=== duflu_ is now known as duflu [05:25] Good morning [07:55] hey pitti [07:56] hey darkxst, how are you? (haven't forgotten your testimonial, will get to it ASAP) [07:56] pitti, I'm good and you? [07:56] thanks btw ;) [07:56] quite fine, thanks! [07:57] fwiw autopilot fails miserably at introspection a gnome-shell session [07:58] I suppose it needs to learn about ST and Clutter widgets? [08:45] hey seb128 [08:45] good morning desktopers [08:45] darkxst, hey [08:46] seb128, not to bug you or anything but can you add something to my application, meeting is on monday! [08:47] sure can, I'm going to do that today [08:47] thanks! [08:52] moin all! [08:53] hey Sweetshark [08:53] seb128: Found two patches on upstream master that should fix the issue. I did not find a ppc porterbox anymore on the wiki -- do we still have one? [08:53] not that I know [08:55] seb128: so we are down to just trying another upload with the patch? [08:56] Sweetshark, I guess so, though I've a feeling Laney is not going to like that one ;-) [08:56] let's just upload before he gets on IRC [08:58] seb128: k ;) [09:02] morning [09:03] also WHAT [09:03] :P [09:04] I've been abusing partch.debian.org [09:04] you might have a case for asking for a native PPA though [09:07] Laney, good morning! [09:08] Laney, Sweetshark asked for native PPA in the past for libreoffice (especially to workaround the disk space limitation of the virtual ones) but they have never been wanting to allocate those resources to libreoffic [09:09] I was thinking for distro uploads only [09:09] like ideally you'd copy with binaries out of it [09:09] but shrug [09:10] I guess that's not much different from direct archive upload if all the preparation work is done in a virtual one [09:11] ho hum [09:11] we do need a porter machine really [09:11] indeed [09:12] tell señor warner [09:12] right [09:13] * seb128 adds to the list of requests [09:34] sil2100, bregma: is anyone working on having an unity7 landing in trusty? [09:36] seb128: backported patches send to review upstream, also added in the source package, which is building right now. [09:36] Sweetshark, great, thanks [09:36] seb128: dont tell Laney ;) [09:36] yeah, shusssh [09:37] you'll have to just try it [09:37] unless infinity wants to donate the use of his private machine to test build libreoffice :-) [09:39] argh [09:39] larsu, could you review https://code.launchpad.net/~attente/indicator-keyboard/fix-tests-gdk-display-error/+merge/202152 ? (that should fix the current fail to build in trusty) [09:40] seb128: sure [09:40] larsu, thanks [09:40] got that nouveau lockup again [09:41] woah, why does it make all those timeouts loop? [09:41] * larsu hates unrelated commits in the same MR [09:44] Laney, btw did you see my autopilot errors with your branch? [09:44] saw the comment [09:44] don't get those errors [09:44] hum, k, I guess it's up to me to debug it then :/ [09:45] can you isolate one test that causes it? [09:46] yey, vala code! :) [09:46] what a non-valaish way to do fixtures :P [09:47] seb128: why was this necessary? Is some test unsetting DISPLAY? [09:48] larsu, not sure, but you can try to build the package on trusty without that commit [09:48] larsu, https://launchpadlibrarian.net/162075569/buildlog_ubuntu-trusty-i386.indicator-keyboard_0.0.0%2B14.04.20131125-0ubuntu1_FAILEDTOBUILD.txt.gz [09:49] larsu, well, not hurry to land that, let's wait for attente to be online if you have questions [09:49] larsu, fixtures the vala-way - http://bazaar.launchpad.net/~unity-team/libunity/trunk/view/head:/test/vala/test-scope-base.vala :) [09:50] seb128: the code looks fine to me, but I'd rather talk to him about the reasoning. If you're fine waiting a couple of hours, I'd prefer that [09:50] mhr3: right, this is what attente is doing as well, no? [09:50] oh! [09:50] larsu, sure, I pinged for review because it was in the queue for a week and I was unsure if you or charles noticed it [09:51] larsu, he does pointers and stuff.. eek :P [09:51] mhr3: ya, just realized. I'll point him to that, thanks :) [09:51] seb128: I didn't, thanks [09:51] yw ;-) [09:51] won't that fix only work if you have a DISPLAY already? [09:53] Laney: yeah but I think the test will fail anyway if there's no DISPLAY [10:02] yeah it does make it pass [10:12] seb128: uploaded -0ubuntu2 to http://people.canonical.com/~bjoern/trusty/ [10:12] seb128: I also uploaded a new -l10n version, just to be safe. [10:12] Laney, nice, the Unity API team is adding a connectivity API (the work is 80% done from what they say) [10:12] what's that for? [10:13] Laney, that should provide the "can I use internet" you asked about previous cycle [10:13] Laney, https://docs.google.com/a/canonical.com/document/d/1BTxpGvhe3pvbwN8vVbu8tQUVV-woHEA6pYo_ERkTP9U/edit [10:13] decent [10:14] seb128: In theory the older -l10n version should work too. But as you know: In theory, theory and practice are the same. In practice ... [10:14] Sweetshark, what are you concerned about? installability issues? [10:15] Sweetshark, it seems like now would be a good moment to check if a version mismatch works? [10:15] seb128: I dont trust anything that I didnt test explicitly. ;) [10:15] seb128: yeah, lets test it. [10:17] seb128: FWIW, ricotz already did a non-rebuild pocket copy of -l10n from trusty to saucy. So if that work with the rebuild saucy build, it is giving some hope to not be totally pessimistic about it ;) [10:17] s/work/works/ [10:17] (and it seems to work: no complaints so far from any saucy users) [10:17] good [10:19] seb128: hold on! [10:19] seb128: lemme check something. [10:24] seb128: ok, go. [10:25] * Laney sneaks wk in there [10:25] (It seemed the patch was not applied, but that was an old jenkins log) ;) [10:28] Sweetshark, ok [10:34] Sweetshark, uploaded [10:34] aww, I didn't get sagari [10:34] oh, Laney uploaded webkit as well [10:34] buildd DoS [10:34] lo didn't either [10:34] fail [10:35] :-( [10:41] buh [10:41] select_many(toolkit_emulators.Base) doesn't work [10:42] seb128: thanks [10:43] seb128: /me sternly watches ppc build. [10:43] * Laney too ... [10:51] oh, my blog might break its all time viewcount record today ... [10:57] aha [10:57] AHA! [10:58] Laney: !?! [10:59] a minor victory over autopilot [10:59] Laney, did you figure out the error I'm getting? ;-) [10:59] Laney: watch your flank, its already plotting its revenge! [10:59] hah, no [10:59] I got the list of installed click packages [11:00] well, I got the number '4' which is the number of installed click packages, so hopefully the list contains the right things ... [11:02] looks like it [11:03] * Laney employs the suggested sorting hack [11:25] Laney, btw, want some non-autopilot-uss-hacking? [11:27] maaaaaaaaaaybe [11:28] Laney, powerd got updated, need to change the power panel to use dbus rather than writes to /sys [11:28] Laney, http://bazaar.launchpad.net/~phablet-team/powerd/trunk/revision/107/data/com.canonical.powerd.xml [11:29] Laney, is that something you would be interested to do? [11:29] I'm not sure where you mean [11:30] don't we just poke the indicator? [11:30] right, because we didn't want to reimplement the "write to /sys" [11:31] do you think we should keep using the indicator, or just do the dbus call directly now that we have those? [11:31] sharing backends always seemed desirable to me [11:31] ok, wfm [11:32] Laney, in fact I think what Bill was mostly interested in was to get the option to enable/disable the auto level [11:33] Laney, that's in https://wiki.ubuntu.com/Brightness#Phone [11:33] Laney, want to have a look to add that panel? the slider should be just copying power, the checkbox is that new dbus call [11:35] okay, should be easy [11:39] Laney, thanks === MacSlow is now known as MacSlow|lunch [12:09] Hey guys. The location indicator if you select one of the 2 options the indicator closes is this the expected behaviour? === alan_g is now known as alan_g|lunch [12:51] bah [12:51] Laney, I can't reproduce the autopilot error from yesterday today, neither with a stock build nor with my hacked geometry (reduced height value) [12:52] the test_storage tests are unhappy with a height: 40 though [12:52] but that's not a blocker for that merge [12:52] * seb128 mark the mp back to approved === Sweetsha1k is now known as Sweetshark [12:53] O RLY [12:53] I only made it 45... could that really make a difference? [12:53] Laney, http://paste.ubuntu.com/6802738/ [12:55] Laney, same issue with 45 [12:56] Laney, seems a flickering issue [12:57] the drag moves the pointer over the "license:" label [12:58] sorry "Legal" [12:59] (they translated to License in french) [13:01] so unreliable [13:01] i've not seen that happen [13:01] :/ === MacSlow|lunch is now known as MacSlow [13:13] exciting [13:13] was that a CI train release? [13:15] Laney, yes, like the few previous ones ;-) [13:16] Laney, was there anything weird/different? [13:16] didn't notice other ones [13:16] I just saw that it included some random merge and not other things [13:17] now I see them [13:17] cool [13:17] ;-) [13:18] Laney, so, if those tests are working for you and not for me, I think it's because our pointer setting are differently (random guess) [13:18] https://mblankhorst.nl/etc/supertuxkart.svg.gz muahahaha (perf timechart recording gpu channels, 60 mb svg, will kill your system) :P [13:18] Laney, e.g the same action on the pointer leads to different amonth of flickering/scrolling [13:18] I guess [13:19] what did you do the main screen which is different from what you do to the about/license one? [13:19] because it doesn't have that issue to reach "about" on the mainscreen [13:20] we launch straight to the about panel === alan_g|lunch is now known as alan_g [13:42] mhr3, larsu, hey [13:50] attente: morning [13:51] larsu, mhr3, i just caught up on the conversation [13:51] mhr3, thanks for pointing me to that file! [13:51] * larsu just wants to know why DISPLAY must be re-set [13:52] it'd be cool if you mentioned that in the commit msg [13:53] larsu, sure [13:53] larsu, it gets cleared in g_test_dbus_up() [13:54] really? Why? [13:57] i'm seeing this in the docs: [13:57] Unset DISPLAY and DBUS_SESSION_BUS_ADDRESS env variables to ensure the test won't use user's session bus. [13:57] This is useful for unit tests that want to verify behaviour when no session bus is running. It is not necessary to call this if unit test already calls g_test_dbus_up() before acquiring the session bus. [14:00] attente, here's the actual magic http://bazaar.launchpad.net/~unity-team/libunity/trunk/view/head:/test/vala/common.vala [14:00] attente: hm, makes sense. thanks [14:01] mhr3, ah, looks useful, thanks [14:08] hiya [14:08] can somebody take a look at https://bugs.launchpad.net/ubuntu/+source/metacity/+bug/907640 please? [14:08] Launchpad bug 907640 in unity-2d (Ubuntu) "alt-backtick doesn't work properly" [Low,Confirmed] [14:12] cyphermox, do you have an opinion on https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/1271567? [14:12] Launchpad bug 1271567 in network-manager (Ubuntu) "Include upstream patch for the ifupdown plugin to honor the source stanza" [Undecided,New] === m_conley_away is now known as m_conley [14:39] dholbach: ohh, shiny [14:39] dholbach: yes, I'm about to upload a new NM, I'll include that [14:45] tseliot: hey dude any news on the optimus driver on 3.13 yet? [14:45] * dholbach hugs cyphermox [14:57] davmor2: not yet, I'm still waiting for news from nvidia... [14:58] tseliot: no worries dude :) [15:01] kenvandine, Laney, tedg, attente: settings hangout? [15:01] charles, ^ [15:01] coming [15:01] Wellark: ^! [15:01] Laney, he's there already [15:01] Laney, as is mpt [15:01] seb128, i don't really have anything to add, let me know if you need me [15:01] ooh fancy [15:02] tedg, you're going to hate me... i've bee rethinking how the hub launches apps... i think i need to use url-dispatcher [15:02] kenvandine, I wanted to ask you some details about what is missing in the background, I'm not sure I grasp it (it seems to be pretty complete to me with the backgrounds grid etc) [15:02] but i probably still need to use that new api i added to ual to determine if it is already running [15:02] seb128, how about you and i chat 1:1 after the meeting? [15:02] we can even do a hangout :) [15:02] kenvandine, ok, wfm [15:02] he's naked [15:02] haha [15:02] hehe [15:02] you can tell by this evasion [15:02] you wish :) [15:02] phwoar [15:03] kenvandine, Hmm, at least you gave me permission to hate you :-) [15:03] tedg, you always have permission to hate me ;) [15:03] seb128, I don't have any settings stuff, do you want me there? [15:03] tedg, no need no [15:03] tedg, i realized we had no way to handle an app that might provide more than one handler of the same type [15:04] kenvandine, Do you need to talk to me about any of that, or is there enough there for you now? [15:04] like system-settings background and sounds plugins could both implement an import handler [15:04] Yeah, you'll have to do that. But how does URL dispatcher help? [15:04] so i think i need to register settings:///background and settings:///sound as separate handlers [15:05] and other apps that need to do something similar could implement a parser for url args [15:05] like appid:///someid/app/1.0/foobar [15:06] tedg, does that make sense? [15:06] kenvandine, I was going to use the fourth item for FD.o actions. [15:06] Though not sure that'll make 14.04. [15:06] will there be an API for parsing that stuff? [15:06] * tedg had hoped, but is too behind at this point. [15:06] kenvandine, No API needed, people just define the actions in their desktop file, so it turns into an Exec line of their choosing. [15:06] it would make sense to have an api for that, so you can reliable grab args [15:07] which the args could be the last position [15:07] ah [15:07] so that would be exactly what i need :) [15:07] * tedg hates kenvandine [15:07] :-) [15:07] system-settings already handles this since it has it's own handler, settings:/// registered [15:08] Uhm, so yeah. It's just a mater of software.... [15:08] but thinking about potential 3rd party apps that want to do something similar [15:08] actions would work though [15:08] For custom URLs I think they'll have to parse them themselves. [15:08] we can get those from inside qml right? [15:08] I think that's a different use-case [15:08] with the ActionManager [15:08] I'm not sure how it's done, but yes. [15:08] You also get notified on second activation as we send them via dbus. [15:08] so for those apps it could be handled that way [15:09] well, remember they will be part of a trusted session embedded with another app [15:09] * kenvandine really fears how that might behave in the wild [15:09] I'm not quite understanding that. The GL contexts will be embedded, but the process would be in it's own confinement, right? [15:10] with the trusted session, it will always start a new instance tied to that trusted session [15:10] same confinement rules [15:10] but it'll appear as a single app with whatever it is embedded in [15:11] my fear is app developers that might do things that aren't multiple instance friendly [15:11] Sure, but most won't be providing content-hub sources. That's an advanced app. [15:11] i'm thinking about things that might do things like locking a db, etc [15:11] But, how are you working around Upstart single instancing stuff? [15:12] i'm not... tvoss and the mir folks are :) [15:12] In theory there could be two instances of the same app in that case. [15:12] Wondering if he's just planning to kill the other instance. [15:13] not kill [15:13] "shutdown" [15:13] but it will be suspended... and potentially killed later [15:13] based on app lifecycle [15:13] but you could switch between them and resume it [15:13] You can call it "put out to pasture" if you like. :-) [15:14] anyway, they said they'll handle it :) [15:14] Hmm, curious. #ubuntu-touch. === dpm_ is now known as dpm [15:20] Laney, going to test on the device in a bit, I didn't touch my phone for some days and it's flat again, so recharging a bit before testing [15:20] righto [15:22] mpt: do you have a mockup of the new header design? [15:23] Sweetshark, shrug [15:23] Laney, no. [15:23] okay [15:24] Sweetshark, https://launchpadlibrarian.net/163197609/upload_5491603_log.txt [15:24] Hmm [15:25] that log could be more useful [15:25] Laney, indeed :/ [15:25] it often means you tried to upload binaries which exist at a newer version already [15:25] which is not the case here [15:25] https://launchpad.net/ubuntu/+source/libreoffice/1:4.2.0~rc2-0ubuntu2 [15:30] rvr: You gave your presence away [15:30] rvr: Off the top of your head, can you tell me how to select something to click on it if I know its display text? [15:31] Laney: Sometime it happens to me that I discover what's the problem just when I was about to ask about it :D [15:31] ;) [15:31] I want to click one of the options in an ItemSelector [15:31] item_selector.select_single(text="blah") doesn't work - StateNotFound [15:31] Laney: Use the introspector [15:32] it's there [15:32] It gives you the tree and clues about how to get where you want [15:32] Tip: if you cannot find it using objectName, then you are doing it wrong ;) [15:32] umm [15:33] It's ItemSelector { model: ["option A", "option B"] } === alan_g is now known as alan_g|tea [15:34] Laney, did you copy that scroll_to_and_click() from somewhere? [15:34] no [15:35] the logic came from uss-online-accounts [15:35] autopilot doesn't provide by itself some "scroll to widget and click" feature? [15:35] it doesn't [15:35] :-( [15:35] that's one of the main sources of annoyance [15:35] Laney, so my license issue it's not the flickering, so I guess something is wrong is that function of yours [15:36] * seb128 tries to understand what that is doing [15:37] scroll until the thing you want is on the screen, and then click it [15:37] self.pointer.click_object(the_object) [15:37] rvr, right, that works when the object you want to click is on screen [15:37] rvr, if it's down a column and you need to scroll, it fails [15:38] Ah, yeah, the scroll problem [15:38] is anyone working on providing a standard solution to that problem? [15:38] I think the emulators will implement it for each thing [15:38] seb128: hmmm, I recently updated to a new gpg-key maybe its that? launchpad seems to know the key though .... [15:38] if I understand right then ListView already works [15:39] Sweetshark, no, launchpad glitch (read #ubuntu-devel backlog) [15:39] Sweetshark, I retried the build [15:39] okay, I'm doing alpha 2 publishing for 30 minutes or so [15:40] seb128: ah, good it hit only a small build and not something like LibreOffice that would squatter on an arm builder for another day. [15:40] Sweetshark, it's only i386 which failed to upload, so that's ok ;-) [15:40] seb128: k ;) === alan_g|tea is now known as alan_g === gatox is now known as gatox_lunch [15:44] Laney, shrug [15:44] # avoid a flick [15:44] sleep(0.2) [15:44] Laney, if I use 0.5 there it works fine [15:44] or 0.3 [15:45] hmm [15:45] yeah for random sleep calls as solution :p [15:45] there might be a property [15:45] while (flicking): sleep(0.1) or something [15:45] while (moving) [15:45] I dunno :( [15:54] Laney, ah, you did it differently from u-s-s-online-account [15:54] Laney, they do "press, move, wait, release" to release on a stopped movement to avoid the inertia/flickering [15:54] bugs meant I couldn't do that [15:54] Laney, you do "drag; sleep" [15:55] the pointer is not pressed thing I was going on about [15:56] on the device I guess? [15:56] yup [15:56] look a bit up in uss-oa [15:56] kenvandine, are u-s-s-online-account autopilot tests working on the device? [15:56] that code doesn't run on device [15:56] oh, right [15:57] last i heard it didn't [15:58] Laney, ok, feels like it's as good we can get on a flacky framework ... [15:58] shall I make it 0.5? [15:58] having them flip flop will be really annoying [15:58] Laney, if you make it >= 0.3 it works here [15:58] safety margin ... [15:58] so, yeah, increasing the value makes sense [15:58] 0.5 wfm [15:58] ok [15:58] can you do that in the mp I already approved? [15:59] will do [15:59] I'm going to reapprove/queue it in CI train for landing [15:59] thanks [15:59] ok, phone charged a bit, let me try things there [15:59] could we get someone to test run it in the ci infrastructure? [15:59] although last time we did that it came out as passing ... [15:59] sil2100, didrocks, kenvandine: ^? [16:11] seb128, do you think somebody can take a look at https://bugs.launchpad.net/ubuntu/+source/metacity/+bug/907640? [16:11] Launchpad bug 907640 in unity-2d (Ubuntu) "alt-backtick doesn't work properly" [Low,Confirmed] [16:12] Laney, seb128: what CI infra do you mean and for what functionality? [16:13] run lp:~laney/ubuntu-system-settings/ap-fixes [16:13] to see if it works [16:13] autopilot [16:15] dholbach: I don't understand it completely but it seems plausible to me [16:15] feel free to sponsor if you want [16:15] dholbach, dunno, don't ask me, I never touched unity2d, that's a question for sil2100 or didrocks [16:16] it's metacity [16:16] right, same difference [16:16] I never touched wms [16:17] seb128, as long as lamont doesn't complain about focus of terminals he's typing in, you should be fine, right? :-) [16:17] interesting [16:17] dholbach, lol [16:17] metacity is still on the desktop image [16:17] the lib is used by the compiz decorator [16:18] is ubiquity still using it, or did they finally move to compiz? [16:18] http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.trusty/rdepends/metacity/metacity [16:18] seb128: we've reverted from compiz back to metacity [16:18] xnox, k, so you are the maintainer for it? good, dholbach has a patch for you [16:19] thursday trolling, it's a new thing [16:19] DON'T BITE! [16:19] seb128: also unity7 depends on compiz which depends on the metacity decorators =) [16:19] maybe have a look at 1263756 too? ;-) [16:19] seb128: if it's about decorators -> I don't care, if it doesn't affect ubiquity -> I don't care +)))))) [16:19] xnox, they are fixing that apparently, there is a branch from Trevinho up moving decoration in unity proper and disabling the decor plugin from compiz [16:20] seb128: excellent =) [16:20] yeah, no more metacity dependency anywhere [16:20] seb128: then i'll switch to something else Xvfb? =) [16:20] * xnox things it would be the most useful way to run installer [16:21] Trevinho, btw, those changes are not small, isn't that risky for a lts cycle? [16:21] seb128: better that, then have constantly crashing decorators process which is not upstart managed / respawned (last time i've checked) [16:21] Laney, hum, http://paste.ubuntu.com/6803664/ [16:21] Laney, 1 error running u-s-s tests on the device [16:22] seb128: well, it works better than old one... THe code is quite simple, and logic is based on the old one... So I've not experienced any regression so far [16:22] oh noes [16:22] Trevinho, ok, good ;-) [16:22] So, let's try to get that in :P [16:22] Trevinho, I'm not going to complain about less old gtk code and deprecated depends [16:23] seb128: it also would allow to grow things with higher dpi screens, or bigger text [16:23] dholbach: seb128: that patch makes no sense to go into trusty =) === gatox_lunch is now known as gatox [16:25] dholbach: seb128: that patch is needed for unity-2d, of which last one was in precise. [16:25] ok [16:27] Laney, Ran 39 tests in 212.210s [16:27] Laney, on another test [16:28] Laney, so maybe I got unlucky on first try [16:28] that makes me worried [16:29] yeah, me neither, I don't like flakyness [16:29] I'm doing another run [16:33] Laney, shrug, got 3 of those this time, http://paste.ubuntu.com/6803705/ [16:33] man, what is up with your device! [16:33] I ran it like 1000 times [16:33]