=== ari_ is now known as arsson === _salem is now known as salem_ === salem_ is now known as _salem === paulliu is now known as paulliu-lunch === marlinc is now known as marlinc|away [06:30] didrocks: ping [06:30] veebers: pong [06:32] didrocks: Good morning. Hey I was running some Unity tests in my Saucy VM to test some Autopilot changes. I appear to be getting some crashes with nautilus (and maybe compiz?). Have you seen the same recently? [06:32] veebers: there are some known nautilus crashes [06:32] veebers: but it's they should be fixed with latest gtk [06:32] didrocks: ah ok, thanks for confirming :-) [06:32] veebers: btw, can we remove the autopilot ppa? [06:33] veebers: as you daily release to distro, it's not needed anymore [06:33] veebers: also, fixing the autopilot tests failing would help to get the latest version to distro :) [06:33] didrocks: remove the actual ppa or remove the need for it in the job? [06:33] veebers: remove the need for it in the job [06:34] veebers: for the failures: http://10.97.0.1:8080/job/autopilot-saucy-daily_release/182/ (since yesterday) [06:35] didrocks: ugh, I hadn't realised they were failing :-P Yeah I'll make sure we're on it (sorry) [06:35] veebers: no worry, thanks! [06:36] didrocks: they weren't failing for me today (on my raring) I wonder if it's nautilus related. nvm, I'll look into it [06:37] didrocks: to confirm, you're asking me to remove the autopilot ppa adding step from the script used in the Unity AP job right? [06:38] veebers: hum? we don't use that script anymore for otto [06:38] veebers: just don't dput on the ppa [06:38] veebers: you should use saucy btw :p [06:39] didrocks: ah yeah sorry, out of touch with otto :-P [06:39] and very good point about Saucy, I'll be upgrading this weekend [06:42] veebers: so, I meant you don't need a staging ppa with latest trunk I guess [06:42] didrocks: ah, understood now [06:42] veebers: https://code.launchpad.net/~fginther/cupstream2distro-config/update-autopilot-qt-ppa/+merge/170708 [06:48] didrocks: ack === marlinc|away is now known as marlinc === mmrazik is now known as mmrazik|afk === mmrazik|afk is now known as mmrazik === mmrazik is now known as mmrazik|afk === ari_ is now known as arsson [08:40] veebers, thomi: hi! I noticed some failures in the autopilot autopilot tests [08:40] veebers, thomi: you guys know any reasons for those? [08:42] mhr3: around? [08:42] mhr3: I think we have a blocked autopilot run, (running for 4 hours, so still one hour of debugging available :)) [08:43] mhr3: ati machine [08:43] didrocks, k, will look into it, in a hangout now [08:43] sil2100: if its autopilot-qt its because qemu segfaults when using a QQuickView [08:43] sil2100: if you're talking about python-autopilot I don't know [08:45] mhr3: so, I confirm, just one hour from now :) [08:46] mzanetti: hm, I don't think it's the case here - I see a UInputError('"/dev/uinput" does not exist or is not a character device file - verify that the uinput module is loaded',) error [08:46] sil2100: where does that happen? [08:47] mzanetti: autopilot autopilot tests: [08:47] http://10.97.0.1:8080/job/autopilot-saucy-daily_release/182/ [08:47] sil2100: that means in a VM on desktop? [08:47] Yes [08:48] sil2100: hmm... so probably uinput module is not there indeed? also maybe check permissions. try adding a chmod 666 /dev/uinput before the test runs [08:54] mzanetti, any luck with the launcher stuff? [08:55] nic-doffay: not with the rotation... I think I need to strech the image because design wants that unrealistic perspective view on the tiles [08:56] mzanetti, could scale it from the centre point without a shader I guess. I think it might look strange with the entire tile scaling though... [08:57] nic-doffay: yeah... need to play around with... [08:59] Saviq, any chance to look at my branch yet. I've had no luck with it. [09:00] tsdgeos, mzanetti, can you help nic-doffay please? [09:00] Saviq: yes sir! [09:00] nic-doffay: what branch? [09:00] mzanetti: you take care of it? [09:00] mzanetti, one sec [09:00] tsdgeos: ack [09:00] greatz [09:01] mzanetti, lp:~nicolas-doffay/unity/messaging-modifications [09:01] Be sure to have demo assets too. [09:01] mzanetti, there was a "please" at the end of it ;) [09:01] Saviq: hehe [09:01] mzanetti, but thanks you took it like a man [09:01] Saviq, >_< [09:02] mzanetti, the main changes are in HeroMessage and IndicatorPage [09:02] A bug is resulting from the changes now which I can't pin point anywhere. I've browsed a lot of code in search of it. [09:03] Select an item in the messages, scroll down past it and deselect. You'll see it scrolls the list back up and selects the first item. This is the piece of code I cannot find perhaps it's part of the Ubuntu component itself? [09:03] mzanetti, ^ [09:04] nic-doffay: one step back... [09:04] what is this about? [09:04] what, suddenly my applications is emtpy on the lvwph branch :-/ [09:04] * tsdgeos tries to find out why [09:04] nic-doffay: the indicators messaging menu? [09:04] mzanetti, yeah. [09:05] nic-doffay: so I need a phone and generate some messages somehow to reproduce/see it? [09:05] mzanetti, yeah [09:05] nic-doffay: ok... how do you generate messages? [09:05] mzanetti, phablet-demo-setup will provide messages [09:07] mzanetti, it's a script from phablet-tools [09:07] * mzanetti didn't know about phablet-demo-setup [09:07] mzanetti, you don't demo much :) [09:07] indeed [09:08] I try to dogfood tho... and badly fail because my Nexus4 only has 3 hours battery lifetime with Ubuntu Touch :( [09:09] mzanetti, Galaxy Nexus battery life seems a hell of a lot better after the newest flash. [09:09] yeah... the Galaxy Nexus doesn't seem to suffer that bug where unity8 spins on 40% CPU all the time [09:09] its only the Nexus4 [09:11] Cimi: ping [09:15] nic-doffay: ok. preapared a device with messages [09:16] nic-doffay: now what? [09:16] mzanetti, go to the messages menu [09:16] Select say the second one. [09:16] Scroll down to the bottom. [09:16] Then press somewhere which should deselect the item. [09:17] Instead it scroll back to the top and focuses on the index at the top. [09:17] mzanetti, if you do this without scrolling down you'll see it deselects the item correctly then you're able to select another without any worry. [09:17] and now i can't repro it :-/ [09:18] nic-doffay: can't reproduce that [09:19] nic-doffay: opened the messaging menu, clicked on the "Otto Greenslade" message => it expands [09:20] nic-doffay: now scrolling down to bottom [09:20] mzanetti, are you running the branch? [09:20] nic-doffay: clicking on "Ellen Arnold" => that message expands === paulliu-lunch is now known as paulliu [09:20] nic-doffay: scrolling back up to beggining, the other message is collapsed as I'd expect it [09:21] nic-doffay: running a stock image flashed yesterday night [09:21] mzanetti, you need to run my branch. [09:21] That has the changes which cause this bug. [09:21] nic-doffay: ah... so you introduced that bug... I see. I thought you were supposed to fix an existing one [09:22] mzanetti, negative. I'm doing changes based on a design doc which weren't implemented yet. [09:22] mzanetti, with the flashed image you noticed it selects another item immediately, right? [09:22] nic-doffay: yeah [09:23] With my changes when you select another item it first deselects the selected item. Then when you press again it selects another. [09:23] This is the change which has introduced the bug. [09:23] uh... thats wanted by design? [09:23] mzanetti, although I can't find any code (after commenting out a lot of stuff and looking) which deals with a refocus on the first index. [09:23] mzanetti, yeah [09:23] :( [09:24] will cause people to tap on messages getting angry because they only expand sometimes [09:27] nic-doffay: ok. got it working now... had to merge trunk [09:27] nic-doffay: I see. while there is an expanded one, all the others are "inactive". that improves things I guess. [09:28] nic-doffay: anywys, I can reproduce the issue now. will come back to you in a sec [09:45] nic-doffay: so... [09:45] nic-doffay: in HeroMessage.qml [09:45] nic-doffay: in onClicked you call "if (!messageActivated) activateMenu()" [09:45] nic-doffay: thing is, menuActivated is always false [09:46] => you call activateMenu always which makes it scroll up [09:46] no sure why it actually collapses [09:47] mzanetti, hmm [09:51] nic-doffay: there is something badly wrong with this [09:53] nic-doffay: http://paste.ubuntu.com/5786370 [09:53] nic-doffay: I get onStateChanged that state changed to "expanded". in onClicked: I print the state and its "" [09:54] * mzanetti is puzzled+ [09:54] mzanetti, in HeroMessage [09:57] smspillaz, hey, do you have any idea why the limitation described there is in compiz: https://bugs.launchpad.net/ubuntu-gnome/+bug/1159430/comments/15 ? [09:57] Launchpad bug 1159430 in Ubuntu GNOME "Nautilus 3.7.92 breaks desktop background on Unity" [Low,Confirmed] [09:58] mzanetti, trying to get my branch to run again today, just running into some probs, going to have a look myself. [09:58] nic-doffay: if you merged your branch with trunk you need a freshly flashed device [10:02] mzanetti, doing it now. [10:02] seb128: not sure off the top of my head, I can have a brief look into it [10:03] seb128: does this version of nautilus use an argb window ? [10:03] smspillaz, yes [10:04] smspillaz, https://bugzilla.gnome.org/show_bug.cgi?id=691359 [10:04] Gnome bug 691359 in Desktop "adapt to gnome-shell rendering the desktop" [Normal,Resolved: fixed] [10:04] smspillaz, they changed nautilus to display icons on a rgba win [10:04] smspillaz, and made gnome-shell draw the background [10:04] ah okay, cool [10:04] but things don't work well with compiz :/ [10:04] indeed, we assume all desktop windows are rgb [10:05] it's blocking the rgba win [10:05] if the limitation is there I guess there is a reason? [10:06] yes - if those pixels are not touched by a window they remain on the backbuffer between paints [10:06] seb128: so my understanding is that the desired path for handling this is to have compiz draw the wallpaper, yes ? [10:07] smspillaz, correct [10:07] well, first idea was to go back to let g-s-d do it [10:07] but having compiz/unity do it in the long run would be goo [10:07] good [10:08] should be relatively straightforward [10:08] it already has code to draw wallpapers [10:08] I tried the wallpaper plugin [10:08] but that draws on top of the nautilus icons [10:10] seb128: the wallpaper plugin creates a fake desktop window if it thinks that there isn't one [10:10] as long as nautilus creates a window that has the hint _NET_WM_TYPE_DESKTOP set in its _NET_WM_WINDOW_TYPE it should work [10:10] I'm just checking to see if core ignores desktop windows that have an alpha channel [10:11] otherwise I just need to fix the wallpaper plugin to draw the background at the right stage, of which it may not be doing at the moment [10:13] Saviq: ping [10:13] seb128: is it possible for me to build this version of nautilus directly from git ? [10:13] dednick, pong [10:13] Saviq: http://pastebin.ubuntu.com/5786411/ [10:13] getting some weird dep error [10:15] smspillaz, sure, it's nautilus 3.8 [10:15] smspillaz, you have it in there if you prefer a ppa: https://launchpad.net/~gnome3-team/+archive/gnome3/+packages?field.name_filter=nautilus [10:15] seb128: ah cool, I'll just dist-upgrade then [10:16] seb128: this one will use the argb window right ? [10:16] dednick, remove libhud-client1-dev [10:16] smspillaz, yes [10:16] dednick, and `apt-cache policy libhud-client1-dev` to find out if you have it in a ppa somewhere [10:17] and remove that ppa [10:19] be back in like 30 minutes [10:20] smspillaz, can you drop the rgb limitation as well? even if we use compiz's wallpaper plugin, I would like g-s-d rendering to work as well (especially as a fallback since it has been proved hard to add compiz plugins to config on upgrade) [10:21] Saviq: ta. seems to have resolved. [10:22] didrocks, guess i'm late already/ [10:22] ? === jamesh_ is now known as jamesh [10:47] seb128: I don't think the rgb limitation is related to what you want to do [10:48] Saviq, what's changed with build -s lately? [10:48] seb128: If you want to use gnome-settings-daemon to set the wallpaper (and I assume that it uses XSetWindowBackground on the root window) then all we need to do is make it so that an argb desktop window is not detected as a desktop window [10:49] nic-doffay, should work the same [10:49] nic-doffay, you can `rm -R ../unity_build` [10:49] that way you'll get nautilus drawing the icons and compiz will use the root window wallpaper [10:49] nic-doffay, and a ./build --setup; ./build --clean is needed once [10:49] nic-doffay, then, ./build is enough [10:49] but it should also take much shorter to ./build -s [10:50] seb128: in any event, adding new compiz plugins is not a particularly difficult exercise, just drop a settings upgrade into /usr/share/compizconfig-1/upgrades [10:53] seb128: a word of warning: if you want this to work both *with* and *without* the wallpaper plugin I'll need to break the ABI [11:00] Saviq, when I try run_on_device ./run: 71: ./run: ./builddir/unity8: not found [11:04] smspillaz, "then all we need to do is make it so that an argb desktop window is not detected as a desktop window" seems like what we want then [11:04] smspillaz, did you seem adam-yorba's patch on the bug I pointed before? [11:04] he did a small hack to compiz that fixed it for him === mmrazik|afk is now known as mmrazik [11:05] smspillaz, I would prefer use g-s-d rather than compiz's plugin, but that mostly because I'm being cautious and going for a codebase we have used for years and I know is doing what we want [11:05] Saviq, I think I'm having packaging issues again for some reason. [11:06] Saviq, qmenumodel isn't found. [11:09] 1 [11:09] oops [11:09] seb128: okay, we'll go with the g-s-d approach [11:09] smspillaz, thanks [11:10] mzanetti: hi! [11:10] seb128: I did see adam's patch - its not the right approach. Its just a workaround to make the wallpaper-plugin created desktop window fully trasparent so that nautilus isn't marked as occluded during the paint pass [11:10] mzanetti: https://code.launchpad.net/~sil2100/indicators-client/fix_dependencies/+merge/170785 <- does this make sense to you? [11:10] (though say I said thanks for the effort :)) [11:10] smspillaz, k [11:11] dednick: can you have a look at sil2100's MR ^^ [11:11] sil2100: I think indicators-client is not used any more [11:11] sil2100: but dednick should know for sure whats going in [11:11] s/in/on/ [11:11] seb128: how do I change the background using g-s-d ? [11:12] smspillaz, you need a patched g-s-d [11:12] sil2100: afaik indicators-client is integrated into the shell now. not sure if that work is already completed tho [11:12] seb128: do you know if its just the same as what xsetroot does ? [11:13] smspillaz, if you want to patch gsd: http://paste.ubuntu.com/5786515/ [11:13] mzanetti, sil2100: looks fine. At the moment, it's still used. [11:13] mzanetti: well, it's still not merged in, we talked yesterday about that with dednick [11:13] basically it has a check "don't duplicate nautilus work, so don't draw if nautilus does" [11:13] mzanetti: so it's used [11:13] dednick, mzanetti: can anyone approve? [11:14] dednick: I'm not up-to-date with this stuff. can you please take care of it ^? [11:14] sil2100: although they're more "recommended", they're not dependant. [11:15] but it wont work without. so... [11:15] sil2100, mzanetti: approved [11:16] \o/ Thanks! [11:18] smspillaz, not sure what is used, it seems to go through libgnome-desktop and cairo [11:19] seb128: ok [11:19] seb128: one last question [11:19] seb128: how do I get this version of nautilus to actually ... draw icons [11:19] (on the argb desktop window) [11:21] smspillaz, it should do by default [11:21] smspillaz, gsettings get org.gnome.desktop.background show-desktop-icons ? [11:21] that should be true [11:21] nic-doffay, hmm, qmenumodel-qml is in distro [11:22] nic-doffay, `apt-cache policy qmenumodel-qml` [11:23] Saviq, Installed: 0.2.6daily13.06.07-0ubuntu1 [11:23] nic-doffay, same here [11:24] nic-doffay, what's your log output, then/ [11:24] ? [11:25] Saviq, https://pastebin.canonical.com/93221/ [11:27] nic-doffay, libqmenumodel-dev should help [11:27] nic-doffay, checking why this isn't in our build deps [11:28] nic-doffay, ah, you're working off of dednick's branch? [11:29] nic-doffay, you might need to merge his branch, libqmenumodel-dev is in build deps there [11:29] Saviq, yeah [11:29] nic-doffay, and ./build should install it [11:29] Saviq, kk [11:29] nic-doffay, /me checks, too [11:30] nic-doffay, ah, remove your builddir [11:30] nic-doffay, and ./build should work again (if you have the latest from dednick's branch) === MacSlow is now known as MacSlow|lunch [11:32] nic-doffay, aah, actualy [11:33] nic-doffay, dednick, please merge trunk in your indicators-client branch [11:33] nic-doffay, as there's been changes to the build scripts that you could benefit from [11:34] nic-doffay, in the mean time `mk-build-deps -i -s sudo` [11:38] Saviq: doing now [11:40] dednick, let me know when it's pushed... [11:40] nic-doffay: pushed === greyback is now known as greyback|lunch [11:42] Saviq, i had some trouble with running unity8 on the desktop yesterday, first it complained that Panel is not known from DashBar.qml, so i added a import "../Panel", that helped, but then it got stuck in a dbus call coming from nm-glib cause it complained that it already has the name that got requested (cause the panel was instantiated twice?), so i just commented out the Panel from DashBar.qml completely and then it worked [11:42] Saviq, and now, what's the proper solution to these issues? :) [11:42] mhr3_, is that lp:unity/8.0? [11:42] Saviq, yep [11:42] mhr3_, that shouldn't happen there [11:43] Saviq, maybe i was missing a dep? [11:43] mhr3_, ah wait [11:43] i'm still on R [11:43] mhr3_, wha!? [11:43] mhr3_, ppa:ubuntu-sdk-team [11:43] let me check if i have that [11:43] mhr3_, ./build should add it [11:44] Saviq, i don't like ./build :P [11:45] yea, got disabled when i upgraded to R, will try if it helps [11:45] mhr3_, should, it wants "Panel" from the SDK, not from unity8 [11:45] mhr3_, and you instantiated Panel twice [11:46] Saviq, aaah, now it makes sense :) [11:55] compiz/unity does weird corruption things to my qt based apps [11:55] in saucy [11:55] like the selection colors get garbage and stuff [11:55] tsdgeos, compiz/unity hates Qt :P [11:55] anyone aware of that? [11:55] tsdgeos, we're fixing that with unity8, thouhg :D [11:56] *cough* *cough* [11:56] though [11:56] :D [11:57] or it may be the qtgtkstyle [11:57] which actually it'd make some more sense [11:57] tsdgeos, oh, about that, it got fixed for me [11:57] than compiz/unity being the ones causing it [11:57] tsdgeos, until tomorrow or so Qt apps were ugly (tk style) [11:57] no, i have the same style i ha before [11:57] but some of the "paintings" go crazy at times [11:57] tsdgeos, and it fixed after an update [11:58] tsdgeos, yeah, but might be related [11:58] awesome, taking a snapshot made it go away :D [11:58] some sort of refresh or something :D [12:02] my shell is constantly segfaulting in libhud-client2 [12:02] anything known already? [12:03] mzanetti, as in unity8? [12:03] Saviq: yes [12:03] mzanetti, no, not known [12:17] seb128: thanks, just building and testing a quick fix now [12:17] smspillaz, great, thanks for working on that ;-) === greyback|lunch is now known as greyback [12:22] ugh! I can't type diacritics in unity7 dash [12:23] compose works, but not AltGr === greyback|lunch is now known as greyback [12:33] Saviq: altgr is a gtk thing, dash is not gtk, is it? [12:33] lunch! === om26er_ is now known as om26er [12:43] tsdgeos, what do you mean a gtk thing?! it's an "everywhere" thing [12:43] tsdgeos, it's just a pl layout - I write łóśąćżźńę [12:44] with altgr [12:44] breakfast! [12:46] Indeed, that doesn't work anymore [12:46] That's clearly a bug! === MacSlow|lunch is now known as MacSlow [12:56] tedg: hi! [12:56] * tedg hides [12:56] ! [12:56] ;) [12:56] I feel bad now! [12:57] Yesterday I update manager upgraded some packages. But my desktop froze and I couldn't switch tty so I forced shutdown'd. The window manager now loads but unity does not. [12:57] tedg: not sure if you'll know how to help here, but we're getting an unit test failure for libdbusmenu for armhf - https://launchpad.net/~ubuntu-unity/+archive/daily-build/+build/4733356 [12:57] I've ran synaptic but there's no mention of any botched packages so presumbly everything upgraded correcltly. [12:58] It started happening after fix LP: #1154701 I suppose? It landed yesterday [12:58] Launchpad bug 1154701 in libdbusmenu (Ubuntu) "dbusmenuitem retains references/callbacks to a widget after the widget is removed" [Undecided,New] https://launchpad.net/bugs/1154701 [12:58] sil2100, Yeah, that'd be my guess. [12:58] sil2100, Do we have any output? [12:58] Seems like something is swallowing all the test ouput. [12:59] pinging charles so he comes and visits when he gets in :-) [12:59] tedg: not sure how we can get any, since it's LP [13:00] Yeah, curious why Jenkins isn't grabbing this as well. [13:00] Hmm, seems it's only on amd64 [13:01] mmrazik: thanks! [13:04] in ccsm it appears unity is un-checked. Checking it just unloads the window manager and doesn't bring unity back [13:10] sil2100, https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1193360 [13:10] Launchpad bug 1193360 in unity (Ubuntu) "Can't type diacritics with AltGr+letter in Dash or HUD" [Undecided,New] [13:10] Magicarp: "unity --reset" help at all? that resets unity's config, in case something confuses it [13:10] Thanks! [13:11] tedg: yep, it's strange... [13:11] greyback, "Reset is not supported anymore. Deprecated option" [13:12] Huh, for some reason arm is turned off for dbusmenu.... [13:13] alesage, In the indicator stack there seems to be a few projects that have ARM builds turned off. Were things failing? Or, why is that? [13:16] Saviq: can you do that altgr thing in Qt apps? [13:16] doh [13:16] ignore me [13:16] i thought you meant windows-like things [13:16] like alt+154 [13:16] tsdgeos, nah [13:16] that afair works in gtk [13:16] tsdgeos, that I can do with Compose [13:17] tsdgeos, but if you type AltGr+w [13:17] you should get ł [13:17] sure [13:17] ł [13:17] at least english layouts do [13:17] yeah [13:17] it's how i get the arrows :D [13:17] yeah exactly [13:17] ←↓→↑ [13:17] you can't do them in dash [13:17] right [13:19] dandrader: ah news to me. Another old reliable gone :( [13:32] come on mumble [13:32] Cimi, standup [13:33] greyback, we'll start [13:34] seb128: working great here now [13:35] smspillaz, \o/ [13:35] seb128: btw, maybe it might be worth considering having compiz handle the wallpaper - you can implement stuff like per workspace wallpapers with it [13:35] seb128: gets you closer to the design spec anwyays [13:35] gah [13:35] Saviq: go ahead [13:36] smspillaz, right, one step at the time ... let's give the compiz plugin some testing to see how it behaves and it supports all our option (like solor color backgrounds) [13:36] smspillaz, does it read the GNOME gsettings background key? [13:37] smspillaz, if it works fine I'm happy to switch [13:37] smspillaz, but when I tried it was not even picking my selected wallpaper, I had to pick one through ccsm [13:38] seb128: there'd need to be some kind of integration story [13:38] but still: http://i.imgur.com/WZYOs1v.jpg [13:39] smspillaz, yeah, as said I'm fine with using it ... once it's integrated with our config tools [13:39] in any case, I have it working with and without the wallpaper plugin [13:39] I probably won't have time to implement the integration story, but its up for grabs if someone is interested [13:40] the only caveat is that if you want to use transparent cube or curved expo you need to use the wallpaper plugin as the default renderer in compiz will draw the wallpaper in a way that doesn't allow for transformations [13:41] tedg: hm, any luck? :) [13:41] smspillaz, ok, noted, maybe MCR wants to pick the integration work ;-) [13:42] smspillaz, do you have a mr with your compiz changes? [13:42] sil2100, I think we're going to have to kick it in on Jenkins, I wanted to see if there was a reason alesage had it disabled before turning it on. [13:42] sil2100, It's still a bit early for him, probably in a little bit. [13:46] dednick, heyo! let's talk plugins [13:46] mterry: ok sure. [13:46] dednick, so what were you saying that the mock plugin did wrong? Exported the wrong thing? [13:47] any taker for reviewing this one? https://code.launchpad.net/~dandrader/unity/8_dragHandle/+merge/170172 [13:47] mterry: so, the UserModel in plugins/LightDM creates a QLightDM::UsersModel, which is defined in tests/mocks/LightDM [13:48] dednick, sure yup [13:48] mterry: and mocks/LightDM creates UserModel private data (which contains the mock into) from tests/mocks/LightDM/[demo|full|single|etc]. [13:49] * mterry nods [13:50] but mocks/LightDM is also linked to MockLightDM-demo [13:50] so it always uses the demo profile. [13:51] dednick, plugins/LightDM is statically linked to demo/ [13:51] dednick, mocks/LightDM is dynamically linked to liblightdm-qt5-2.so, wherever it finds it [13:51] dednick, that's where LD_LIBRARY_PATH comes in [13:51] dednick, the expectation is you point to the backend data you want with LD_LIBRARY_PATH [13:52] dednick, you can see this happening in tests/qmltests/CMakeLists.txt [13:52] each test points to its desired backend [13:52] mterry: i understand the intent, but i can't seem to get it working [13:52] mhr3_: ping [13:53] dednick, if you ldd on the mocks/LightDM plugin, is there a line that points at liblightdm-qt5-2 [13:53] ? [13:54] greyback, pong [13:54] mterry: yes. not found. [13:54] dednick, good, that's expected [13:54] mhr3_: hey, when application is started, zeitgeist is notified somewhere. I guess there's code in gtk that sends that notification. Any idea where I'd find it? [13:55] dednick, so how are you running unity that you end up with the demo one? (sounds like you're running with the plugins/LightDM plugin rather than the tests/mocks/LightDM one) === _salem is now known as salem_ [13:55] greyback, it's deep inside gio [13:55] in g_app_info_launch iirc [13:55]