=== duflu_ is now known as duflu === Malsasa_ is now known as Malsasa === Malsasa_ is now known as Malsasa [09:20] bah [09:20] unity8-desktop doesn't start anymore after today upgrade :/ [09:20] seb128, saw that, too :| [09:20] any idea what changed? [09:21] we really need to make sure that image get tested before landings [09:22] oh, unity8 segfaults [09:25] shrug [09:25] segfault in libmirclient8driver.so [09:27] seb128: there was some discussion about too many threads being spawned last night [09:27] maybe that's related? Let me try to find the bug [09:27] larsu, no, it's not [09:27] that issue is happening on the phone for over a week [09:27] but thanks for the hint [09:27] ah okay [09:33] seb128, hmm it actually worked fine here [09:33] Saviq, what worked fine? [09:33] seb128, desktop unity8 [09:34] worked or works? [09:35] seb128, works, like 2 minutes ago [09:36] seb128, had to come back to do IRC [09:36] :-( [09:36] why is that not working for me [09:37] seb128, that's fully updated vivid, in case that matters [09:37] it segaults in libmirclient8driver.so [09:37] same here [09:37] ouch [09:37] the previous function is libEGL.so [09:37] is it supposed to use EGL on amd64? [09:37] seb128, hmm no [09:38] * seb128 tries to power down the box and do a clean boot [09:39] seb128, http://pastebin.ubuntu.com/8800650/ [09:39] seb128, how's that look for you? [09:41] Saviq, similar versions [09:42] bah, after reboot the segfault is in libaalmediaplayer [09:42] in setAudioRole [09:43] unity8-dash abort in libqpa-ubuntumirclient.so [09:47] bah, that time it's unity8 abort in libmirserver [10:14] bah [10:14] Saviq, it works again after downgrading unity8{,private,common} binaries [10:14] Saviq, what unity8 version do you have? [10:14] https://launchpad.net/ubuntu/+source/unity8/8.01+15.04.20141030-0ubuntu1 ? [10:14] seb128, yeah, that [10:15] :-( [10:20] Saviq, what info would be useful in a bug report? [10:22] seb128, apport-cli the .crash, not sure what else... although the fact that it crashes in different spots for you is reaaal weird :/ [10:23] Saviq, on #ubuntu-mir alan wrote that " That is weird. It happens while allocating the memory in which to construct an object. (Which suggests the stack is already screwed.)" [10:24] Saviq, I seem to have consistently the bt on http://paste.ubuntu.com/8800848/ now [10:25] greyback, can you have a look at seb128's crash ↑ [10:25] greyback, I'm not sure we're even getting involved there [10:27] Saviq: that's a bizarre place to crash, would agree with alan_g. Sure, can have a look [10:27] seb128: this is on vivid? [10:27] greyback, yes [10:27] ok [10:27] * greyback kicks his mir box into life [10:28] greyback, it started today with the recent dist-upgrade, goes away when downgrading unity8{,private,common} binaries [10:28] so could be something in http://launchpadlibrarian.net/188743245/unity8_8.00%2B15.04.20141030-0ubuntu1_8.01%2B15.04.20141030-0ubuntu1.diff.gz [10:28] greyback: let me know if/when I can help. (But it will take a while to get vivid set up to play). [10:29] alan_g: thanks! === _salem is now known as salem_ [11:28] Holas [11:40] greyback: that might be a bug if you have to kick mir to make it work ;) [11:42] dandrader, it's the design team that changes status of ubuntu-ux after they've confirmed the fix [11:42] Saviq, oh, sorry [11:42] dandrader, is fine [11:43] Saviq, do you have the bug number handy. I've just deleted all my mails about it [11:44] Saviq, found it in the trash bin [11:44] Saviq, oh, cannot change it back to "Fix Commited"... [11:44] it's all greyed out [11:58] dandrader, yeah, leave it be [12:17] mzanetti, Hi! on image 140 on krillin, I don't see the unity8-dash app, if I close all apps I just see black screen with top panel [12:17] this happened when I picked my phone after a night' sleep. [12:18] om26er: ~/.cache/upstart/unity8-dash.log [12:18] mzanetti, only: [12:18] invalidateScopeResults: no such scope ' "smart-scopes" ' [12:18] I don't see anything else in there. [12:19] Saviq: are we aware of this yet? ^ [12:19] I have heard from people about this, but never happened to me so far [12:19] om26er: can you reproduce? [12:20] om26er, mzanetti, looks like the dash was respawning too fast and the job got stopped [12:20] om26er, what does `initctl status unity8-dash` say? [12:21] Saviq, the process is running [12:21] unity8-dash start/running, process 2482 [12:21] om26er, what PID is unity8? [12:21] Saviq, 2299 [12:22] hmm, so qtmir seems to have lost the connection somehow... [12:22] om26er, `ps aux | grep 2492` ? [12:23] Saviq, phablet 31673 0.0 0.0 4836 660 pts/31 S+ 17:22 0:00 grep --color=auto 2492 [12:23] no process with that pid. [12:23] om26er, typo, 2482 [12:23] Saviq, phablet 2482 0.1 11.2 510188 110440 ? Ssl 12:51 0:21 unity8-dash --desktop_file_hint=/usr/share/applications/unity8-dash.desktop [12:24] om26er, ok, it seems to be running just fine, but unity8 doesn't know, `restart unity8-dash` should bring it back [12:24] om26er, and if you can find steps to reproduce, that would be best [12:25] yeah, that would, I was just keeping it in that state. Will see if I can find steps to reproduce that. [12:27] om26er: could you plz pastebin the ~/.cache/upstart/unity8.log file, it may have a clue [12:29] greyback, http://paste.ubuntu.com/8802368/ [12:31] lots of "qtmir.applications: Dash doesn't seem to be running... Ignoring." messages, so it appears qtmir did not get an incoming connection from unity8-dash [12:33] unity8: pthread_mutex_lock.c:80: __pthread_mutex_lock: Assertion `mutex->__data.__owner == 0' failed. [12:33] line 802 - so unity8 crashed [12:34] om26er: ^ [12:35] greyback, there is a unity8 crash file if you want [12:35] so unity8 respawned, and dash also seems to be respawned. Are there 2 instances of dash running by and chance [12:36] om26er: please share it, in case it helps [12:37] greyback, I am not sure now, I have restarted unity8-dash, so there is only one pid now. [12:37] om26er: ok [12:40] Saviq, greyback, so 3 people confirmed that unity8-desktop stop starting issue, do you need more debug info? [12:41] seb128: I'm installing vivid now, will be able to repro it soon [12:41] greyback, thanks [12:42] let me know if you hit the issue or not [12:42] Saviq doesn't have the problem for some reason [12:42] so it might not be impacting everyone [12:43] seb128: ok, I'll let you know [12:43] thanks === dandrader is now known as dandrader|afk === MacSlow is now known as MacSlow|lunch === alan_g is now known as alan_g|lunch === dandrader|afk is now known as dandrader [13:34] seb128: ok vivid installed, unity8-desktop starting fine for me too :( [13:34] :-/ [13:43] dandrader, poke -- bug 1388359 looks like something I might be related too, have you started looking at it yet? If not, I can help [13:43] bug 1388359 in unity8 (Ubuntu) "User metrics can no longer be changed by double tap " [Critical,In progress] https://launchpad.net/bugs/1388359 [13:44] mterry, I'm working on it and already know what's happening. It's a regression caused by revision 1385 (which is why I guess Saviq assigned it to me) [13:45] mterry, the TouchGate I added broke the double click propagation between the MouseArea in Greeter.qml and the one in Infographic.qml [13:46] dandrader, ah OK -- looks like we are missing a qmluitest for that :) [13:46] mterry, writing one in tst_SingleGreeter at this very moment :D [13:47] mterry, I didn't know about this feature of composite event propagation between mouseareas [13:47] dandrader, well awesome, glad you got it sorted :) [13:48] mterry, will ask you for a review once I'm done [13:48] k [13:48] dpkg guru here? I think dpkg maintains that 0.4.5 > 0.4.5-0~280~ubuntu14.10.1 - that true? [13:48] mterry: ^ [13:48] dpkg --compare-versions 0.4.5 gt 0.4.5-0~280~ubuntu14.10.1 && echo "y" - prints y [13:49] greyback, right because of the -0. If it were -1 it wouldn't [13:49] greyback, because ~ means "less than the previous thing" [13:49] so you are doing "less than 0" [13:50] mterry: ok, I didn't know that. Thanks [13:50] greyback, you can do +280 to do "more than" [13:50] greyback, is 280 a revision number? that's typically how you would append a revision with + [13:50] greyback, or even +bzr280 for clarity [13:51] mterry: was using the default LP recipe for a branch, think it needs tweaking after all [13:52] greyback, assuming 280 is a revision, I'd do {upstream}+bzr{revno}-0ubuntu0ppa{packaging:revno} [13:52] if there is a separate packaging branch. But I bet there isn't, eh? [13:52] so just {upstream}+bzr{revno}-0ubuntu0 [13:52] mterry: nope [13:52] ok, thanks! [13:52] Not sure I got the upstream version variable right. Might be {upstreamversion} or something [13:55] {debupstream} === MacSlow|lunch is now known as MacSlow [14:06] mzanetti, it was your birthday! Belated Happy Birthday! [14:06] mterry: heh, thanks :) === salem_ is now known as _salem === alan_g|lunch is now known as alan_g [14:24] Saviq, greyback, so, you guys can't reproduce the issue, but willcooke didrocks and I can, what's next? [14:24] seb128: what are we doing that's inherantly different. I did a dist-upgrade to vivid, did you? [14:25] greyback, I did a fresh install [14:25] from the desktop-next iso [14:25] greyback, I did dist-upgrade, willcooke did a fresh install from the desktop-next iso [14:25] didrocks did an install of vidid a week ago and updated unity8 only on it [14:25] which was enough [14:25] do you guys have special unity8-hackers builds? [14:25] nope [14:25] didn't do anything fancy [14:26] could you pastebin me a "dpkg -l" and I'll try identify differences [14:27] greyback, http://cdimage.ubuntu.com/ubuntu-desktop-next/daily-live/current/vivid-desktop-amd64.manifest [14:27] greyback, willcooke get it with a fresh install so you can use the manifest [14:28] greyback, btw, downgrading unity8 to the utopic version fixes the issue for me [14:28] didrocks confirmed that upgrading to the version published on friday trigger the bug for him as well [14:28] I wonder if that could be depending of the config, like I think we tested on laptop [14:29] Ahhh... Europe and US are in sync wrt daylight savings again... [14:29] mterry, yeah, for over a week ;-) [14:29] that happened while we were in Washington [14:29] seb128, in Europe, yeah. But US just changed ours yesterday [14:29] mterry: heh, that actually came in handy flying back home from the US last week :) [14:29] seb128, my meetings have been off by an hour all week [14:30] * mterry wants a world clock, where it's the same time everywhere all the time [14:30] mterry, oh ok, I though u.s had it a easlier oct, like before we came to Washington [14:31] mterry: can you reapprove please? https://code.launchpad.net/~mzanetti/unity8/rework-quicklist-visuals/+merge/238149 [14:31] mterry: http://www.timeanddate.com/time/internettime.html [14:32] mzanetti, stupid merges! [14:33] mzanetti, :) I don't need metric beats, I just need everyone to live and breath UTC [14:38] mzanetti, code looks the same and fine, just testing a bit to avoid rubber-stamp syndrome [14:38] yep [14:42] seb128: no obvious dpkg difference, will wipe the box & install the daily live, will take 1-2 hours (it's a slow box) [14:42] greyback, thanks [14:42] greyback, http://cdimage.ubuntu.com/ubuntu-desktop-next/daily-live/ is what you want [14:43] e.g desktop-next [14:44] mzanetti, whenever someone says that base 12 is worse than base 10 I want to slap in the face ;p [14:45] :D [14:46] * Saviq wants 12 fingers [14:47] Saviq: oh you're obviously using emacs ;) [14:47] davmor2, nope! [14:51] mterry: no, those tests are failing everywhere, not mzanetti's branch fault [14:51] or the other way around [14:52] tsdgeos: you talking about this? https://code.launchpad.net/~mterry/unity8/dialer-app-lock/+merge/240293/comments/591130 [14:52] mzanetti: yes [14:52] oh... how did we commit non-compilable tests... [14:53] stuff happens [14:53] and then the fix is waiting for you guys to discuss if the fix is good [14:53] or we prefer to paint the shed of a different color [14:53] https://code.launchpad.net/~aacid/unity8/multimediaMocks/+merge/240248 [14:54] you preferred it purple as far as i remember [14:55] :D [14:55] not really === _salem is now known as salem_ === Pici is now known as to === to is now known as Pici === alan_g is now known as alan_g|tea [15:26] mterry, https://code.launchpad.net/~dandrader/unity8/doubleTapUserMetrics-lp1388359/+merge/240453 === alan_g|tea is now known as alan_g === salem_ is now known as _salem === _salem is now known as salem_ === salem_ is now known as _salem === dandrader is now known as dandrader|lunch [15:42] mzanetti: regarding anti-aliasing, my first attempt was going to be copying how QImage implements antialiasing - it has a special scenegraph node (QSGDefaultImageNode) with a setAntialiasing method on it. We could maybe (ab)use some of its code [15:43] greyback: tbh I don't have any clue about this yet... was gonna try to do what duflu proposed in the bug report [15:43] mzanetti: QSGDefaultImageNode::updateGeometry() is doing the actual heavy lifting [15:43] mzanetti: ah, I didn't see his suggestion. [15:44] i'll take https://code.launchpad.net/~dandrader/unity8/doubleTapUserMetrics-lp1388359/+merge/240453 [15:46] mzanetti: ah he's not wrong, we're indeed using Linear filtering, my bad [15:46] greyback: yeah, I removed that condition on "smooth" just to see what happens [15:46] still compiling atm [15:46] mzanetti: mirbuffersgtexture.cpp, setFiltering() <- change that [15:47] to Nearest instead of Linear [15:47] greyback: that's my diff so far: http://paste.ubuntu.com/8804543/ [15:47] lets see what it does [15:48] mzanetti: looks reasonable [15:48] greyback: obviously if that does the trick, I need to see how it impacts rendering, and maybe make it conditional when the spread is open or whatnot [15:49] mzanetti: yep, good to check, but I don't expect a huge perf impact === _salem is now known as salem_ [16:37] seb128: ok fresh install has me reproducing the problem, can now dig [16:38] greyback, excellent [16:39] greyback, I'm curious how you debug it ... so if you can share some wisdom on what you did, once you are done debugging, it would be great [16:39] greyback, thanks for looking at it!m [16:39] seb128: sure. Once I've an idea what's wrong, I'll share my process [16:40] greyback, thx ;-) === dandrader|lunch is now known as dandrader [17:06] mterry, are you reviewing that guy? https://code.launchpad.net/~dandrader/unity8/doubleTapUserMetrics-lp1388359/+merge/240453 [17:06] has been promoted to a "top blocker" [17:06] dandrader, yes, building on my mako [17:06] dandrader, / waiting for jenkins to do it for me [17:06] mterry, ok, thanks! [17:07] mterry, oh, just saw that tsdgeos claimed the review [17:07] an hour ago [17:07] oh shoot I forgot to press that button, never remember [17:07] well whatever, I'll still test it out, already building :) [17:07] mterry, maybe you could review the greeter part and tsdgeos the touchDispatcher part [17:08] greeter part is trivial refactoring, looks good :) [17:21] mterry: if you want to do it, i'm eod'ing now [17:21] tsdgeos, ok [17:21] otherwise i'll do tomorrow morning first thing [17:22] you choose :) [17:22] * tsdgeos waves [17:22] tsdgeos, see ya! === alan_g is now known as alan_g|EOD === salem_ is now known as _salem