=== kdub is now known as kdub^holiday === duflu_ is now known as duflu === larsu is now known as crowd === crowd is now known as larsu [09:16] Mirv: any chance you can add https://bugs.launchpad.net/ubuntu/+source/qtbase-opensource-src/+bug/1253120 to the Qt 5.2 beta2/RC packages so we get egl again on the desktop? [09:16] Ubuntu bug 1253120 in qtbase-opensource-src (Ubuntu) "Enable build of egl on desktop gl in Qt 5.2" [Undecided,New] [09:24] tsdgeos: right, I tried and it didn't apply for some reason, I mean the configure sections looked different. I was thinking to wait RC, but then it looks the RC isn't similar to what you submitted either. [09:24] hmmm ?¿ [09:25] that's for the stable branch [09:25] how can it have diverged that much [09:25] tsdgeos: yeah, I know. it's not much but the if logic is different so I was a bit cautious. [09:26] tsdgeos: http://pastebin.ubuntu.com/6488707/ - there's that top if [ "$CFG_OPENGL" != "es2" ]; then [09:27] but I guess it would be maybe removing similarly the "$CFG_OPENGL" = "no" and elif [09:28] by the way, still no real RC http://download.qt-project.org/development_releases/qt/5.2/ [09:29] yeah... [09:31] Mirv: yeah same thing applies [09:35] tsdgeos, that doesn't seem to have worked completely https://jenkins.qa.ubuntu.com/job/ubuntu-ui-toolkit-trusty-amd64-ci/295/console [09:35] Saviq: yeah [09:36] Saviq: didn't realize i was just running the test manually [09:36] not thorugh the make check thing [09:36] that passes -minimal [09:36] tsdgeos, yeah, looks like you need to swallow the -platform arg [09:36] or make it not be passed by the make [09:36] tsdgeos, or well, QT_QPA_PLATFORM=minimal instead would work, too [09:36] yeah [09:37] or swallow it :D [09:38] tsdgeos, yeah, it's probably a common set of args? [09:39] all the other ones work fine [09:39] i.e. -maxwarnings 10 -o ../../test_tst_arguments.xml,xunitxml -o -,txt [09:39] yup [10:26] mhr3: there? [10:26] tsdgeos, yep [10:26] mhr3: i don't understand how https://code.launchpad.net/~mhr3/unity8/ensure-currentitem-non-null/+merge/196948 makes set_current_index not get lost [10:27] tsdgeos, cause it doesn't use setCurrentScopeAtIndex [10:27] that is what invalidates it [10:28] or well... resets it [10:29] mhr3: but why do you need to set it at all? [10:30] when it gets loaded it'll be set to the correct value anyway [10:30] or [10:30] in case we don't have one [10:30] right [10:30] and you answered yourself :) [10:31] mhr3: i think the code would be a little easier to read if you had an outer if (count > 0) [10:31] but that may as well just be my brain [10:32] so this may actually be the proper fix to the dash content test unstability we have [10:32] and not cimi's 5 to 10 seconds [10:32] i was thinking about it [10:32] but meh, it's 4 loc :P [10:33] mhr3: what i'm not so sure is about how you [10:33] dashContentList.currentIndex = -1; in the test [10:33] should this be a verify and have that code in the onCountChanged? [10:33] tsdgeos, not visible in the diff, what the -1 is in cleanup() [10:33] s/what/that/ [10:34] sure [10:34] i knew it was there [10:34] still [10:34] it means that dashContentList it's not cleaning up itself much [10:34] so will the code you added in DashContent.qml work in real life [10:34] without the extra cleaning you're doing in there? [10:34] tsdgeos, listview itself is very odd how it behaves there [10:35] listview itself is very odd [10:35] yep :D [10:35] tsdgeos, basically i added it so the test_current_index would actually pass [10:35] cause if it wasn't run as the first thing it'd fail on the currentIndex == -1 [10:36] yet if it was run as the first test it passed [10:36] right [10:36] and that's why i am saying that it makes sense to have [10:36] dashContentList.currentIndex = -1; [10:36] as else [10:36] in that count > 0 of your countChanged [10:37] i'm not sure, what if you want to reset the model? [10:37] ie you'll see count 7 -> 0 -> 7 [10:37] if you don't force currentIndex to -1 you'll actually keep the same active page open [10:39] if you to 7 -> 0 -> 6 -> 7 [10:39] you won't [10:39] right, but at least it'll be approximately there :) [10:40] and that's probably what will happen anyway 7 -> 0 -> 1 -> 2 -> 3 ... [10:40] well if we don't forcefully set it to -1 we have the option of keeping the same page active [10:40] if we do, we loose it [10:40] end of story :) [10:41] who has bugs for me? [10:41] sure [10:41] mhr3: i don't see how that counters the fact that you had to add extra code in cleanup to make your test work [10:43] tsdgeos, tests should start in the same base environment, the component itself can remember state if it chooses to [10:43] putting the reset in the component seems wrong [10:44] well you're testing it [10:44] cause you're closing the doors on keeping the state [10:44] you test says [10:44] compare(dashContentList.count, 0, "DashContent should have 0 items when it starts"); [10:44] compare(dashContentList.currentIndex, -1, "DashContent's currentIndex should be -1 while there are no items in the model"); [10:44] but that is simply not true [10:44] you're totally arguing that the component is allowed to have dashContentList.currentIndex as not -1 even if there are no items in the model [10:45] it is if all the tests were isolated [10:45] so either your test is wrong [10:45] or your component is wrong [10:45] no [10:45] you specifically wrote [10:45] "DashContent's currentIndex should be -1 while there are no items in the model" [10:45] no [10:45] "DashContent's currentIndex should be -1 while there are no items in the model at the beggining of the program but not after because i think it makes sense to keep it" [10:45] right? [10:45] then i'll change it to "while there have been no items..." ;) [10:46] mhr3: well, do it [10:46] at least it won't be lying [10:47] tsdgeos, fine with me, do you want the count if split too? [10:47] mhr3: i'll leave it to you [10:53] pushed [10:54] tx ;.) [10:56] btw tryCompare() really silly name [10:57] when it's actually waitUntil() [10:57] waitUntil with a timeout [10:58] waitUntilWithTimeout() would be silly too :) [10:59] mhr3: actually... [11:00] mhr3: maybe we can have two tests? [11:00] have one do dashContentList.currentIndex = -1; at the beginning (as if coming from a clean startup) [11:00] and the other do dashContentList.currentIndex = 900; at the beginning (as if coming from a weird model) [11:00] and then on both you test that at the end dashContentList.currentIndex == 0 or something? [11:02] whatcha think? [11:04] tsdgeos, well, that's kindof already tested in the set_current_scope_index_async [11:04] but not directly, so ok... [11:07] tx! === MacSlow is now known as MacSlow|lunch === _salem is now known as salem_ [11:42] Saviq: ci still borked, right? [11:44] Saviq: can you tell me what to do in order to run the real application manager instead of the mock with run_on_device? I just can't seem to find the right place [11:50] tsdgeos, Q_SCRIPTABLE? Can't seem to find any info on it. (In dbusscreen.h) [11:50] makes it dbus friendly amonsgt other things [11:51] like making it also a Q_INVOKABLE [11:58] tsdgeos, yeah, we're fighting, should be good soon [11:58] mzanetti, right! that's actually a bug [11:59] mzanetti, it was working before [11:59] mzanetti, somehow QML2_IMPORT_PATH is borked [11:59] mzanetti, unity-mir installs it in /usr/lib/arm-linux-gnueabihf/qt5/imports/Unity-Mir/ [11:59] mzanetti, and main() adds it [12:00] mzanetti, main.cpp:155 [12:00] * Cimi flashing nexus 10 [12:00] mzanetti, and somehow that started to fail [12:04] mzanetti, most likely the set of import paths is constructed in the wrong order [12:04] mhm... I was playing around with that. but it always fails to find it [12:07] tsdgeos: Q_SCRIPTABLE doesnt seem to mean anything. src/corelib/kernel/qobjectdefs.h defines it as just "#define Q_SCRIPTABLE" [12:08] sure [12:08] it's just moc and others that make magic happen [12:08] ah [12:08] it's like emit [12:08] well not really [12:08] yep [12:08] emit is really nothing [12:08] noone cares [12:17] btw i lost the user indicator and the volume indicator in unity7 [12:18] oh [12:18] it's unity8 playing up with them [12:18] damnit [12:18] run unity8 [12:18] have sound and wifi indicators [12:18] kill it [12:18] indicators gone [12:18] :D [12:38] greyback, landings are b0rked until mir gets published, so no point in top-approving unless you kick generic-land by hand [12:38] greyback, which I'll do in a minute [12:39] Saviq: ah really, wasn't aware of that [12:39] greyback, bug #1255578 [12:39] bug 1255578 in Ubuntu CI Services "dependency issues on libunity-mir1 in testrunner-otto" [Undecided,Confirmed] https://launchpad.net/bugs/1255578 [12:40] greyback, it got fixed long-term by removing daily-build from the jobs [12:40] but we're still deadlocked until Mir's published [12:40] and then there's bug #1255948 which means we can't even use daily-build for a one-off job ;? [12:40] bug 1255948 in Ubuntu CI Services "upstream merger hooks do not propagate to downstream jobs' builder_hooks" [Undecided,New] https://launchpad.net/bugs/1255948 [12:40] so will push manually [12:41] Saviq: I am as confused as you though. How did it happen? [12:41] greyback, daily-build was enabled by default for unity-mir -ci and -autolanding [12:42] greyback, so it built, got put into the local per-stack repository for unity8 stack [12:42] greyback, but unity8 isn't using daily-build (and shouldn't) [12:42] nothing should [12:42] so then you couldn't install libunity-mir1 from the mbs repo [12:42] Saviq: finally :) https://code.launchpad.net/~mzanetti/unity8/fix-appmanager-on-device/+merge/197053 [12:42] 'cause it wanted new mir [12:43] mzanetti, that should be there from /etc/environment [12:43] mzanetti, I mean imports [12:43] Saviq: yeah... it is when I query it in the shell. somehow exec_with_ssh doesn't pick that stuff up [12:43] Saviq: what's wrong with daily-build? [12:43] mzanetti, also, will conflict https://code.launchpad.net/~saviq/unity8/tweak-run-scripts/+merge/196871 [12:44] greyback, it's unsafe - stuff in there might never end up in distro [12:44] Saviq: ah ok, perfect. I'll drop the conflicting change [12:44] mzanetti, there's also a possible fix to do in exec_with_ssh then [12:44] mzanetti, instead of what you did [12:45] mzanetti, does exec_with_ssh use sudo -u phablet -i? [12:45] mzanetti, that's the way to pick up the correct environment [12:45] mzanetti, let's fix it at its root [12:45] Saviq: but when I just ssh into the phone it's there too [12:45] mzanetti, yeah, that's because it's interactive [12:45] Saviq: understood. Pity though [12:45] mzanetti, i.e. bash [12:46] greyback, why? === MacSlow|lunch is now known as MacSlow [12:46] Saviq: handy if you wanted to live on cutting edge :) [12:46] greyback, you can still use it [12:46] greyback, but -ci nor -autolanding should live on cutting edge, should they? [12:47] * Saviq always gets confused with the rules around neither / nor / or / not [12:47] was that sentence syntactically correct ↑↑? [12:48] Saviq: neither/nor usually [12:48] greyback, yeah, but if there's "not" somewhere already, neither/nor is wrong to use, right? [12:48] greyback, i.e. something should not do foo, neither should bar [12:49] Saviq: right, then you'd use "nor" [12:49] yeah, that's why I said I'm confused ;) [12:49] it all makes perfect sense :P === dandrader is now known as dandrader|afk [12:53] i think english should have xor as well then [12:53] example "you either agree or not" -> "you agree xor not" [12:53] Saviq: this does look correct to me: ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -t phablet@localhost -p 2222 sudo -u phablet -i bash -ic \"echo $QML2_IMPORT_PATH\" [12:53] guys, a little more killing https://code.launchpad.net/~aacid/unity8/remove_unused_applications_folder/+merge/197058 [12:54] and that said [12:54] lunch! [12:55] mzanetti, yeah, does this work with exec_with_ssh? [12:56] mzanetti, if not - we should find out why - as we're bound to actually lose more env then just import path [12:56] Saviq: yeah... it doesn't work [12:56] strangely [13:02] mzanetti, exec_with_ssh 'echo ${QML2_IMPORT_PATH}' [13:02] mzanetti, that drops /usr/lib/arm-linux-gnueabihf/qt5/imports for me [13:03] not here [13:03] mzanetti, TBH I think my branch should fix the issue [13:03] mzanetti, the pkg-config was the cause [13:03] mzanetti, as it was overwriting the path [13:04] * Saviq tries === dandrader|afk is now known as dandrader [13:09] Saviq: nope... tried with your branch... not in there [13:11] mzanetti, yeah, I mean I'm trying run_on_device for real with my branch [13:11] mzanetti, AFAICT it should come up with real apps plugin [13:11] no. it doesnt [13:11] 200MB deps... [13:11] that's what I just did [13:11] mzanetti, ok, but yours does not echo the import path to start with ;) [13:12] mzanetti, just let me try [13:13] ok. grabbing something to eat [13:14] mzanetti, ok, not working indeed [13:14] * Saviq prints the import path [13:14] s === alan_g is now known as alan_g|lunch [13:31] Saviq: what's the verdict? [13:31] mzanetti, something's stupid [13:31] told ya :D [13:32] no clue why it doesn't pick up the env [13:32] probably some very restrictive check in some of the env scripts [13:32] mzanetti, but really, exec_with_ssh "\${QML2_IMPORT_PATH}" doesn't work? [13:32] mzanetti, it's completely stupid [13:32] mzanetti, in: [13:32] * mzanetti tries again [13:32] exec_with_ssh "cd $CODE_DIR/ && ./run $ARGS -- $RUN_OPTIONS" [13:33] I can add '&& echo \${QML2_IMPORT_PATH}' [13:33] and it's there [13:33] Saviq: right... forgot to escape the $ [13:34] then it's there [13:34] mzanetti, right, what I though [13:34] t [13:35] mzanetti, it's the ./run that doesn't get it somehow [13:37] wtf ;? [13:37] +1 [13:44] Saviq: well, it loads the environment but doesn't export it [13:45] mzanetti, right [13:51] mzanetti, thing is the whole sudo/bash atrocity should not be needed [13:51] mzanetti, if I go [13:51] Saviq: that's what I would have said too. but ubuntu is behaving a bit different than the other distros I know [13:52] ssh phablet@mako 'cd ~phablet/shell; ./run' [13:52] so I wasn't entirely sure [13:52] that works fine [13:52] mzanetti, but then if I make exec_with_ssh like that, it fails for some reason :/ [13:54] ah wait [13:54] grrr no, without bash we don't have the whole env from upstart [13:55] Saviq: I guess ogra would be the right one to help here [13:55] mzanetti, yeah, was about to go to him, wanted to try a few things [13:57] Mirv: RC packages \o/ [14:07] tsdgeos: \o/ [14:21] Mirv: can we close this one down since we're not aiming at 5.1.1 anymore? or do these still fail under 5.2?¿ [14:35] tsdgeos, could you review my MP then? [14:35] tsdgeos: cool, can you link me to the gerrit page of that? [14:35] dandrader: shoot the url [14:36] tsdgeos, https://code.launchpad.net/~dandrader/unity8/runningAppsEndClose/+merge/196257 [14:36] greyback: https://bugreports.qt-project.org/browse/QTBUG-30493 [14:36] Cimi: what do you think about this? https://bugs.launchpad.net/ubuntu-ux/+bug/1255926 [14:36] Ubuntu bug 1255926 in Unity 8 "If there is only one category in a scope it should not use the carousel" [Undecided,Incomplete] [14:37] I know I'll miss those lazy days next week :))) [14:37] Cimi: yeah... thought the same :D [14:37] mzanetti, ok I'll look at it [14:38] tsdgeos: no I meant the gles & opengl patch [14:38] greyback: silly me i copied pasted from the wrong tab [14:38] greyback: https://codereview.qt-project.org/#q,71717,n,z [14:38] tsdgeos: perfect, thanks === alan_g|lunch is now known as alan_g [14:40] Cimi: https://bugs.launchpad.net/unity8/+bug/1255926/comments/2 [14:40] Ubuntu bug 1255926 in Unity 8 "If there is only one category in a scope it should not use the carousel" [Undecided,Incomplete] [14:44] Is the run_on_device script meant to work for the N10 at the moment? [14:44] mzanetti, pushed the . /etc/environment to https://code.launchpad.net/~saviq/unity8/tweak-run-scripts/+merge/196871 === dandrader is now known as dandrader|lunch [14:47] Saviq: me... if I print the engine's importPathList it's still not there [14:47] +h :D [14:47] mzanetti, yeah, I'll track that down - at least it's there in the run script ;) [14:47] indeed [14:48] mzanetti, prolly not exported [14:48] yeah... same issue, one level down :D [14:48] mzanetti, yup [14:49] mzanetti, pushed [14:50] Saviq: \o/ [14:51] mzanetti, as for the whole thing - http://paste.ubuntu.com/6489818/ [14:51] mzanetti, that's what's there [14:51] mzanetti, it's just the Unity.Application plugin in surfaceflinger + mir versions [14:51] Saviq: yeah, seen that...I'm actually a bit confused by the splitting of QT_INSTALL_QML and QT_INSTALL_IMPORTS [14:51] Saviq: in Qt4 there was only QT_INSTALL_IMPORTS and everything in there [14:52] it was named slightly different iirc [14:52] mzanetti, yeah, I don't know of anything that's supposed to go into imports [14:52] tsdgeos, ↑ do you? [14:53] nope [14:54] Saviq: shouldn't we export whole environment instead of just that one variable? [14:54] mzanetti, nah, everything that *should* be there is already there [14:54] ok [14:54] mzanetti, i.e. everything else comes from /etc/profile and friends [14:54] greyback, on that note, why does libunity-mir1 have all the symlinks? sounds like it's built like a shared lib instead of plugin? [14:55] greyback, there's nothing that links against it is there? [14:55] Saviq: part of it is a shared lib [14:55] Saviq: the QML plugin links against it [14:56] greyback, installed with the QML plugin? [14:56] Saviq: yes [14:56] greyback, I mean http://paste.ubuntu.com/6489818/ [14:56] Saviq: oh, I see. Yeah that's not necessary [14:56] Can you guys help me install qmake on Ubuntu phone? would I install qt5-default if I want to compile a scope? [14:56] root@ubuntu-phablet:/home/phablet/scope# qmake [14:56] qmake: could not find a Qt installation of '' [14:57] karni, or export QT_SELECT=qt5 [14:57] karni: have you qt5-qmake installed? [14:57] greyback, that's qtchooser complaining about not knowing which version to look for [14:57] Saviq: actually this seems a new issue too. installing qt5-default shouldn't require you to export QT_SELECT [14:57] greyback, not that it can't find it [14:57] mzanetti, yes, but karni said he doesn't have it installed [14:57] I have qt5-default installed [14:58] karni, oh [14:58] I just reinstalled it [14:58] ok /me misunderstood [14:58] Saviq: I noticed this too a few days ago [14:58] Saviq: my run_on_device_scripts from my apps stopped working [14:58] Saviq: no, you were actually right, I just reflashed the device and tried again without it [14:58] now I installed it, exported QT_SELECT=qt5 aaand.. make: /usr/lib/x86_64-linux-gnu/qt5/bin/qmake: Command not found [14:59] karni qt5base-dev etc installed? [14:59] karni, yeah, that'd be qt5-qmake not installed [14:59] mzanetti: will check [14:59] karni, btw, scope? scope with qmake? [14:59] E: Unable to locate package qt5base-dev [15:00] karni, is there debian/control? [15:00] karni, apt-get install devscripts equivs [15:00] oh wait, I think I have the Makefile from my PC [15:00] * karni recreates that [15:00] karni, yeah, saw x86_64 there [15:01] Saviq: you got it. installing devscripts and equivs. [15:01] Thanks guys [15:01] First time I'm trying to actually compile a scope on the phone. [15:01] karni, mk-build-deps -s sudo -i will install the build deps for a local package (assuming you've got debian/control there) [15:02] On that note, need to tell Daniel H. (Cc mhall119 ) that we have a scope tutorial, but we don't have a scope tutorial for the phone, explicitly. It's the same code, but you do need some initial setup. [15:02] Saviq: yeah, that's just a bare scope code, so no debian/control [15:02] but I'll note that down :) [15:07] greyback, bug #1256014 [15:07] bug 1256014 in unity-mir (Ubuntu) "The Applications plugin should not be built as a shared library, but as a plugin instead, no SONAME etc." [Undecided,New] https://launchpad.net/bugs/1256014 [15:08] Saviq: ta [15:08] MacSlow, greyback, ricmm, bug #1256011 too [15:08] bug 1256011 in unity-notifications (Ubuntu) "Should install the applications plugin in plugin path from libunity-api-dev" [Undecided,Confirmed] https://launchpad.net/bugs/1256011 [15:08] MacSlow, you just got hit 'cause Satoris isn't here ;) [15:09] Saviq, ah ok :) [15:09] MacSlow, should be an easy fix, too [15:11] Saviq, will look into it [15:17] can't run apps?¿ [15:20] Saviq: ↑↑ ? [15:20] tsdgeos, try on console: [15:21] upstart-app-launch webbrowser-app [15:25] Saviq: nope [15:25] tsdgeos, output? [15:25] none :D [15:26] phablet@ubuntu-phablet:~$ upstart-app-launch webbrowser-app [15:26] phablet@ubuntu-phablet:~$ [15:26] tsdgeos, anything interesting in .cache/upstart/unity8.log ? [15:26] only osk stuff [15:26] maybe i should merge dandrader's branch before running [15:26] let me see if that helps [15:26] i mean merge with trunk [15:27] tsdgeos, one more interersting place to look [15:27] -r [15:27] tsdgeos, .cache/upstart/application-foo.log [15:27] where foo is the app name / app_id [15:28] ok, will heck after merge/recompile/run finishes [15:33] Saviq: hmmm, it indeed seems to be running [15:33] it's just not on the screen nor on the currently running apps :D [15:34] tsdgeos, and that's across unity8 runs? [15:34] tsdgeos, weird... stock unity8, too? rebooted? [15:34] that's when running dandrader|lunch's branch [15:34] let me try the stock one [15:37] stock one does actually work [15:38] silly me for not having tried that [15:38] sorry [15:41] and now it works on dandrader's one too [15:41] * tsdgeos confused [15:47] sil2100, anything blocking unity-scopes-api? [15:48] mhr3: hey, looking into scope.cpp, I see that you are trying to find a desktopFile for some "uri". where does this uri come from? [15:49] mzanetti, scopes [15:49] :) [15:49] mhr3: can we have activateApplication to pass the appId instead of the path? [15:49] i know that was unexpected, right? :) [15:49] mhr3: hi! I guess we just need someone to NEW it now, since it's in the queue - we'll be poking people soon [15:49] mhr3: we have quite some workarounds spread around to trim that down to an appId again [15:50] mzanetti, iirc the uri format you expect isn't the same as what the scopes provide [15:51] mhr3: well, we expect the appId. not really an uri format [15:51] see? :) [15:51] still doesn't answer my question. can we haz appIds there? [15:52] what are appIds in your part of the code? [15:52] "address-book-app" for example [15:53] mhr3: we have a discussion ongoing to drop this version stuff from click's appId's too [15:54] address-book-app doesn't sound overly unique to me [15:55] shouldn't it have like full publisher name and stuff? [15:56] mhr3: yeah. click apps still have com.foo.bar.appname [15:57] mhr3: but stuff installed with apt doesn't [15:57] mzanetti, iirc there's the appid schema-ed uri, imo we should use those [15:57] mhr3: tbh I don't really mind. but we used the appIds before that schema came up and now we have a mess of appId, url-schemas and full desktop file paths [15:58] so I'm trying to find where we could align that stuff [15:58] mzanetti, i'm all for unifying it, just don't think we're there yet [15:58] mhr3, there's also bug #1251635 [15:58] bug 1251635 in click (Ubuntu) "drop version numbers from users' .desktop file names" [Undecided,Confirmed] https://launchpad.net/bugs/1251635 [15:59] obviously we're note there yet :) but I'm trying to get there :D [16:01] Saviq: IMO we should tackle this at the sprint together. greyback, you and me sitting down and fix that appId everywhere [16:01] do you think thats feasable? [16:02] mzanetti: yes let's make time to do that [16:02] mzanetti, well clearly you do see what's involved, has to work for clicks, has to work for regular .desktop from apt, should be compatible with unity7 in some way [16:02] mzanetti, once you have a solution, open bugs :) [16:02] mzanetti: Saviq: did we ever decide if we'd handle appIds with or without the version string? [16:03] or push patches :) [16:03] greyback: I'm afraid we didn't get a response from the click people [16:04] mzanetti: was there a mail sent or something? [16:04] greyback: https://bugs.launchpad.net/ubuntu/+source/click/+bug/1251635 [16:04] Ubuntu bug 1251635 in click (Ubuntu) "drop version numbers from users' .desktop file names" [Undecided,Confirmed] [16:04] the day when we reported it, the people I pinged were either at a conference or in holidays [16:04] let me try again [16:05] ah.. today US is not around either [16:05] oh darn [16:06] cjwatson is online [16:06] mzanetti: ^^ [16:07] greyback, doesn't mean he's responsive [16:07] oh not for a while [16:07] greyback, he did send an email that he's gone [16:07] Saviq: ah ok [16:07] greyback, mzanetti FWIW internally we should go no-version-string [16:08] agreed [16:08] +1 [16:08] but then there is still the full desktop path coming from the scopes [16:08] and some places where we expect or get application:// [16:08] greyback, mzanetti it should be stripped as soon as it reaches us, and re-introduced when going out [16:08] it's close to impossible to change something in Stage.qml without breaking something [16:08] mzanetti, basename minus extension? [16:09] Saviq: would you do that in scope.cpp or when it reaches us on the QML side? [16:09] mzanetti, we just need to normalize to ".desktop file name minus extension, minus version" [16:09] mzanetti, cpp [16:09] ok... good [16:09] mzanetti, basically as close to process boundary as possible [16:13] do we have some easy ./run method now with scopes in their own repo? [16:13] mzanetti, same [16:13] mzanetti, ./build -s will install them [16:14] mzanetti, and then they're picked up automagically [16:16] Saviq: I'm feeling stupid. but can't find anything like that [16:16] mzanetti, hmm? [16:17] Saviq: unity8's ./build -s doesn't do anything like that [16:17] mzanetti, the package us unity-plugin-scopes [16:17] is [16:17] Saviq: ah. so you mean it will be installed with apt [16:17] mzanetti, yes [16:17] I meant a ./run if I want to change the code in there [16:17] mzanetti, just export QML2_IMPORT_PATH [16:18] mzanetti, with my branch that is [16:18] ok [16:18] mzanetti, anything you export before run [16:18] mzanetti, will be the first fallback [16:18] ok. cool [16:18] mzanetti, so builddir/plugins is first, then QML2_IMPORT_PATH, then mocks [16:36] tsdgeos, ping [16:36] Saviq: yes? [16:42] mhr3: https://code.launchpad.net/~mzanetti/unity-scopes-shell/activate-appid/+merge/197098 [16:44] mzanetti, doesn't handle the full path case [16:45] mhr3: no? [16:45] mzanetti, thank you ;D [16:45] ah wait... [16:45] Saviq: ? [16:45] mzanetti, I wanted to get rid of that chunk of code a long time ago ;) [16:45] mzanetti, there's too much g_ ;D [16:45] noone like gio, snip :P [16:45] * mzanetti is always happy to remove g_'s [16:46] evil, evil people! [16:46] g_free *shrug* [16:48] noone shrugs at g_free [16:48] g_free is awesome! [16:48] cause it does if(ptr) free(ptr) [16:49] clearly that's the best function since sliced bread... i mean vmsplice? [16:53] mhr3: better now? [16:54] mzanetti, /usr/share/applications/kde4/kate.desktop will give you "kate", are you sure that's enough? [16:55] hmm... fair point. [16:56] I think those are handled as kde4-kate [16:56] usually, yes === dandrader|lunch is now known as dandrader [17:07] mhr3: actually... can there be something like application:///usr/share/applications/kde4/kate.desktop ? [17:07] I mean the full path + application uri schema [17:07] yes [17:08] it's the same as application://kde4-kate.desktop but both are valid [17:08] yay for yet another combination :) I thought application has to be appid [17:08] it's for resolving ambiguities iirc [17:08] ie if you have foo.desktop in both /usr/share/applications and in ~/.local/share/applications [17:12] pushed another version [17:13] hardcoding paths? [17:13] eeeek! :P === boiko_ is now known as boiko === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader === salem_ is now known as _salem === seb128_ is now known as seb128 === thomi_ is now known as thomi