=== alan_g is now known as alan_g|afk [08:00] greyback, ping [08:02] Saviq: pong === didrocks1 is now known as didrocks === alan_g|afk is now known as alan_g [08:42] unity8 and several apps have a "search entry" (e.g one with a magnifying glass icon on one side and a crossed circle to clear it on the other side) ... is that using the standard TextField? or is there a custom widget doing that which can be re-used? [08:44] ok, seems like those are TextField with a primaryItem [09:33] Saviq, why's unity-shell-launcher=2? [09:33] *what's [09:33] doesn't get installed with ./build -s [09:33] Cimi, did you merge from trunk? [09:34] Cimi, you need ppa:ubuntu-unity/next [09:34] Cimi, instead of the other PPAs [09:34] ah [09:34] Saviq, I have trunk [09:34] Cimi, do you have ppa:ubuntu-unity/next? [09:34] Cimi, it's been merged in http://bazaar.launchpad.net/~unity-team/unity8/trunk/revision/106 [09:35] Saviq, installing now [09:35] weird [09:36] Saviq, which package? [09:36] Saviq, still complaining [09:37] damn ^W [09:37] Cimi, libunity-api-dev [09:37] Saviq, let's add it to the build deps [09:37] Cimi, orly? [09:37] Cimi, it's there [09:38] Saviq, :\ [09:38] Cimi, ./build will build a unity8-build-deps package [09:38] Cimi, and install it, installing all the build deps [09:38] Cimi, ./build -c, btw [09:38] Saviq, indeed it's in debian control [09:38] Cimi, when in doubt ./build -c [09:39] Saviq, maybe we need to update the version of this package? [09:40] Cimi, no [09:40] does it get reinstalled? [09:40] Cimi, yes [09:40] Cimi, but only if debian/control is newer than builddir/unity8-build-deps*deb [09:40] http://paste.ubuntu.com/5880241/ [09:40] Cimi, so if it failed once, you need --clean so that it gets reinstalled [09:40] ok [09:41] it's newer but didn't work [09:51] Saviq: I'll start making people pay when they use the term "autolanding" :p [09:51] or CI as well ;) [09:51] didrocks, :D [09:51] like a swear-jar? [09:51] right [09:52] people are always confused with the term autolanding or CI [09:52] didrocks, what shall we use? [09:52] Saviq: upstream merger [09:52] and daily release [09:52] Saviq: want me to do a followup email? [09:53] didrocks, sure, but upstream merger covers both CI and autolanding, really [09:53] or it doesn't cover CI at all, depending how you look at it [09:53] Saviq: well, it's the same in the way you are using it [09:53] CI = autolanding :p [09:55] didrocks, not in my book :P [09:55] maybe not in your book [09:55] but in your email, yeah :p [09:55] * autolanding failed due to HW issue [09:55] * decided to fast-track without CI [09:55] didrocks, autolanding ( CI, merging ) [09:56] Saviq: what is CI for you? [09:56] didrocks, the thing that unity8-ci does [09:56] didrocks, so testing [09:56] didrocks, but I agree that's the wrong name to use [09:56] as there's no integration going on there [09:56] Saviq: right ;) [09:56] let me send a followup email [09:56] didrocks, I'm about to send my comments [09:56] didrocks, might want to wait for that [09:57] Saviq: well, I want to give a rationale first and add fghinter :) [09:57] didrocks, sure [09:58] Saviq: daily release was triggered manually, right? [09:58] Saviq: seeing the hours [09:58] * didrocks checks the hours [09:58] didrocks, I didn't, maybe Sergio did [09:59] didrocks, OTOH it was over 2hrs between landing and daily release [09:59] Saviq, seb128 https://code.launchpad.net/~unity-team/unity8/unity8.background_gsettings/+merge/174958 [09:59] Saviq: the daily release starts at 2h30 for unity8 [10:00] Saviq: sent [10:00] Cimi, lines 22 23 of the diff look weird [10:00] brb [10:01] Cimi, ah, know I get it... [10:01] Cimi, why do you build-depends on gsettings-desktop-schemas ? shouldn't you depends on it rather? [10:01] seb128, don't ask me about debian stuff :P [10:01] seb128, feel free to tell me what to change boss [10:01] Cimi, same for qtdeclarative5-gsettings1.0 [10:01] right [10:02] didrocks, shouldn't have […] snipped if you added fginther to CC [10:03] didrocks, unless you Fwd, too? [10:03] Saviq: oh, good point, silly me :p [10:03] Saviq: let me fw him the first discussion [10:03] Saviq, seb128 pushed [10:04] Saviq: "fixed" :p [10:04] didrocks, I wonder if we should rename the Jenkins jobs, then [10:04] didrocks, 'cause really that's where it all comes from [10:04] Saviq: yeah… I know :/ [10:04] didrocks, and the bot votes "continuous-integration", too [10:04] Saviq: I have no strong opinion on what should be done TBH, but right… [10:05] didrocks, btw, would triggering manually at 1am have helped anything? [10:05] didrocks, triggering daily, that is [10:06] Saviq: not sure if the platform stack would have been finished by then [10:06] Saviq: and no image build anyway [10:06] Saviq, info graphics don't work with it [10:06] so I would say "no" [10:06] Saviq, it is of a fixed purple [10:06] Saviq: and can't tell you, seems jenkins on magners decided to shut down… [10:06] Saviq, shall I fix it here or new bug and separate branch? [10:08] Saviq, also, the image should be cropped not scaled.. [10:11] Cimi, you mean that infographics don't change their colour with the background? yeah I know, it's nic-doffay's shortcut ;P [10:11] Cimi, so no, not the same MP [10:11] Cimi, as for scale vs. cropped, yeah, fix it in there [10:12] Saviq, http://imgur.com/GzRBHD1.png [10:12] weird [10:12] Saviq, :P [10:12] Saviq, this is with Image.PreserveAspectCrop :-\ [10:13] Cimi, doesn't look like it :D [10:13] Cimi, remember it's a CrossFadeImage [10:13] Cimi, maybe something to change there? [10:14] Saviq, mmm yeah might be [10:14] Cimi, while you're at it [10:15] Cimi, didn't CrossFadeImage get into the SDK? [10:15] * Cimi checks [10:15] Cimi, we could maybe get rid of it from our code [10:15] I am still fighting with the massive abundance of italian ice cream I had last night as personal welcome :P [10:26] seb128, you need to increase max flicking velocity in "Software licenses" :D [10:26] seb128, getting to unity8 is tricky ;) [10:27] Saviq, hehe, indeed, I noticed that as well [10:27] seb128, and it doesn't integrate with the header well (header doesn't scroll off screen and list isn't clipped) [10:27] Saviq, I was wondering if the slowness is due to the fact that the list is built "on demand" when you scroll [10:27] seb128, no [10:27] seb128, http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-flickable.html#maximumFlickVelocity-prop [10:27] Saviq, will try that, thanks [10:28] do you have a value to recommend? [10:29] seb128, we went for: [10:29] maximumFlickVelocity: height * 10 [10:29] flickDeceleration: height * 2 [10:29] seb128, in the dash [10:29] ok, thanks [10:29] Saviq, noted for the header ... what about the list clipping? [10:29] do you know what's the issue there? [10:29] seb128, well, the list does not have clip: true at least [10:30] seb128, but then it should push the header off screen [10:30] seb128, and am not sure how that works in the SDK [10:30] the header problem is there in the storage panel as well [10:30] I need to look at that [10:31] the pagestack is supposed to do that for us [10:31] Saviq, mmm [10:31] Saviq, small question [10:31] Cimi, stop eating icecream [10:31] lol [10:31] unless you have enough to share! [10:31] Saviq, so the sourceSize set to dimensions of the shell [10:31] Saviq, makes the image blurred [10:32] because there's a massive crop [10:32] Cimi, yeah, we probably shouldn't have sourceSize there [10:32] Saviq, maybe I should just set sourceSize width or height depending on the crop? [10:32] Cimi, or just one dimension at most [10:33] Saviq, I'd say just height maybe? [10:33] wallpapers should be landscape no? [10:33] Cimi, not necessarily, no [10:33] or how can I be smarter? [10:33] Cimi, I'd probably drop the sourceSize there [10:33] Saviq, we get 2MB of men usage no? === sil2100_ is now known as sil2100 [10:34] *mem [10:34] in case of big image [10:34] s [10:34] Cimi, hopefully Qt is smart enough to drop the parts of the image that are out of bounds [10:34] Cimi, otherwise we'd have to load the image, check its aspect ration [10:34] -n [10:34] Cimi, and then decide [10:34] Saviq, I am doing it [10:35] with TestImage [10:35] already [10:35] so what can I add here? [10:35] help me thinking [10:35] we have aspect ratio of the phone/tablet [10:35] with width and height of shell [10:35] and aspect ratio of the image [10:36] mmm [10:36] Cimi, if shell aspect ratio < image aspect ratio [10:36] if aspect ratio of phone is smaller than the image aspect ratio [10:36] yes... [10:37] :P [10:37] Cimi, sourceSize { height: }, otherwise sourceSize { width: } [10:37] Cimi, you'll get there [10:37] yes you type faster [10:37] Cimi, depending on the definition of aspect ratio :) [10:37] but yeah, you know the drill [10:37] aspect ratio is a ratio with/height [10:39] Saviq, you think the testImage might consume memory or not? [10:39] Cimi, it will, you need to unset the source [10:39] Saviq, dunno what qt does with visible: false [10:39] ok [10:57] Saviq, I need to use a ShaderEffect to render an Image component. The source for an image component is only a url though. How can I work around this? [10:57] The ShaderEffect lives in the theme atm. [10:57] The image is in the actual component class. [10:58] nic-doffay, ShaderEffect needs a ShaderEffectSource, and an Image can be the sourceItem of ShaderEffectSource's, if I get your question correctly [10:59] Saviq, nope. I'll ask the SDK guys about it. It's a bit trickier since it involves the theme stuff. [10:59] ta though [10:59] nic-doffay, k === greyback is now known as greyback|food === MacSlow is now known as MacSlow|lunch [11:18] Saviq, I think I am missing something [11:18] Cimi, hit me [11:19] Saviq, if (shell.width / shell.height <= testImage.sourceSize.width / testImage.sourceSize.height) { [11:19] backgroundImage.sourceSize = undefined [11:19] backgroundImage.sourceSize.height = shell.height } [11:19] Saviq, image is still blurred :-\ [11:20] seb128, gsettings is before qmenumodel in alphabetical order [11:20] Cimi, is it a JPEG image? [11:20] Saviq, yes [11:20] Cimi, I think we identified an issue with JPGs that the scaler in Qt is bad for them [11:20] Saviq, I expect the source size to behave correctly [11:21] Saviq, shall I try smooth false? [11:21] Cimi, can you try in a simple .qml for qmlscene? [11:21] Cimi, and see the impact [11:21] mm ok [11:26] Cimi, I think we saw that with MacSlow|lunch and Kaleo before in notifications - but never followed up [11:32] didrocks, didrocks, http://paste.ubuntu.com/5880522/ known? [11:33] mhr3: slangasek made a transition on uity-common, it should work when dist-upgrading though [11:33] mhr3: do you dist-upgrade and get that? [11:33] didrocks, i did dist-upgrade, and didn't even notice that it removed unity :/ [11:33] had to install it [11:33] mhr3: maybe check with slangasek? I have way too much on my plate right now to debug that [11:33] didrocks, ok [11:34] (merged 10 branches since this morning, 2 packages NEWing, a lot of jenkins debugging, a patch to cupstream2distro, lot of configuration changes…) [11:34] sounds like one of those "fun" days :) [11:34] and now unity-system-compositor is giving me hell [11:34] mhr3: well, I have too many "fun" days TBH… :/ [11:35] fun couple of months then ;) [11:35] Saviq, works here :-\ http://paste.ubuntu.com/5880531/ [11:37] Cimi, try to see what's the difference between that and unity8 [11:37] Cimi, smooth shouldn't be used for images, though [11:37] Saviq, there is no difference [11:38] Saviq, I copied code [11:38] Cimi, can you push to your branch please? [11:41] Saviq, lp:~unity-team/unity8/unity8.background_gsettings-sourceSize [11:41] (pushing...) [11:45] pushed [11:51] Saviq, ^ [11:51] Cimi, doh, you are right ;-) [11:52] Cimi, yeah, got it, trying to find out what's happening [11:52] Cimi, g/q they look the same :p [11:52] Saviq, it's weird isn't it? [11:53] Cimi, yeah, I removed most of the shell and it's still like that [11:54] mhr3, dpkg -l | grep libunity [11:54] Saviq, I'm checking if it's a bug on the sourcesize [11:55] Saviq, due to phone_background [11:55] Saviq, yes [11:55] Cimi, ok, so you're on it [11:55] Saviq, no idea though [11:55] Saviq, why it happens [11:55] Cimi, what's the solution then? [11:56] Saviq, property url default_background: "graphics/tablet_background.jpg" and the bug is gone [11:56] Saviq, if you can understand what could be the issue... [11:56] Cimi, how do I access the label colour for a theme again? [11:57] nic-doffay, which label? [11:57] Cimi, do the themes have a set colour for headers? [11:57] Or similar labels at the top of the component. [11:57] should be Theme.palette.selected.foregroundText iirc [11:57] Cimi, cool I'll try it now... [11:58] Cimi, where does that colour file live again? [11:58] So I can check the others (foreground text wasn't right for this) [11:58] nic-doffay, modules/themes/... [11:58] iirc [11:58] search inside here [11:59] Cimi, got it. [11:59] ta [12:00] Cimi, btw, you need to use Qt.binding(), so that sourceSize gets updated when shell size changes === alan_g is now known as alan_g|lunch [12:03] Cimi, sourceSize = undefined doesn't work [12:04] Cimi, ah [12:04] not sourceSize = undefined [12:04] Cimi, sourceSize.width/height = undefined [12:09] Wellark, but yeah, come here with that [12:09] Wellark, what's the problem? [12:09] Saviq: I have a meeting starting [12:09] Cimi, got it [12:10] Cimi, sourceSize.width/height = 0 [12:10] Cimi, not undefined or null [12:12] Cimi, http://pastebin.ubuntu.com/5880631/ fixes your issue [12:13] Cimi, it quietly ignores non-numeric values and keeps the original one === greyback|food is now known as greyback [12:16] Saviq: ok, the meeting got postponed [12:16] Saviq: so, application identifiers and application identity [12:17] Wellark, yes [12:17] greyback, ↑ you want to be here [12:17] * greyback is [12:17] Saviq: first off, some of our platform service like the hud-service needs to identify applications which push data to it [12:18] currently the hud-service expects the application to provide it's application ID which is then used to match the application windows with the data the application provides [12:19] yes [12:19] clearly there is a flaw here as the service should not simply trust the application to tell it's appid (can be faked), but instead the hud-service should be able to ask from *somewhere* "who is this application that is pushing me data" [12:20] Wellark, yeah, you can find the PID over DBus and then look through what the shell will expose to match the application [12:20] Saviq: indeed. I can get the PID [12:20] not sure how that'll work for multi-surface apps, but we'll manage [12:21] Saviq: I imagine the interface for this does not yet exist [12:21] Saviq: then this is also the same inside the shell [12:21] Wellark, it *almost* does === MacSlow|lunch is now known as MacSlow [12:21] when writing the launcher backend (which I currently am) [12:21] Wellark, greyback is working on it [12:21] Wellark: ApplicationManager can give you the application ID for a PID. Will need to add a dbus interface for it [12:21] greyback: is ApplicationManager a singleton inside the shell? [12:22] Wellark, yes [12:22] great. [12:22] even better [12:22] Wellark: yes [12:23] Wellark: so you need the app_id corresponding to a PID [12:23] greyback: indeed [12:23] greyback: and also the absolute path to the .desktop file [12:23] inside the shell application manager probably provides the desktop file contents directly === _salem is now known as salem_ [12:24] but the external services might need to read the .desktop file also [12:24] and now that we will have click application [12:24] there are actually multiple places the .desktop file may live [12:25] not just under /usr/share/applications/ [12:25] greyback: so, when doing the launcher backend (inside shell) I will utilize your api directly [12:28] Wellark, we need a common .desktop file reader, yes [12:29] Saviq: yes, that is a known fact. :) [12:29] Wellark, one that will respect XDG etc. [12:29] Saviq: yes, but click .desktop files are not respecting XDG [12:29] Wellark, why do you say that? [12:30] Saviq: https://lists.launchpad.net/ubuntu-appstore-developers/msg00253.html [12:30] Wellark, aren't the click hooks going to "install" .desktop files? [12:30] Saviq: "he core difference being [12:30] that it'll have relative paths that aren't resolvable using standard XDG [12:30] directory definitions (they'll all be in the package)." [12:31] Wellark, mhm [12:31] Saviq: so we need special handling with the icons [12:31] the file format follows the standard [12:31] Wellark, mhm, got it [12:32] Saviq: one options is simply convert the relative paths to absolute ones when parsing [12:32] Wellark, yeah +1 [12:32] Wellark, I don't want to have to care in the shell [12:32] Wellark, so the .desktop file reader, supporting click packages, will need to handle that [12:33] Saviq: indeed [12:33] sorry dropped connection [12:33] Saviq: but then again there is always the question how we handle the "themed" icons [12:33] Wellark, we should probably not call it a .desktop file reader anyway [12:33] so with this we have three options [12:33] Wellark, but "application data provider" or something [12:34] Icon=/absolut/path/icon.png [12:34] Wellark: http://studio.sketchpad.cc/EEd2PSjTRn <- is list of dbus APIs requested of ApplicationManager. I've added yours, does it look ok? [12:34] Icon=from-theme [12:34] icon=./click-relative [12:34] but sure. I need to run to a meeting now [12:35] Wellark, I imagine we'll need to add the app's dir to the theme search path [12:41] Saviq, one remaining bit [12:41] Saviq, setting the source of testImage to undefined [12:42] Saviq, if I do this I think I'll break the binding with the gsettings... [12:44] Cimi, yeah, you will, but that's ok - you don't need a binding, you need an on*Changed instead [12:50] Saviq, mmm hard [12:50] Saviq, I wanted to do onPictureUriChanged [12:50] but it doesn't work [12:50] Cimi, just proxy through an alias [12:50] Cimi, or through another property [12:50] ok [12:50] indeed [12:52] seb128, what am i looking for? [12:52] we have too many libunitites :) === alan_g|lunch is now known as alan_g [12:53] mhr3, libunity-core-*'s version [12:53] seb128, 6.0-7 is installed [12:53] mhr3, https://launchpad.net/ubuntu/+source/unity/7.0.2+13.10.20130705.1-0ubuntu3 [12:53] do you have the provides added there? [12:54] guys, can you please fix your barrier :P [12:55] kthxbye [12:55] seb128, 7.0.2+13.10.20130705.1-0ubuntu3 [12:55] seb128, the problem now is that the compiz schema got weird, it doesn't start the unity plugin [12:56] mhr3, ok, dunno about that [12:57] mhr3, I was looking at your upgrade pastebin earlier saying that unity-common was not installed [12:58] seb128, strangely it was a warning [12:58] oh, ok [13:00] Saviq, cannot set the url to undefined [13:01] Saviq, so how can I unload the image? [13:01] Cimi, "" [13:01] ok [13:05] Saviq, ok I finished the branch, it's ready for review... [13:05] Saviq, I have afternoon off, but I think I can come to the standup [13:05] Saviq, in any case, you know what I did :) [13:06] Cimi, yup [13:17] hmm.. upgrading with unity-next enabled was not that good idea..' [13:17] didrocks: any known breakage? [13:17] Wellark: what ppa exactly are you upgrading to? [13:18] unity-next doesn't exist :) [13:18] didrocks: :D [13:18] didrocks: let me check (once I get to the VT...) [13:19] yeah, either ubuntu-unity/daily-build-next (but as the description tells, you shouldn't use that one) [13:19] or ubuntu-unity/next (this one should be safe) [13:20] didrocks: ubuntu-unity-next [13:20] didrocks: not it works [13:20] it actually probably was not unity related at all [13:20] Wellark: this isn't a ppa :) I think you mean ubuntu-unity/next [13:20] didrocks: hmmm [13:21] might be that I hadn't updated in a while [13:21] didrocks: why didn't otto collect failing videos of autopilot tests? [13:21] reinstalling virtualbox drivers fixed it [13:21] sil2100: I guess jibel reenabled that with his patch, isn't it? ^ [13:21] didrocks: saw error on starting X that GLX driver had old ABI version [13:21] Wellark: you should just develop on your system with latest saucy ;) [13:21] 13 vs. 14 [13:22] sil2100, when running unity testsuite? [13:22] didrocks: not sure, the latest unity test runs don't have any videos [13:22] jibel: yes [13:22] sil2100, because I forgot to reenable them [13:22] doing now [13:23] didrocks: talking with mhr3 about that now [13:23] jibel: ok [13:23] Thanks [13:23] Saviq, I'll skip the standup, gotta run! [13:23] Cimi, k, have fun [13:23] Saviq, I'm on holiday tomorrow afternoon [13:24]