[08:59] o/ [09:09] mzanetti: welcome bac [09:09] k [09:09] mzanetti, \o [09:20] mzanetti, hey there [09:20] mzanetti, happy new year btw [09:21] MacSlow: hey, thanks. happy new year you too [10:28] tsdgeos, I tested your branches, they seem fine [10:28] tsdgeos, which tests you said you wanted to add? [10:29] the one i added to https://code.launchpad.net/~aacid/unity8/departmentListHorizontalScroll/+merge/245580 [10:29] ok [10:29] which failed ^_^ [10:29] tsdgeos, yeah saw that [10:29] it should be running again after my last commit [10:29] let's see how it goes [10:29] i couldn't reproduce it failing here even in a loop [10:30] do uou think is possible to fix the pageheader height instead clipping? [10:30] on the other branch [10:31] not as the code stands now [10:32] rebooting router, back in a min [10:34] Mirv: ping [10:43] tsdgeos: pong [10:45] Mirv: about the qt5.4 issue i was wondering if you could rebuild the stuff with http://paste.ubuntu.com/9686974/ and removing the -reduce-relocations from debian/rules [10:45] since this was one of the things that was causing problems in 5.3 and then our linker got fixed [10:45] but maybe not totally fixed? [10:46] i tried compiling myself but qtdeclarative-opensource-src is failing to dpkg-package for some reason :/ [10:48] grrr stupid https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-vivid/245/console failed [10:49] * tsdgeos hits rebuild [10:51] tsdgeos: right.. so that paste is actually the manual our patch reverting upstream lines since we got the updated binutils. [10:51] so I can comment out the patch. plus comment out the configure option. [10:51] right [10:51] see if it helps [10:51] I'm in a middle of enabling qtbase tests but I can also comment out that.. [10:52] oki === vrruiz_ is now known as rvr [11:04] greyback: did you have time to test https://code.launchpad.net/~aacid/qtmir/create_observer_sooner/+merge/244622 ? [11:05] tsdgeos: hey, no I must do that now [11:05] had slipped my mind [11:05] tsdgeos: 5.4.0+dfsg-4ubuntu1~vivid1~test2 building with http://pastebin.ubuntu.com/9687032/ [11:06] oki [11:09] greyback: oki [11:46] MacSlow: how's https://code.launchpad.net/~macslow/unity8/swipe-dismiss-snap-decisions/+merge/233347 going? [11:46] tsdgeos, working on it to get to work [11:46] :) [11:46] tsdgeos, setting actions is giving me a headache atm [11:47] getting segfaults where I didn't expect them... especially since the real (non-mock) implemenation works flawlessly [11:48] :( [11:52] tsdgeos, but I'm getting there [11:52] two of three segfaults are already eliminated === MacSlow is now known as MacSlow|lunch === alan_g is now known as alan_g|afk [12:15] MacSlow|lunch: https://code.launchpad.net/~pitti/unity8/notify-api-fix/+merge/245643 is for you, right? [12:21] tsdgeos, nah, actually for me [12:21] oh ok [12:21] ACKed === alan_g|afk is now known as alan_g [13:04] tsdgeos: could you add checklist (https://wiki.ubuntu.com/Process/Merges/Checklists/QtMir) to https://code.launchpad.net/~aacid/qtmir/create_observer_sooner/+merge/244622 plz === MacSlow|lunch is now known as MacSlow [13:37] greyback_: sure [13:37] ta [13:58] Is MediaPlayer a QML-object defined unity8 or Qt itself? [13:58] doh... defined by unity8... [14:15] Mirv: is there a simple way to trigger a rebuild of unity8 and qtdeclarative in https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-005/+packages ? [14:15] now that the new qtbase has been built [14:16] Hola [14:16] rsalveti, ping; having problems with rest scopes? [14:19] tsdgeos, Mirv will have to upload bumped versions [14:19] mterry, test failed on jenkins [14:20] tsdgeos, you can't rebuild in PPAs as they prevent two different binaries with the same version [14:20] Cimi, yeah I've been looking at that this morning [14:20] makes sense i guess [14:20] ok! [14:37] Cimi, please leave a vote on https://code.launchpad.net/~mterry/unity8/wizard-passphrase-osk/+merge/244358 apart from top-approving [14:37] Saviq, ok [14:45] greyback_, few more reasons for hash -d ;D → you can type "foo" to `cd ~foo` (where foo is how you named the "shortcut"), but even better - it also works in tab-completion, foo will autocomplete to files from foo etc. [14:45] ;) [14:45] hmmm [14:46] Saviq: can you use that in a command. e.g. to overwrite a file, can you do "cp thisFile foo" [14:46] greyback_, yeah [14:47] greyback_, zsh expands foo [14:47] hmm, that is nice [14:47] greyback_, *might* need to prepend ~ in this case [14:48] no ~ needed for some built-ins like cd apparently [14:48] so ~foo [14:51] at least notifications still work on the device... testing unity8's notifications locally on the desktop always fails [14:54] hey, I can't ./run.sh any more... anyone knows already what's going on? [14:54] i haven't run run for a long time [14:54] directly the binary i want [14:55] Cimi, ok tests fixed, let's see what jenkins says. I also filed the MP for the tutorial-new-screens branch yesterday, it's ready [14:56] facubatista: the problem I had with the today scopes is that it wasn't adding the weather data, even when I was online (not sure if this is the one you're talking about) [14:56] and the remote scopes were also not available, so cwayne1 said that this could be the reason [14:56] mterry, what shall I review in the tutorial refactor?? [14:57] rsalveti, __lucio__ told me you weren't seeing any remote scopes... [14:57] mterry, design too? It looks nice but doesn't work well [14:57] Cimi, doesn't work well? [14:57] facubatista: yeah, worked after a clean flash, will let you know if I get to reproduce this again [14:57] mterry, I kept tapping the orange circle to swipe [14:57] mterry, before I realised I need to swipe from the edge on the ubuntu phone [14:58] mterry, has this been user tested? [14:58] Cimi, I don't know whether they've user tested [14:59] mterry, let's wait for user testing [14:59] Cimi, but as for what to review, I'd say just that my code isn't crazy and that it works. I think design can handle their side [15:00] mterry, ok, but let's be sure we ask for user test [15:00] Cimi, no let's not wait for user testing... Design will review the MP and sign off or not on whether they like it [15:00] rsalveti, ok! just let me know about *any* issue with remote scopes, I should be of help :) [15:01] mterry, I think we should user test, to avoid what happened with the scopes overview [15:01] facubatista: great, thanks [15:01] mterry, so ok, I will approve code wise, but let's make sure we request that before merging [15:01] Cimi, I'm only going to wait for Design to say it's OK. I don't want to second guess their methods [15:01] Saviq: ok, i know what's going on with the crash on https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-vivid/246/consoleFull something is going to slow and it's crashing on a waitForRender [15:02] we should reallly make waitForRender not crash on empty items [15:02] tsdgeos, sounds like an easy thing to do in UnityTestCase [15:02] with the exception that it's going to be tricky to call the "upstream" waitForRendering [15:03] mterry, might go to office tomorrow if my back feels better, I can ask... [15:09] Saviq: nah it's easy too [15:09] i'll add it in and propose the fix upstream, see what they think [15:09] coolz [15:14] tsdgeos: how do you do with the lightdm plugins then? [15:14] tsdgeos: you always type the full LD_LIBRARY_PATH thing? [15:14] don't need them for unity8-dash [15:14] don't run unity8 much ^:^ [15:14] ah [15:15] Saviq: https://code.launchpad.net/~aacid/unity8/waitForRenderingNoCrash/+merge/245747 [15:15] tsdgeos, wanna add semicolons? [15:16] Saviq: to where? [15:16] tsdgeos, to end of lines ;) [15:16] ah [15:16] i just copied from the qt code [15:16] but sure will add [15:16] it's not like we're consistent about them :/ [15:17] added [15:17] tx [15:22] Saviq: so, does ./run.sh work for you? [15:22] for me it runs the binary, but it doesn't show up, and it can't be killed any more except with -9 [15:23] mzanetti, same here yeah [15:27] mzanetti, ah! [15:28] pkill -SIGCONT unity8 [15:28] hmm wonder why upstart isn't doing it [15:29] Saviq, initctl works for me, but takes ages... [15:29] ah ok [15:29] mterry: hey, do we still need all that lightdm LD path fun? I wonder if this couldn't be dropped at some point [15:30] mzanetti, there's a branch from dandrader dealing with that [15:30] ah, great [15:31] Saviq, mzanetti: except it doesn't drop the LD stuff, just consolidates it a bit. I think we still want the LD stuff [15:31] mzanetti, that's unless you want to type your password in every time you start unity8 locally ;) [15:32] fair enough [15:36] mzanetti, so what happens is that we emit SIGSTOP twice for some reason [15:36] o_O [15:37] greyback_, ideas ↑? [15:37] greyback_: didn't we change something there before hoidays? [15:37] yep [15:37] we messed up [15:37] yep [15:37] ugh [15:38] yep [15:38] lol [15:38] it's there in main as well as AppMan [15:38] I guess we didn't update the fake AppMan, only the real one? [15:38] who's prepping an MP? [15:38] and why did we skip standup? :D [15:38] we didn't [15:38] just you did [15:39] :P [15:39] :P [15:39] qtmir emits SIGSTOP only if UNITY_MIR_EMITS_SIGSTOP==1 [15:39] I suspect that should not be set in the run script maybe? /doesn't remember [15:40] greyback_, yeah, but we have another raise in main() as well [15:40] Saviq: yeah we needed that for when a mock appman was used instead of real [15:41] greyback_, yeah, but mock appman has raise too now [15:41] oh [15:41] then main can drop it [15:41] yes [15:42] I just wonder how does it work outside of run.sh... [15:42] how do AP tests work?? [15:42] indeed [15:43] ah [15:43] they're ran on Mir [15:43] and the main() one only acted on !Mir [15:43] and we're not running the AP tests on x86 for a while now [15:43] not all AP tests run on Mir tho [15:44] they do [15:44] oh yeah, ignore me [15:44] greyback_, you MP or do I? [15:44] some just use mock appman [15:44] yeah, which is why we needed the raise in mock appman too [15:44] Saviq: I'm trying ot focus on a critical bug right now, can do after [15:45] greyback_, I'll do, then [15:45] ok thanks [15:45] greyback_, let me know if you need help with the wakelock [15:46] Saviq: wakelock the easy bit, hard bit making sure it held only when needed [15:47] greyback_, bird's eye view here ;) [15:47] a drone more comes to mind... [15:49] Cimi: https://code.launchpad.net/~aacid/unity8/departmentListHorizontalScroll/+merge/245580 passes now [15:50] Ugh, my krillin is Out Of Juice after the holidays [15:53] mterry, it lasted a whole two weeks of dogfooding? our battery usage improved a lot! ;P [15:53] easy cleanup https://code.launchpad.net/~aacid/unity8/cleanupTypeString/+merge/245756 [15:53] Saviq, :) [15:54] Saviq: need to land stuff! https://code.launchpad.net/~unity-team/unity8/trunk/+activereviews is growing :D [15:54] tsdgeos, yup, was just restarting failed ci jobs earlier today [15:54] tsdgeos, want a green on every MP [15:57] oki [15:57] we still have unstability [15:57] mzanetti, https://code.launchpad.net/~saviq/unity8/no-sigstop-main/+merge/245758 [15:57] tsdgeos, yeah, I saw a few weird Mir crashes today, trying to see if I can reproduce locally [15:58] but stuff's a bit difficult as we're 160 packages behind vivid in latest devel-proposed image :/ [15:58] thostr_: does the Dash cache GPS data for scopes? I noticed while traveling over the holidays that my scopes were getting old GPS data from 30 miles away, where I was over an hour ago [15:59] it was happening on all scopes that use location data, so I don't think it's the scopes doing it [15:59] (I would expect at least one of them to not do it if that were the case) [15:59] even pulling down to refresh didn't change their location-based results [15:59] Saviq: wow 160 :S [16:00] tsdgeos, yeah [16:00] mhall119: scopes themselves cannot access location data directly but get that passed on by the scopes FW, that's why all scopes have same location [16:00] mhall119, I imagine the location service does the caching [16:00] mhall119: so, question is rather why the location was not updated in either the system or in scopes machinery [16:01] pete-woods: ^ when do we trigger/request a new location? on every query? [16:02] thostr_: the sensor status app was showing updated GPS coordinates [16:08] thostr_: when any scope that has requested location is active, we start a GPS session [16:09] while a session is active the backend is always tracking the current location [16:10] however last time I checked, it took quite some time before a real GPS location was received [16:10] so we most often fall back to geo IP [16:11] each query dispatch we retrieve the current location from the GPS session [16:11] so in theory it should be current [16:35] tsdgeos: so is qt using the disk writes somehow for animations/event handling ? [16:35] something i [16:35] s [16:35] wrt your bug....it's just weird [16:35] i didn't investigate [16:35] odd...means there should really be a reserve [16:36] kgunn, well, the problem with reserve is that your shell is running as you, so if the shell can write, so can you [16:37] kgunn, we'd have to make sure we're writing outside of /home for whatever we need to be writing [16:37] and not write at all what we don't actually need to be writing [16:37] yep [16:37] don't we cache in ~? [16:37] and well, write logs and everything [16:37] mzanetti: qml cache? [16:37] tsdgeos: I think we have at least an image cache for unity + dash in there [16:37] mzanetti: not being able to write logs should not make everything not work :D [16:37] and yeah, prolly qml caching too [16:37] same for the cache [16:38] yeah, anything we *really* need write access to should be outside of /home partition [16:38] tsdgeos: i agree about logs...but i could see cache being an issue [16:38] and any caches, logging should just fail [16:38] sure, the question is if we actually expire the cache [16:38] otherwise at some point it'll just stop working, no matter where [16:38] mzanetti, whether we do or not doesn't matter (for this problem) [16:39] mzanetti, the cache can stop working [16:39] mzanetti, the shell can not, even if cache is not writable [16:39] sure [16:39] it's a cache after all [16:39] mmm yeah, you're right...cache should just fail, i see that [16:39] but I would go as far as saying the cache being full with old stuff and not cache new stuff any more is a bug too :) [16:39] that's not going far [16:39] ...and could be a bug someone is using cache wrongly [16:40] mzanetti, in any case, as far as the QML cache goes, it's cleared on upgrade [16:40] the network cache is max 50MB right now, so old things are expires [16:40] d [16:40] ok [16:40] (and tsdgeos was on vivid, where the QML cache doesn't exist yet IIRC) [16:41] really? [16:41] so anyway, one problem is why it filled up [16:41] * Saviq checks [16:41] +1. really? [16:41] :) [16:41] yeah, the other is it should still work somewhat [16:42] work, just shitty :) [16:42] Saviq: well it filled up because i copied lots of stuff [16:42] that's totally my fault [16:42] hehe [16:42] * kgunn wishes more users were like tsdgeos [16:43] tsdgeos, sure, but there's other stuff we're writing to $HOME [16:43] logs are rotated, but only on session start, if your session lasts for a year or something [16:43] this seems like a bug that's to be fixed thoughout the system actually [16:43] sure [16:44] kgunn, yeah, no QML compilation in vivid [16:44] noted [16:44] glad to hear, explains poor perf i saw when testing too === alan_g is now known as alan_g|EOD