok [08:41] tsdgeos, ah now I see what you mean [08:41] tsdgeos, weird [08:42] no, patch doesn't fix it [08:42] which is weird, seems like it should :/ [08:42] tsdgeos, what's worse, this will mean one more UbuntuShape again :| [08:43] ah, it actually does [08:43] i was merging from the wrong place [08:43] Saviq: oh yeah, i know [08:43] but I think Cimi tested that it's not significant? [08:43] testCardBenchmark gave "similar numbers" [08:44] i think he said [08:52] errrr [08:52] can't compile on the phone anymore :/ [08:52] runs out of space [08:55] so we have to x-compile now? [09:00] tsdgeos: yes, which project? [09:00] unity8 [09:00] tsdgeos: yes, should x-compile fine [09:00] tsdgeos: https://wiki.ubuntu.com/SimpleSbuild [09:01] tsdgeos: https://wiki.ubuntu.com/CrossBuilding [09:01] i know it x-compiles fine [09:01] it's just infinite amounts of work more than ./run_on_device.sh [09:01] for a 1 line change [09:01] +1 [09:01] tsdgeos: in that case, you can free up a little disk space on the phone [09:01] makes my developer life hard without any need [09:02] tsdgeos: http://pastebin.ubuntu.com/7119790/ [09:02] oor we could make a mroe developer friendly image [09:02] which tbh are the 99% of people using the phone now [09:03] that's a 2 sided sword [09:03] but whatever if people want me to be slower developing, i'll be slower developing [09:03] well, the solution to this would be to have run_on_device using the x-compile environment really [09:04] but it would still create a deb? [09:04] that's sloooooooooooooooooooooooooow [09:04] no... just compile in the chroot and rsync binaries [09:04] rsync the builddir basically [09:06] probably not even using sbuild... probably click-chroot to set it up, then call schroot make etc and then rsync /path/to/chroot/builddir to the phone... that's how I'd start tackling the issue... [09:13] Saviq: what do you think of https://code.launchpad.net/~aacid/unity8/obeyArtShapeFixedSize/+merge/222294 ? [09:14] tsdgeos, whoa, I was hoping you'll tell me I overcomplicated it and you'd simplify it :D [09:14] ^_^ [09:14] it's just adding an extra if :D [09:14] *just* ;) === pete-woods is now known as pete-woods-away [09:15] tsdgeos, I'll have to focus to read through this, will need food first! ;) [10:05] tsdgeos, Saviq what was the question? [10:05] Cimi, aren't you on holidays? [10:05] Cimi: no question i think [10:05] Saviq, I am [10:05] Cimi, then go away [10:05] :P [10:05] Saviq, but I receive pings on my pebble :P [10:05] Cimi, then log out of IRC [10:06] and yes, we know you've got a pebble :P [10:06] B) [10:06] and we know you eat scorpion chilly [10:06] tsdgeos, you missed that! [10:06] chilli [10:06] http://en.wikipedia.org/wiki/Trinidad_Scorpion_Butch_T_pepper [10:07] "was formerly the most piquant pepper" [10:07] lol [10:07] oh no, i didn't miss that at all :D [10:07] i may die just by being on the same room as that thing [10:07] tsdgeos, I *did* die [10:08] or maybe it was http://en.wikipedia.org/wiki/Trinidad_moruga_scorpion [10:08] I have the pic [10:08] hold on [10:11] Saviq, https://drive.google.com/file/d/0BzO8ZDfBk9UKNTBPNkVWbm5rRlU/edit?usp=sharing [10:12] so it's probably the latter [10:15] Saviq, "You take a bite. It doesn't seem so bad, and then it builds and it builds and it builds. So it is quite nasty" [10:15] Saviq, your description :D [10:15] from wikipedia [10:15] the "oh fuck" started indeed 5 minutes after the first bite [10:15] ahaha [10:17] mzanetti, can you drive https://code.launchpad.net/~mzanetti/unity8/dont-crash-on-invalid-app/+merge/220636 into the ground (or trunk, for that matter)? [10:17] Saviq: you mean landing it? [10:17] mzanetti, I mean getting someone to review and ACK [10:17] I'll take care of the landing [10:18] Saviq: hey, can you review https://code.launchpad.net/~mzanetti/unity8/dont-crash-on-invalid-app/+merge/220636 ? [10:18] :P [10:18] nah... I'll ask greyback, he saw it already anyways [10:18] greyback: small one https://code.launchpad.net/~mzanetti/unity8/dont-crash-on-invalid-app/+merge/220636 [10:18] ;) [10:19] dammit that bug is annoying, I just can't figure it out [10:19] greyback: yesterday I came by some lambdas that looked suspiciously thread unsafe [10:20] but not on appStarted... only on appclosed :/ [10:20] the crash seems to happen on started too [10:21] /food (and scrub first) [10:21] lol [10:22] mzanetti: my only guess is perhaps an app which immediately dies on start. Obvoiusly I've must've made an error in serializing all the mir & upstart events, I just don't yet see it [10:22] greyback: can we get infinite amounts of debug added so that when it happens we have debug logs to pull? [10:22] greyback: do you know how how QObject::connect() behaves in when connecting a lambda whith something from another thread? [10:22] tsdgeos: working on it now [10:22] :) [10:23] greyback: I think it might fail to detect the different thread and makes it a directConnection [10:23] mzanetti: that lambda run in the thread of the signal. No queued connections for lambdas [10:23] yeah... so I think in qtmir the applicationRemoved is dangerous [10:24] because that one in turn emits rowsRemoved etc [10:28] mzanetti: there's no such lambda in unity-mir's AppManager tho [10:29] mzanetti: and yeah the qtmir code is an old version of unity-mir's - it does need updating [10:29] I've a branch with that mostly done somewhere [10:32] paulliu: can you merge https://code.launchpad.net/~paulliu/unity8/logout/+merge/216373 ? [10:34] mzanetti: https://code.launchpad.net/~gerboland/unity-mir/enable-logs/+merge/222300 <- that the right thing to do? [10:37] greyback: looks ok, but would need to test [10:37] doing that now [10:39] thanks [10:50] Saviq, I've implemented most scopes-api/unity-api/scopes-shell changes for https://bugs.launchpad.net/unity-api/+bug/1326292 ; can you assign unity8 to somebody? [10:50] Ubuntu bug 1326292 in Unity 8 "Unity8 should not request preview for the upcoming 'store' category in Apps" [Undecided,New] [10:50] s/most// [10:51] Saviq, also, reviews would be much appreciated ;) [10:52] greyback: nope... installed that package, no unity-mir debug prints :/ [10:53] mzanetti: darn. Any ideas? [10:53] greyback: just checking, but I guess dpkg-buildpackage actually sets the build type [10:53] I totally didn't think it did. Oh well [10:54] greyback: I'm verifying that [10:58] pstolowski, if you want reviews, we need MPs! [10:58] Saviq, ah, indeed :), 1 minute [10:58] mzanetti, greyback, it does - RelWithDebInfo [10:59] yep... [10:59] unless you override dh_autoconfigure of course [10:59] ok... so I guess that's what we want to do [10:59] mzanetti, see what unity8 does [11:00] thanks [11:03] Saviq, done. when implementing it in unity8, I think it'd be best if you leave the hack we currently have to ease the "transition", and remove it once new click scope lands [11:03] pstolowski, +1 [11:03] pstolowski, or well, we can just land them together [11:03] but k [11:04] Saviq, when does the old tabbar in unity8 get removed ? [11:04] Saviq, yeah, sure, I just think that way may be faster [11:04] om26er, you mean the new header? we'll get on it soon [11:04] Saviq, yeah [11:05] om26er, probably along with the dash overview, to not regress in the speed you can access your favourite items with [11:05] Saviq, overview means bottom edge swipe to show all scopes ? [11:05] om26er, yes [11:06] ok [11:06] Saviq, when the screen is turned off the volume keys still work, is that handled by unity ? [11:06] om26er, yes [11:06] there goes a bug report [11:07] om26er, there is one already [11:07] om26er, bug #1291458 [11:07] bug 1291458 in unity8 (Ubuntu) "[platform design] Volume buttons change ringtone volume when suspended" [High,Triaged] https://launchpad.net/bugs/1291458 [11:08] /food for real [11:21] greyback: https://code.launchpad.net/~gerboland/unity-mir/enable-logs/+merge/222300/comments/532564 [11:21] greyback: I just tested this, works [11:21] just make sure to turn spaces into a tab when copy/pasting it [11:22] mzanetti: ack === MacSlow is now known as MacSlow|lunch [11:25] greyback: mzanetti: do either of you guys have an n10? i've heard reports of the n10 being very flaky with the shell-rotation ppa, but i don't have one to verify [11:26] cwayne: can you define "flaky" better? The main issue I know of is the browser scrolling badly (it just jitters until the scroll completes - not something we can fix, is either Mir or browser issue) [11:27] greyback: all I got was "it crashed 3 times in 5 minutes" [11:27] i can try and prod for some logs [11:27] cwayne: running packages from yesterday/today? [11:28] greyback: yesterday, yeah [11:28] greyback: also apparently video playback doesn't work on n10 [11:28] cwayne: if you could get logs, it would help. Contents of /var/crashes especially. [11:29] greyback: I think the problem is still that it gets crashy after the screen has been off [11:29] it works fine at first, then it suspends, we wake it up again and things start crashing more often [11:31] mzanetti, Hi! since the right edge switcher landed, now if you have a single app opened and you swipe the right edge nothing happens. Previously the app zoomed out to show no apps are running. Intentional or regression ? [11:31] cwayne: ouch. I hadn't tried video at all. I would suspect that doesn't work with stock images either... [11:31] mzanetti: interesting [11:31] that behavior can still be seen on dash (with no apps open) [11:31] om26er: will be fixed soon [11:31] om26er: design wasn't really sure yet what should happen in that case [11:32] cwayne: ok I'll play with the N10 and try to repro & fix crashes. [11:32] mzanetti, need me to report a bug ? [11:32] greyback: i've asked for logs to be sent your way as well [11:32] cwayne: I doubt there's anything I can do about video playback though. Will see if I can figure out what's wrong [11:33] om26er: not really... there's that big QtCompositor rework coming up, also turning the dash into an app... that will force us to touch that code anyways [11:33] mzanetti, cool, that means dash in the right edge switcher ? [11:33] cwayne: in the PPA this morning we have app splash screens working. It makes the UI look nicer, could you check it out please? [11:33] om26er: yes [11:33] always wanted that. [11:33] yay! [11:33] greyback: yep, was just updating [11:40] greyback: cool! splash screen seems to work really well so far [11:41] cwayne: let me know if you spot any problems [11:49] tsdgeos: Saviq: FWIW, for the short right edge flick Vesa's prototype didn't have the 2 gu minimum either but we still added it in the end. I think its safe to go ahead and add that check. I had the same feeling as Saviq when I tried it before. [11:49] greyback: will do [11:50] hmm, seeing a weird overlay on one of our scopes when in landscape on n4: http://people.canonical.com/~cwayne/overlaid.png [11:50] mzanetti, TBH 2gu is nothing, I'd say it should be like 1/4th of the screen [11:50] Saviq: no... that's explicitly not wanted by design [11:50] mzanetti, but if Vesa made it like that, then that's probably fine [11:51] Saviq: design wants to have it really short, if we're detecting its really what the user wants, and that's done by the speed [11:51] mzanetti, k, feels a bit jumpy to me [11:51] Saviq: jumpy in what sense? [11:51] too easy to close? [11:51] tsdgeos, it's too easy yeah [11:51] it's what the iphone does [11:52] it's ultra easy to close stuff in their "app carousel" view [11:52] cwayne: can you show me how to install that scope? === MacSlow is now known as MacSlow|lunch [12:00] greyback: push this to the phone and untar it: http://people.canonical.com/~cwayne/scope.tgz [12:00] cwayne: you said its flaky with the shorts app [12:00] cwayne: for me the shorts app doesn't start up at all [12:01] still nothing crashes [12:01] mzanetti: yeah me neither, i meant flaky as in sometimes that caused it to crash [12:01] not sure why the app itself will never launch though [12:01] messaging app & dialer take _ages_ to start up [12:02] greyback: right... dialer doesn't start for me at all either on the tablets. it does work on the phone [12:02] we hide it in the tablet images for the demos anyway [12:03] oh, there it is :) === alan_g is now known as alan_g|lunch [12:03] indeed it just takes _ages_ [12:03] greyback: found an issue with the app splash [12:03] its click through [12:03] mzanetti: ah nice catch [12:04] but I didn't have a single crash today so far [12:05] cwayne: I pushed a fix so videos play from scopes [12:05] greyback: on n10? [12:05] cwayne: well I'm playing "yes men fix the world" in the browser on my n10. But the video is far from smooth. [12:06] there's a frame ordering bug on n10 that is very obvious [12:06] Mir & binary blob GPU driver issue - nothing we can do [12:06] damn [12:07] well at least there's some playback now :) [12:07] always a silver lining ;) [12:13] it's too bad split greeter + shell rotation can't go together, would be real nice to use a promoted image [12:14] cwayne, they can, it just needs time that you don't have ;) [12:14] cwayne, it's gonna be there in two weeks or so [12:14] :) [12:14] greyback: i love this app splashscreen [12:31] tsdgeos, oh I wonder, does the "swipe to close" work with multiple fingers? :D === pete-woods-away is now known as pete-woods [12:35] * greyback thinks its more natural to swipe away from you to close apps [12:36] webos did that [12:36] or still does that [12:36] true [12:37] oh man i loved webos [12:37] i frequently do that by accident on ubuntu .. especially when in the right edge app switcher [12:38] tsdgeos: ok. done. [12:39] tsdgeos: I really don't understand bzr merge. I didn't modify main.cpp at all but I need to solve the conflict on main.cpp === _salem is now known as salem_ === MacSlow|lunch is now known as MacSlow === alan_g|lunch is now known as alan_g [13:03] Saviq: lol, you want to close 3 at the same time? [13:03] tsdgeos, or 8 :D [13:03] dunno if I can do 10 [13:04] Saviq: is this a requirement? [13:04] tsdgeos, no, ;) [13:05] tsdgeos, but I can't imagine why it wouldn't work tbh [13:05] actually i should fix it [13:05] you can drag two [13:05] or more, just tried two [13:05] good ;) [13:05] but i think i have code that assumes just one [13:05] and the second breaks [13:05] and is left lying in the middle of the drag [13:05] so either limit it to 1 [13:05] or make it work [13:09] Saviq: i think i'm going to limit it to one [13:09] for now [13:09] http://nooooooooooooooo.com/ [13:11] * tedg wonders how many words Saviq has bookmarks for [13:11] eofl [13:11] rofl [13:12] end of the floor, laughing [13:12] good I'm not on my balcony [13:12] Heh, yes! [13:12] tedg, bookmarks are so 80's [13:12] tedg, awesome bar FTW [13:13] Saviq, What do you see as the difference? I just see that as searchable bookmarks. [13:13] tedg, I don't *put* stuff in there [13:13] tedg, it just is [13:14] Saviq, Ah, it forgets too much for me to do that. [13:14] yeah, my fallback is the Google machine [13:14] but rarely need it [13:14] I use pinboard.in, it has nice chrome awesomebar integration "p ${search}" [13:15] https://pinboard.in/u:ted.gould [13:15] chrome? maybe that's why it forgets too much :D [13:15] Ifttt integrates it with other stuff [13:38] tsdgeos, remember how the carousel deals with clicked and pressAndHold? [13:38] Saviq: what do you mean? [13:38] tsdgeos, carousel itself emits clicked and pressAndHold signals, with index and delegate.y [13:39] yes [13:39] Saviq: for some reason here [13:39] the dragging area of the second card [13:39] was getting the drag y of the first one [13:39] how to autopilot tests with the split greeter ? just stop unity and start the tests ? [13:39] so that's why it wasn't working [13:39] tsdgeos, how about instead we say that the delegate needs to have a clicked and pressAndHold callables (function / signal), and let it deal with what it needs internally? [13:40] om26er, unlock, and run phablet-test-run -n unity8 [13:40] om26er, you can use the unlock_script (on your host) to unlock programatically (but easier to just unlock yourself) [13:40] Saviq: what's the benefit of that? [13:41] tsdgeos, no knowledge in Carousel of what CardCarousel needs [13:41] tsdgeos, the delegate knows what it needs instead [13:41] Saviq, ok, btw is the silo being tested ? [13:41] om26er, yes, I'm running ap now [13:41] om26er, and will go through testplan just after [13:42] Saviq: can do i guess, seems a big-ish refactory for only a "small" teoretical/architectural imprivement [13:42] Saviq: but sure, makes kind of sense [13:43] tsdgeos, huh? http://paste.ubuntu.com/7601829/ [13:43] tsdgeos, not a huge refactor :) [13:43] tsdgeos, otherwise we'd need to pass the whole delegate, or build knowledge into Carousel [13:44] (sure, tests need fixing) [13:44] tsdgeos, ideally we'd just pass the input to the delegate, but that's kinda impossible now [13:45] why are you passing the result to pressAndHold? [13:46] yeah smaller than what my head was thinking [13:46] tsdgeos, scopes plugin requires it for activation [13:46] tsdgeos, and we need to allow for activation on long press now (scope decides per-category) [13:46] ^_^ [13:46] tsdgeos, so we're getting rid of the if (scope == "click" && category == "local") [13:46] tsdgeos, not like it will ever get used [13:47] Saviq: there's this thing called CarouselDelegate or something [13:47] document the functions there [13:47] BaseCarouselDelegate [13:47] or actually create empty ones [13:47] which we're not using [13:47] we are [13:47] in the tests [13:47] :D [13:51] ;) [13:54] tsdgeos, btw, this could be an addition to QML [13:54] inheritance? [13:55] virtual functions? [13:55] multiple inheritance? [13:55] tsdgeos, sometimes having to create a CardCarouselDelegate just to base it off of a BaseCarouselDelegate is painful [13:55] tsdgeos, so yeah, inheritance [13:55] something like [13:55] multiple inheritance + abstract functions would be cool [13:55] basically [13:55] interfaces [13:55] Loader:BaseCarouselDelegate { foo } [13:55] yeah === alan_g|tea is now known as alan_g [14:45] Did the system-wide theme used by Qt-apps break for anybody with any recent update pulled? [15:22] mzanetti: greyback: dandrader: have you guys seen this on n4 landscape? http://people.canonical.com/~cwayne/overlay.png [15:23] tsdgeos: ^^ [15:24] i didn't even know we could put the n4 in landsacape [15:24] cwayne, no... [15:24] * greyback EOW [15:24] bye all! [15:27] bye greyback [15:27] cwayne: I've seen this on trunk too [15:27] with portrait [15:27] tsdgeos: re landscape, check out the QtCompositor ppa [15:28] or well, there's the rotation anyways... no need for you to check it out atm [15:39] dandrader: ping [15:39] tsdgeos, pong [15:39] dandrader: about the EdgeDragEvaluator [15:40] it doesn't do what i'm doing (which is what designers designed), no? i just "finish" the animation if they moved more than 1gu and the velocity is > 600 [15:40] the EdgeDragEvaluator seems to have a distance that even if you are not moving at 600 it will be completed [15:40] tsdgeos, yes, it's more sophisticated [15:41] tsdgeos, the farther away away you drag, the less velocity is needed for the action to complete [15:41] right [15:41] not sure that's what deisgners want there [15:41] dandrader: is that what we use for launcher? [15:41] tsdgeos, up to a point where no velocity is needed. where actually you would need a negative-direction velocity to avoid auto-completion [15:42] tsdgeos, in short: the smaller the drag, the bigger the velocity must be for the action to be accepted [15:42] right [15:42] i unerstand [15:43] doesn't seem to be what designers wanted [15:43] or not what the prototype does [15:44] dandrader, but if it feels better, it can be proposed. I believe it yields less unwanted actions and is still easy to achieve an intented action [15:44] s/dandrader/tsdgeos :) [15:45] tsdgeos, I actually made this up by playing with android's indicator panel (opening and closing it) :) [15:45] tsdgeos, so it's used in ubuntu's panel, lock screen and stage [15:46] tsdgeos, you known, design is an experimental, always changing, thing [15:47] dandrader: i understand you think it may be better, i'm eow'ing nowish though and next week i'll be busy with QtCS, maybe you can find time to replace the parts of code i have that use draggingArea with that? [15:47] dandrader: we're reaching that point where imho it's better if we "do" something than if we "perfect" something [15:48] tsdgeos, sure. incremental evolution/improvements is always a good approach [15:49] agreed, i'd vote for landing this if the code is not awful and design thinks its ok [15:50] and then we can propose using the EdgeDragEvaluator after RTM [15:51] tsdgeos, not for me to say. I just made a suggestion over there [15:51] haven't actually tried that branch yet [15:52] I mostly wanted people to be aware of this component so that it might get more wide spread use and improvements [15:53] i see [15:57] dandrader: tsdgeos: yeah... I had the same issue for the short right edge flick too [15:57] its a pity that we're not using a component we prepared just for this kind of stuff [15:58] dandrader: mzanetti: some crash files from the n4 running with shell-rotation: https://chinstrap.canonical.com/~joc/n4-crashes/ [15:58] dandrader: I think we should make EdgeDragEvaluator support this use case and then make use of it in the stages code [15:59] cwayne: hmm... I don't think those are related to unity8 crashing. Do you still see many unity8 crashes? [15:59] cwayne: its working quite smoothly here by now. don't think I had a crash today [15:59] well, no visible one at least [16:00] mzanetti: joc has seen a few on the n4, let me get him in here to share how he did it [16:09] dandrader: https://code.launchpad.net/~aacid/unity8/swipe_down_close_no_remove_dash/+merge/221996/comments/532741 [16:12] mzanetti, ok === gatox is now known as gatox_lunch === dandrader is now known as dandrader|afk [16:32] tedg: are you about? [16:32] * tedg checks about me [16:33] xnox, What's up? [16:33] tedg: i have a PPA with upstart+cgroups that is usable on amd64/i386/armhf both for system and user level jobs [16:33] Oooh, sexy. [16:33] xnox, Link/ [16:33] ? [16:34] tedg: there are a couple of tricks (a) one needs to call initctl notify-cgroupsmanager-address (b) one needs to specify at least one "cgroup" stanza in the job to get it cgroupsify [16:34] (e.g. "cgroup cpu") [16:34] tedg: it's in https://code.launchpad.net/~canonical-foundations/+recipe/upstart-daily-nonvirt [16:34] xnox, Do I need to notify in the user session? [16:35] Seems like there's no way it could be started there. [16:35] tedg: yes, you do need to notify in the user session as well. Which i think is a miss-normer. [16:36] tedg: i might add a user-session job "start on startup" which does a sanity check that cgroupmanager is available, and if yes sends notification to user-session init. [16:36] tedg: but yeah, it shouldn't be necessary on user-session at all. I'll chat with jodh about it. [16:37] xnox, Cool, makes sense. If it was a job that'd work too really. Not sure which makes the code simpler. [16:37] xnox, When's the next upstart release? [16:38]