[05:12] mzanetti: you're making me into a framebuffer driver? :) [05:14] I for one will be the first one to use mirvfb [08:21] Mirv: haha [08:36] Saviq: so cleanup is gone, only clean-to-trunk as of now, right? [09:01] meh, the autopilot tests are still too unstable :/ [09:08] Saviq: hey, are you going to implement Ricardo's fix or do you expect a fix in qmenumodel? === davidcalle_ is now known as davidcalle [09:22] tsdgeos, yeah [09:22] tsdgeos, saw loads of failures on the isCurrent and on the index [09:22] tsdgeos, will they be fixed with new tabs? [09:23] has nothing to do with tabs [09:23] MacSlow: ping [09:24] Cimi: i have no idea why they showed up now since nothing changed in there [09:24] tsdgeos, so they might be races [09:24] yes [09:24] same as for unlocking the greeter [09:25] if you look at the video failure [09:25] tsdgeos, what's up [09:25] it tries to unlock the greeter what seems to be a few milliseconds before it's there [09:25] MacSlow: you gave me a unity-notifications thing to review yesterday that says "better with this MR", but that MR is not yet approved, what do you want me to do [09:25] ? [09:26] tsdgeos, that would be nice *very.broad.grin* [09:27] confused :D [09:28] what me to review+approve? or wait? [09:28] tsdgeos, I asked you for this here lp:~macslow/unity-notifications/multiple-snap-decision-example right? [09:28] yes [09:29] tsdgeos, mzanetti already did look over lp:~macslow/unity8/snap-decisions-states and commented (all issues fixed by now) [09:29] MacSlow: nope. on the list [09:30] mzanetti, well sure... I meant it's not totally fresh (like nobody took a look at it yet) [09:30] tsdgeos, then wait I guess [09:30] mzanetti, ^ ok? [09:30] oh... I read it as a question if I re-reviewed after the fixes already [09:30] so yeah. will try to get to it before your lunchtime [09:31] ok [09:32] thx [09:42] tsdgeos, yes, no cleanup [09:43] didrocks, we need to chat whether we can do it lower level easily [09:44] Saviq: as it's a blocker, can you get back to us quite quickly, please? [09:44] didrocks, definitely [09:44] thanks :) [09:46] larsu, you know things about qmenumodel, don't you? [09:46] Saviq: not that much anymore, dednick pretty much took over [09:47] larsu, yeah, but it's actually maybe more of a dbus question - https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1293478/comments/5 [09:47] Ubuntu bug 1293478 in unity8 (Ubuntu) "Volume buttons no longer function correctly since #238 mako/flo/manta" [High,Confirmed] [09:47] larsu, Qt 5.2 decided that "1" is a double, not an int in QML [09:48] Saviq: so what's in for me, anything you want me to look in the clean-to-trunk branch or go back to having a look at the split branch by mterry? [09:49] larsu, we can do what Ricardo wrote there and pray that this'll work (i.e. JS respecting the type), or actually handle this lower level (now that I think of it - maybe it should simply be the sound indicator that accepts doubles?) [09:50] Saviq: on other note [09:50] Saviq: I don't think we should get the transition to new scope [09:50] before getting a promotable image [09:51] didrocks, that's ok, I expected this [09:51] Saviq: can you punt them out? (we still want the other fixes for the indicator thingy) [09:51] didrocks, I just wanted to get a silo, can back it out (or force unlock or something) [09:51] and maybe the workaround in unity8 if the sdk team don't think it shuld be lower [09:51] sure [09:51] Saviq: but then next time qt decides that -1 is an int again we break again? [09:51] that's fine if you reconcile the conflicting silo then [09:52] larsu, well, it should just accept both? [09:52] didrocks, yeah, I'll take care of rebasing / merging and such [09:53] Saviq: action groups don't work like that. Parameters are of a fixed type, so that implementing an action doesn't involve a lot of type checking [09:53] Saviq: we could amend qmenumodel that it knows about the type and does the cast for us [09:54] larsu, right, so that was my other idea, do you know how much work that'd be? I remember talking to dednick about this before and it seemed like a significant amount? [09:54] Saviq: I'd trust him on that, he did most of the action work in qmenumodel [09:54] Saviq: is he out of office? [09:55] larsu, should be around afaict [09:55] just late [09:55] ok, looks like we'll get the workaround in for now [09:56] didrocks, ↑ [09:57] tsdgeos, so, only thing would be test stability, we did go green at some point with the cleanup branch, but that doesn't seem to happen any more :/ [09:57] Saviq: ok, let me see what the clean-to-trunk results are [09:58] tsdgeos, one weird thing I noticed was that "Desktop Nexus 10" tests were running a small geometry window in a few cases [09:58] yeah [09:58] now that you mention [09:58] didn't realize but those videos for N10 look like regular [09:58] tsdgeos, https://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/3533/? seem like actual failures, although no idea how new scopes could affect it [09:59] tsdgeos, and well... qmltests locking up... [09:59] tsdgeos, the 90 min limit, I mean [09:59] * Saviq goes into meeting [09:59] Saviq: we still don't know if they are locking up or just initial sync took ages [10:00] and that's why they go to 90min [10:02] Saviq: great! [10:02] Saviq: so… [10:02] bad news [10:02] blame davmor2 [10:02] there is a crash of unity8 if you search in the music scope [10:02] he's starting bisecting [10:03] didrocks: it's my job to break stuff, don't blame me when I do :P [10:05] tsdgeos, right, mzanetti should be able to increase that, but maybe we can find some timestamps in the log (probably not) [10:05] * mzanetti reads backlog [10:05] Saviq: adding timestamps is trivial [10:05] meh... phone call.. will take a bit [10:05] it's just a plugin [10:06] sil2100, elopio, KUDOS [10:06] we have it in kde jenkins, we just have to convince the right people to install it :D [10:06] Saviq: it crashes unity8 on music and home scope, we are assuming because the home scope is also showing music [10:06] tsdgeos, we tried that already, didn't we ;) [10:07] Saviq: sure, but being repetitive sometimes has some success :D [10:07] davmor2, didrocks, sounds like album art provider, then [10:07] tsdgeos, file a bug with ubuntu-ci-services-itself [10:07] Saviq: right [10:08] Saviq: could be [10:08] Saviq: good news, testShell just hanged in here in the same place it hangs on the CI machines [10:10] tsdgeos, oh good [10:10] davmor2: I don't get a crash though [10:10] but as popey and you are getting it, can be the format the UK providers are sending unity8 a bitmap [10:11] Saviq: FYI, popey didn't reproduce it on latest promoted image, maybe you were safe-guarded against that [10:11] didrocks: how much music do you have on the device I think popey and I have over 4GB [10:12] davmor2: ah, I only have the online providers [10:13] 1.3G /home/phablet/Music [10:13] popey: so just me with over 4 GB then :) [10:13] didrocks, https://bugs.launchpad.net/unity8/+bug/1268172 btw [10:13] Ubuntu bug 1268172 in Unity 8 "Music preview crash" [Undecided,Incomplete] [10:14] Saviq: I think you'll need davmor2's crash file? [10:14] and then https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1262711 [10:14] Ubuntu bug 1262711 in unity8 (Ubuntu) "Unity crashes with lots of music displayed in expanded music category in home scope" [Undecided,Incomplete] [10:14] didrocks, I can try and reproduce locally after adding some music to my device [10:15] didrocks, we had those before, never managed to get rid of them, was hoping 5.2 would help - but maybe it did not [10:15] Saviq: excuse for running your torrent! :p [10:15] or it's a different one [10:15] Saviq: seems to be a different one or popey has different music on the two devices [10:15] i also have had a long standing crash when just expanding the music scope [10:16] popey, yeah, look up ↑↑ [10:16] popey: can you try to copy your music from latest proposed to promoted? [10:16] biab [10:16] Saviq: and now it's even crashing :-S [10:16] https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1262711 [10:16] Ubuntu bug 1262711 in unity8 (Ubuntu) "Unity crashes with lots of music displayed in expanded music category in home scope" [Undecided,Incomplete] [10:16] ah yes [10:16] popey: if it's a dedicated music crashing it, I think unity8 is trying to teach you good music tastes :p [10:16] didrocks: i have same music on all devices [10:17] ok, so Saviq, a safeguard seems to have been removed during the transition (or other formats…) [10:20] didrocks, yeah, we know it to be pretty random [10:21] re [10:21] Saviq: so, should I increase the timout for jenkins jobs? [10:21] didrocks, Saviq, popey: okay so with 238(first image after popey's working one) search is working on home and music with no music on the actual device, I'm transferring music over now so I'll retry once that is done [10:22] mzanetti, no, tsdgeos reproduced the failure locally [10:22] ack [10:22] Saviq: I can trigger the crash as well now, tried the same configuration on latest promoted image and it wasn't the case [10:22] davmor2: yeah, seems you need to transform the same music [10:23] Saviq: i was suggested that in the future we run our processes thorugh annotate-output if we want timestamps [10:24] didrocks, so Qt 5.0 vs 5.2? [10:24] Saviq: that or (hopefully) one of the 2 unity8 landings [10:24] which will make it way easier to corner [10:25] wouldn't count on it :/ [10:27] Saviq: I have a regular string search failing here [10:27] Saviq: I'm trying to pop out music slowly [10:27] to see if I can track down to an album/music [10:27] didrocks, thanks [10:27] yw ;) [10:31] Saviq: That's why I've started on the 5.2 image, next I'll move to the first of the unity landings [10:31] davmor2, thanks [10:35] Saviq: hum… can't get a nice reproducer, I will suggest that we start from the crash file rather [10:36] sometimes it crashes with the same string, sometimes not [10:43] didrocks, I saw recently that the 30s timeout wasn't enough... :/ [10:43] didrocks, at times, at least [10:43] Saviq: do you want me to attach the CoreDump? There are 37 threads… [10:43] Saviq: (gdb) bt full [10:43] #-1 0xaf1004f8 in ?? () [10:43] No symbol table info available. [10:43] warning: Unable to restore previously selected frame. [10:43] doesn't seems yummy :/ [10:43] mzanetti, answered your question in the MR-comment [10:47] Saviq: so i have a fix for that lockup i think, shall i commit to clean-to-trunk? [10:47] tsdgeos, yes please [10:48] didrocks, try increasing the kill timeout to 60 in /usr/share/upstart/session/unity8.conf [10:48] didrocks, maybe the crash is truncated again [10:48] pushed [10:49] Saviq, didrocks, popey: so 238 has the issue on home but I believe that is the expanding on the music section. On the music scope it is finding music no issues [10:49] didrocks: do you happen to know what image had the first landing on unity8 after 238? [10:50] Saviq: hum, can be [10:50] davmor2: hum, why do you need to upgade more, if 238 has the same issue? [10:51] as you have the crash on home [10:51] didrocks: 238 has the issue on home which is different it is expanding the music section which is the known bug, musci search is working fine on the music lens [10:52] davmor2: are you sure you don't have the issue at all on 238? [10:52] it's really random for me [10:52] didrocks: ah no I do see it now [10:53] didrocks: so it looks like initial search adele didn't trigger it but mea did so I'll try that again now [10:55] didrocks: yeap retried mea and it died again I have a theory I am about to try it out [10:55] ok [10:55] davmor2: I couldn't get something tight enough to find a particular music triggering it [10:55] can be the number as well [10:57] didrocks: search for something in the carousel over searching something that isn't [10:58] meh nope just proved that theory wrong [10:58] davmor2: yeah, no :/ [10:59] davmor2: mind updating the bug that Saviq pointed at? [10:59] Saviq: I can't make it recrash now… [10:59] ah that was the mir locking up so it might be that my theory was right [10:59] didrocks, after increasing the timeout? great [11:00] Saviq: yeah, it's clearly the timeout increase that fixed it… hem hem :p [11:02] Saviq: which of the 2 bugs do you want updating? [11:03] davmor2, whichever one fits more - one is for opening preview, the other for expanding the list of music [11:03] davmor2, although IIUC you're describing something completely different [11:04] * Saviq loves Dell... complained yesterday about my keyboard dying, have a new one in my hands already [11:04] Saviq: yeah I'll open a fresh one if it turns out it linked we can always link them [11:05] davmor2, indeed [11:09] larsu, do you know anything about the volume indicator? [11:10] Saviq: yes [11:10] larsu, so, I'm looking at "if you use the slider to change volume, using buttons then won't reach you 0% or 100%" [11:11] larsu, it looks like it becomes 0.5 - 1 → -0.5, rejected [11:11] larsu, and same at the other end [11:12] Saviq: interesting. Let me have a look [11:13] Saviq: hm, the code looks right to me. It clamps after doing the calculation [11:14] larsu, on the "volume" action, too? [11:14] larsu, that's what we activate with -1 / +1 [11:14] ya [11:15] let me try it out [11:15] larsu, you probably won't be able to due to the int vs. double issue [11:15] Saviq: I'm trying on the dbus interface directly [11:15] larsu, ah ok [11:17] Saviq: works fine here [11:17] larsu, ok, I'll dig more to see what's happening [11:17] does unity check the current value before activating the action? [11:17] larsu, no [11:17] hm, weird [11:17] larsu, https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1293478/comments/7 is what Ricardo said [11:17] Ubuntu bug 1293478 in unity8 (Ubuntu) "Volume buttons no longer function correctly since #238 mako/flo/manta" [High,Confirmed] [11:17] larsu, I didn't have time to repro, so let's leave it for now [11:18] okay [11:25] Saviq: the nexus10 being weird is something that happens in trunk too [11:25] in autopilot i mean [11:26] let me bisect [11:26] tsdgeos, yeah, right, forgot to mention that I saw that outside of new scopes... [11:28] didrocks, ok, added row 51 with just small fixes to unity8, can we get a silo (overriding the lock)? [11:29] Saviq: giving to you even a luxury silo (number 7!) [11:30] (I turned ready to yes) [11:30] didrocks, right, thanks [11:30] yay, I don't have to scroll 3 times to get to my silo ;P === MacSlow is now known as MacSlow|lunch [11:34] Saviq: weird, even going back one month in unity8 i get that bad shaped N10 :S [11:39] Saviq: bug #1294605 I've added the 2 crash files I see and I'll try to get a full BT on them now if I can [11:39] bug 1294605 in unity8 (Ubuntu) "Searching on the music scope is causeing unity8 to crash" [Undecided,New] https://launchpad.net/bugs/1294605 [11:42] davmor2, you might need to increase the kill timeout in /usr/share/upstart/sessions/unity8.conf [11:42] davmor2, I saw 30s not being enough [11:42] tsdgeos, I'm feeling 5.2 [11:43] biab [11:43] Saviq: will do [11:45] Saviq: you think 5.2 is causing the sizing thing? [11:45] may be [11:49] with new js, maybe [11:52] tsdgeos, what's the magic cmake declaration that puts qml files inside qtcreator project dir? [11:52] I'm trying to do the same with the wizard [11:53] Cimi: see src/CMakeLists.txt [11:54] Saviq: changing the 30 to 60 makes things searchable again with no more crash [12:07] davmor2, lol [12:07] Saviq: home scope is searchable now too [12:08] davmor2, not like that can have any impact, but didrocks reported exactly the same ;) [12:08] Saviq: don't even dare propose that as a fix! :) [12:08] Saviq: and if I drop down the music section that is no longer crashing either [12:09] didrocks, :D [12:09] Saviq: do it i dare you ;) [12:09] well, actually, that can be useful for CI or other people reporting [12:09] but clearly not what's going to fix it :) [12:10] Saviq: I can also preview music too so fixes that also :D [12:11] didrocks: but it fixes it ;) [12:11] pffff :p [12:11] didrocks: 3 bugs gone from one setting change it has to be a fix right :D [12:11] sure sure [12:12] * davmor2 picks up a tissue to wipe the tears from his eyes from laughing so much :) [12:13] didrocks: so now we know where the issue evolved from can I update to latest and do some daily testing, or Saviq do you need anything else? [12:13] davmor2, no, thank you, we'll have to dig into that [12:15] didrocks: right I'm updating to current then and starting the daily testing then [12:28] larsu, can you give me a dbus-monitor command to monitor the things going back and forth between unity and the sound indicator so we can see what's going on at the slider bounds? [12:30] Saviq, try "gdbus monitor --session --dest com.canonical.indicator.sound --object-path /com/canonical/indicator/sound" ? [12:31] seb128, thanks [12:32] Saviq, wait to see if that does what you want before thanking me ;-) (I didn't follow the details) [12:32] seb128, that's a start :) [12:32] Saviq, seb128: that only gives you signals from indicator-sound [12:32] larsu, I was unsure if you were at lunch so I replied with what I knew :p [12:32] if you want method calls as well, use `dbus-monitor destination=com.canonical.indicator.sound` [12:32] seb128: thanks :) [12:33] larsu, thanks! [12:35] Saviq, will test https://code.launchpad.net/~saviq/unity8/workaround-lp1293478/+merge/211702 when CI will build the package :) [12:35] Cimi, k [12:35] Cimi, actually you can grab silo 007 [12:35] Cimi, it's there already [12:35] Saviq: yeah, defenitely qt, want me to have a go at fixing it? [12:36] tsdgeos, would rather you do the crash / lockup first [12:36] Saviq: the qml one? that's done [12:37] or should be [12:37] tsdgeos, ah ok [12:37] has it run again? [12:39] tsdgeos, can you check the autopilot upstart tests that failed in desktop jenkins? [12:39] tsdgeos, that seemed not flaky, 'cause failed on both scenarios [12:39] Saviq, you mean https://launchpad.net/~ci-train-ppa-service/+archive/landing-007 ? [12:39] Cimi, yes [12:39] Saviq: do you have the url at hand? [12:40] tsdgeos, https://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/3533/? [12:41] Saviq, were ap tests run on the new-scopes silo? [12:41] oki, that was new-scopes right? [12:42] tsdgeos, yes [12:42] mhr3, they are in -ci, I didn't get around to yet [12:43] Saviq, how does it look [12:43] mhr3, positive [12:43] mhr3, just two failures tsdgeos is looking at now, unrelated to new scopes [12:43] mhr3, but we were green at some point, so we just need to get back to that [12:44] Saviq: so these two tests fail if you don't have scopes running [12:45] so i had the scope-registry stopped, run them -> fail [12:45] manually started it -> works [12:46] there's something out there that is making it not auto start the scopes i guess [12:46] anyone still has Qt 5.0 around? [12:47] nope [12:47] mzanetti, do you need it on phone or desktop? [12:47] doesn't matter [12:47] desktop more easily [12:47] something is broken in the right edge stuff which was definitely working last week [12:47] nice, now i stopped the scope-registry and also passes ^_^ [12:47] and going back revision by revision doesn't fix it [12:48] ah no false alarm, still fails [12:48] Saviq: so how do we make sure scope-registry is running for that test to succeed? [12:48] tsdgeos, upstart should take care of that [12:49] awesome, my mako just started remounting every 5 s [12:49] * Saviq gains a nautilus window ever 5s [12:49] *every [12:50] tsdgeos, if we have "restart" anywhere, we need to change it into stop/start [12:50] tsdgeos, bug #1294230 [12:50] bug 1294230 in upstart (Ubuntu) "initctl restart does not reload job config, so does not equal stop + start" [Undecided,Invalid] https://launchpad.net/bugs/1294230 [12:52] Saviq: maybe the wrong unity8.conf is being used?¿ [12:52] tsdgeos, weird, looking at that test it should be fine [12:52] tsdgeos, there should be no "wrong unity8.conf" nowhere [12:53] Saviq: well at least for me locally there's the system one and the one of new-scopes [12:53] that are different [12:54] tsdgeos, yes, but upstart won't use it unless you copy it to ~/.config/upstart/ [12:54] tsdgeos, which you probably should in your case [12:54] Saviq: right, so is using the system one that doesn't do "emits scope-ui-starting" [12:54] so it makes sense it fails, no? [12:54] tsdgeos, but for otto this should "just happen" [12:54] right [12:54] otto is a different matter [12:54] tsdgeos, yeah, but on otto it shouldn't matter, since the thing is installed [12:54] or should be, at least [12:55] * Saviq reinstalled trusty yesterday with nvidia-prime... no idea how/why but (dying) battery life improved like twice... [12:56] Saviq: other than that maybe we need to wait a bit to give time for the scopes to load [12:56] going for lunch [12:56] will try that later [12:56] enjoy === alan_g is now known as alan_g|lunch [13:11] tsdgeos, yea, the timeouts to wait for scope startup are pretty low, but seems to work mostly ok === _salem is now known as salem_ === alan_g|lunch is now known as alan_g [14:03] Saviq: any idea what's this? https://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/3546/console [14:04] just general jenkins flakyness? === MacSlow|lunch is now known as MacSlow [14:14] tsdgeos, the branch lp:~macslow/unity8/snap-decisions-states is approved, but jenkins still needs to settle... so it'll probably still take a bit before you can more easily review/test lp:~macslow/unity-notifications/multiple-snap-decision-example [14:15] ok, thanks :) [14:22] Saviq: you said yesterday that otto wasn't being reinstalled between runs or something? maybe that's the cause of https://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/3533/ ? [14:28] tsdgeos, yeah, that's just jenkins going crazy [14:28] tsdgeos, that got fixed, and it wasn't otto but devices [14:28] ah _D [14:28] ok [14:28] not that then [14:56] rsalveti, larsu so, looking at dbus monitor the backend actually does set the volume to 0.0/1.0, it's just the slider that doesn't update: http://paste.ubuntu.com/7120192/ [14:57] * Saviq files another bug [14:58] dednick, can't assign to you: bug #1293478 [14:58] bug 1293478 in unity8 (Ubuntu) "Volume buttons no longer function correctly since #238 mako/flo/manta" [High,Confirmed] https://launchpad.net/bugs/1293478 [14:58] Saviq: yeap. looking at it now [14:59] looks like that code was overlooked. Didnt know we still used it [15:01] dednick, how long for a fix do you think? should we merge the workaround or drop it? [15:02] Saviq: hour or 2 probably. [15:02] "hopefully" . if the fix works [15:02] didrocks, wdyt ↑ land unity8 workaround or wait for proper fix in qmenumodel (volume up/down) [15:03] sil2100, btw, how did we not see a meaningful error message for the temp app do you think? [15:06] Saviq: you mean, in the case of url-dispatcher? :) [15:06] sil2100, yes [15:06] sil2100, sounds like we should improve the test somewhere to get a meaningful error [15:06] Saviq: I have no idea, actually once you open up the whole console log you can see the error message ther [15:07] sil2100, oh [15:07] Saviq: but the big console log file is for all the tests being ran on the CI infra [15:07] Saviq: so it's a big big file - and I was not aware that not all stderr is not being collected [15:08] sil2100, what was the actual issue? couldn't execute url-dispatcher 'cause it wasn't there or? [15:09] Saviq: it was simply not there - we couldn't reproduce it locally since we were installing unity8-autopilot [15:09] Saviq: so yes, in the end the infra was doing things differently - but this time not phablet-test-run was the problem, but the actual test setup ; [15:09] ;/ [15:09] sil2100, the command, you mean? so url-dispatcher-tools? [15:09] Yep [15:09] sil2100, so yeah, test should bail out when trying to exec it, shouldn't it... [15:10] Saviq: well, the test is not written like that... it assumes it's there, and just executes it from the system shell [15:10] Saviq: it assumes it's there as it's a dependency of the test package [15:10] sil2100, doesn't matter, a call to a non-existent executable should raise [15:11] It seems it wasn't thought that someone would actually execute the test without that - maybe we should fix that anyway [15:11] sil2100, let me mention... bug #1262879 again [15:11] bug 1262879 in Ubuntu CI Services "There should only be one, documented, way to run tests on devices" [Undecided,Confirmed] https://launchpad.net/bugs/1262879 [15:12] Saviq: ha, -geometry applies only to qwidgets (no idea why) [15:12] tsdgeos, hmm are we not parsing it explicitly? should we be? [15:12] Saviq: we trusted Qt to do it [15:13] Saviq: not sure if it should raise, as it's done by os.system() [15:13] tsdgeos, sounds like a bad idea :) [15:13] Saviq: but since there's no documentation on what the -geometry applies (or i can't find it) not sure we can blame them for changing the behaviour [15:13] sil2100, probably that's the wrong thing - should use Subprocess.check_call [15:13] Saviq: so it's like, executes a command, cannot find it, noop [15:14] sil2100, yup, that's what I'm saying is wrong - should use check_call so that it bails out if the dispatcher fails for whatever reason [15:14] Saviq: indeed - now that you mention it we should basically first check if it's there and bail out gracefully [15:14] Maybe even skip the test then [15:14] sil2100, no, ask for forgiveness, not permission :) [15:14] sil2100, nope, it should fail [15:15] ;) [15:15] sil2100, or well, there's plenty of things we could do - use liburl-dispatcher or whatever, assuming there're py bindings for it [15:15] sil2100, or call dbus directly [15:15] whatever we do, it should fail if url-dispatcher fails [15:15] I'll make it happen [15:16] tsdgeos, ;) [15:16] Saviq: thanks! [15:32] Saviq: https://code.launchpad.net/~aacid/unity8/windowgeometry/+merge/211760 [15:33] tsdgeos, so Qt was swallowing -geometry? [15:33] s/was/started/ [15:33] yay [15:34] it's interesting that they eat args that they don't read anyway... [15:34] grr [15:35] Saviq: they read [15:35] tsdgeos, just ignore ;) [15:35] to apply later [15:35] but the later is only in QWidget [15:35] tsdgeos, yeah, understand === alan_g is now known as alan_g|tea [15:47] Cimi, top-ack https://code.launchpad.net/~saviq/unity8/workaround-lp1293478/+merge/211702 ? [15:48]