[04:32] <m4n1sh> jbicha: so what do you suggest where in the app to add Lock and Temporary files purge features?
[04:33] <m4n1sh> a separate tab? name it what?
[04:38] <jbicha> I think if you add screen lock options to the privacy panel, it basically needs to be a separate tab
[04:39] <jbicha> it might be cool if you could test for gnome-control-center 3.8 (either build-time or run-time) and only show that tab for that version or higher
[04:40] <jbicha> since there's quite a bit more work to be done before we can upgrade to gnome-control-center 3.8 without regressions on Ubuntu
[04:41] <jbicha> perhaps Purge Temporary Files could go in Clear Usage Data
[04:42] <jbicha> in lieu of anything better, I think I'd just call the tab "Screen Lock"
[04:43] <jbicha> I'm not really a designer; I just make some suggestions some times :)
[04:45] <m4n1sh> jbicha: or just "Behavior"
[04:45] <jbicha> I know you were trying to get rid of tabs and this would bring us back to 4 tabs on Ubuntu (since I've kept the Search Results tab because I hadn't been told specifically whether that was to be killed or not)
[04:45] <m4n1sh> Behavior can also take in "Search Results" tab
[04:45] <m4n1sh> so "Search Results" + Screen Lock  = Behavior tab
[04:46] <m4n1sh> jbicha: I guess that would be better - Search result and Screen Lock are both behavior.. again I am not designer
[04:46] <jbicha> I think Behavior is really ambiguous and we already use Behavior for 1 of the 2 tabs in Appearance
[04:47] <m4n1sh> jbicha: I can
[04:47] <jbicha> on the other hand, the Search Results tab is nearly empty
[04:47] <m4n1sh> cant remember Appearance tab anywhere
[04:47] <m4n1sh> right now I have only Background
[04:48] <m4n1sh> Brightness and Lock, Displays etc
[04:48] <m4n1sh> yeah. I was trying to merge Search Results and Screen Lock
[04:48] <jbicha> it looks like www.omgubuntu.co.uk/wp-content/uploads/2012/04/Screen-Shot-2012-04-26-at-13.42.jpg
[04:48] <jbicha> I wish Unity would have taken inspiration from GNOME 3.8's new Search panel and moved the on/off switches for 100 scopes to System Settings where it's a lot more discoverable
[04:49] <jbicha> I had no idea how to turn on/off individual scopes without "reading the manual"
[04:49] <m4n1sh> looks like my gnome-control-center is from ricotz or gnome3 team ppa
[04:49] <m4n1sh> so it looks vanilla
[04:50] <jbicha> m4n1sh: which desktop are you running?
[04:50] <m4n1sh> jbicha: it shows only on unity? I am running shell atm
[04:50] <jbicha> yeah, close settings then run XDG_CURRENT_DESKTOP=Unity gnome-control-center
[04:51] <jbicha> yes it's bad usability to have the settings panel work differently depending on what desktop you're using but it's a compromise :|
[04:52] <m4n1sh> plus I am not very convinced that allowing to change icon size is actually useful
[04:52] <m4n1sh> if I was the maintainer, I would have just dropped that setting
[04:53] <jbicha> some people do really like that option and it was another compromise in lieu of directing people to ccsm or dconf-editor at the time
[04:54] <m4n1sh> oh yeah. it is still way better than asking people to open ccsm and blow up everything
[04:54] <jbicha> we may even still have been using gconf at the time which had a horrible UI for tweaking Unity
[04:54] <jbicha> dconf-editor works well now for Unity
[04:55] <m4n1sh> so what should be done?
[04:55] <m4n1sh> how to integrate those settings?
[04:55] <jbicha> we need to tweak the "When searching in the Dash" section to only show when Unity is running (we can check dbus for it or check for XDG_CURRENT_DESKTOP)
[04:57] <jbicha> I suggest you ask the Unity design guys whether they still want the online search toggle for saucy
[04:58] <m4n1sh> actually Appearance > Behavior should be renamed to Appearance > Launcher and Workspaces
[04:58] <jbicha> unless a designer suggests something else, it's up to you whether you do it in one tab or two
[04:58] <m4n1sh> makes more sense
[04:59] <m4n1sh> and Behavior for Lock Screen etc
[04:59] <m4n1sh> jbicha: unity dev happens on ubuntu-devel or some other list?
[05:01] <jbicha> there's #ubuntu-unity (US or Europe hours); I'm not sure that Unity designers really use mailing lists
[05:04] <m4n1sh> jbicha: do you think I should release the current code so that it is included in saucy and tested
[05:04] <m4n1sh> since getting rest of the code might take a few weeks
[05:05] <m4n1sh> as this new UI  has never been tested on a wider scale. It might contain dragons
[05:05] <jbicha> m4n1sh: well I have 2 bugs I filed today and there's my 3.8 merge proposal I'd like to see handled first ;)
[05:06] <jbicha> but as long as there's no known regressions I agree it's better to land sooner :)
[05:06] <m4n1sh> jbicha: true. those will be included in next release. I am talking about getting the gnome privacy thing at later stage
[05:06] <m4n1sh> I am going to merge gnome 3.8 branch now and then work on making diagnostic available in standalone too
[05:07] <jbicha> yes, gnome-control-center 3.8 is likely at least a few weeks away and the lock settings aren't the only blocker
[05:07] <m4n1sh> and then release
[05:07] <m4n1sh> what else blockers are there?
[05:09] <jbicha> from me, there's bug 1192777 too; either it should work or be hidden; everything else seems to look fine although I use GNOME Shell a lot more than Unity
[05:09] <ubot2> Launchpad bug 1192777 in Activity Log Manager ""Send a report automatically" feature doesn't work" [High,Confirmed] https://launchpad.net/bugs/1192777
[05:10] <m4n1sh> for #1192777 I would need to ask evan dandrea as he manages the diagnostic thing
[05:13] <m4n1sh> jbicha: sent him a mail. He is in London Time, so he might get it done in 12 hours. (Hopefully)
[05:38] <pitti> Good morning
[07:26] <mlankhorst> bonjour
[07:27] <RAOF> pitti: Good morning!
[07:28] <RAOF> pitti: I hope that https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1192842 contains enough information for you.
[07:28] <ubot2> Ubuntu bug 1192842 in systemd (Ubuntu) "logind loses track of sessions when multiple sessions share a VT" [Undecided,New]
[07:29] <mlankhorst> RAOF: can you accept xxv-intel and mesa in quantal, raring, and the lts versions in precise?
[07:31] <pitti> hey RAOF
[07:32] <RAOF> mlankhorst: Sure.
[07:32] <pitti> RAOF: so the effect is that c2, c3, c5 don't get ACLs?
[07:33] <RAOF> pitti: Correct.
[07:33] <RAOF> Although they *do* have ACLs when I first log in. It's only after switching to the guest session that logind gets confused, and no longer thinks any session is active.
[07:39] <RAOF> Hm. Do you find the bug blurry?
[07:47] <seb128> good morning desktopers
[07:52] <didrocks> hey seb128!
[07:53] <seb128> didrocks, lut! happy thursday
[07:53] <didrocks> happy thursday as well :)
[07:53] <seb128> how is the weather for you?
[07:53] <seb128> no ice falling from the sky? no storm?
[07:53] <seb128> it has been raining a lot here, some ice as well
[07:53] <didrocks> no, a nice and sunny weather, but with more resonnables temperatures
[07:53] <didrocks> oh?
[07:53] <didrocks> just during the night here
[07:53] <seb128> I was up at 5am with the noise that was making on my windows
[07:54] <didrocks> waow
[07:54] <didrocks> at least, the temperatures are better I guess?
[07:54] <seb128> it's better this morning, though heavy raining now
[07:55] <didrocks> well, if that can result for you a cooler week-end…
[07:55] <didrocks> seb128: I think I finally found a case where unity-gtk-module is regressing us! Dart editor (so, I guess eclipse as well), has its menu exported, but there is nothing in them, just the first level is available
[07:58] <seb128> interesting
[07:58] <seb128> is that java and gtk2?
[07:58] <didrocks> yeah, I'm going to open a bug
[07:58] <didrocks> I guess it's java + the sdk theme
[07:58] <didrocks> gtk*
[07:59]  * didrocks tries installing eclipse and pulls the world
[08:01] <seb128> didrocks, were we exporting menus for that one before?
[08:01] <seb128> I know we had eclipse blacklisted at some point
[08:01] <didrocks> seb128: yeah, it was exported and working fine
[08:01] <didrocks> for dart
[08:01] <didrocks> which I guess didn't use eclipse blacklist
[08:01] <seb128> right
[08:01] <seb128> in any case please open a bug ;-)
[08:02] <didrocks> yeah, let me try to confirm with eclipse, seeing if we still have that list
[08:02] <didrocks> seb128: on another note, I'm starting to regret having pulled -proposed in the ppa
[08:02] <seb128> too much transitional breakages?
[08:02] <didrocks> a lot of FTBFS this night (and still now), because of archive mismatch due to new gtk3
[08:02] <didrocks> yep
[08:02] <seb128> like gtk not built on arm?
[08:02] <didrocks> exaclty!
[08:03] <didrocks> exactly*
[08:03] <seb128> shrug
[08:03] <seb128> still didn't fail, it failed again on some stupid test
[08:03] <didrocks> yeah, seems like shacky ones
[08:03] <Laney> howdy
[08:03] <didrocks> hey Laney!
[08:03] <seb128> Laney, good morning
[08:03] <seb128> I retried 3 times yesterday
[08:03] <didrocks> seb128: eclipse is still blacklisted (or exported the env var)
[08:04] <didrocks> seb128: should I disable -proposed for now?
[08:04] <didrocks> meanwhile? (I wanted to do the Xorg transition today though)
[08:04] <seb128> didrocks, I need to fix gtk today in any case, but your call
[08:04] <didrocks> seb128: do you need help on those tests?
[08:04] <seb128> didrocks, the xorg transition? 1.14 to saucy you mean?
[08:05] <didrocks> no, on the gtk failures
[08:05] <seb128> didrocks, thanks for the offer, let me already have a look to what's the issue with gtk
[08:05] <didrocks> ok
[08:05] <seb128> didrocks, sorry, the 1.14 to saucy question was for "I wanted to do the Xorg transition today"
[08:05] <didrocks> meanwhile, let me remove all this java things that got pulled :p
[08:05] <seb128> what transition?
[08:05] <didrocks> seb128: oh right, the 1.14 version
[08:05] <Laney> wasn't gtk failing in the rebuild test already?
[08:05] <seb128> Laney, it was
[08:06] <didrocks> seb128: we need to get in sync for the barrier push change
[08:06] <seb128> didrocks, right, I'm not sure 1.14 is ready to be pushed to the archive yet though
[08:06] <didrocks> seb128: oh? mlankhorst is telling me he's waiting for days
[08:07] <seb128> didrocks, we promised asac and olli that we would make a call for testing before upload, once we have a ppa with everything including unity to test
[08:07] <seb128> mlankhorst, did you get fglrx ready for the new abi and tested as well?
[08:07] <seb128> it was not in the ppa last time I asked
[08:07] <didrocks> seb128: ah, so independant from us then, I think mlankhorst already has all with the special unity
[08:07] <didrocks> sil2100: FYI ^
[08:08] <seb128> Laney, I wonder if the GTK build issue is a real bug, there are some FAIL with those warnings
[08:08] <seb128> Gtk-WARNING **: Theme parsing error: Raleigh.css:394:140: Cannot animate property 'background-image'
[08:08] <mlankhorst> seb128: i think we do, but tseliot was busy making the driver ready for 3.10
[08:09] <didrocks> mlankhorst: please ping sil2100 once we have a definitive ack to transition then :)
[08:09] <Laney> Don't know, that soundsl ike something which could make tests fail
[08:09] <Laney> maybe a question for upstream
[08:09] <Laney> unless the same messages appear on good arches too
[08:09] <seb128> mlankhorst, can you check and be sure? we will be in trouble in we break anybody's box on upgrade, we should make sure we have all drivers ready, including nvidia and fglrx and we should do a call for testing on ubuntu-devel@ with the ppa once it's ready for saucy copy
[08:10] <didrocks> Laney: you mean armhf is a bad arches? :p
[08:10] <Laney> haha you won't catch me out there
[08:10] <Laney> I mean arches in which the test is good :P
[08:10] <seb128> no cookie for armhf!
[08:10] <didrocks> ;)
[08:10] <didrocks> heh
[08:10] <Laney> like ppc :-)
[08:10] <mlankhorst> seb128: the call for testing already happened before raring was released, someone was optimistic that fglrx would be ready in time for that..
[08:11] <didrocks> Laney: you can try as hard as you want, that won't change my view on ppc though :p
[08:11] <Laney> yeah no warnings there
[08:11] <Laney> I guess that is pretty relevant then
[08:11] <seb128> mlankhorst, well, we need a new call for testing on saucy, things changed since, and we need fglrx to be part of the batch which is ready to be tested
[08:11] <didrocks> +1
[08:11] <seb128> Laney, there was no such warning on armhf in the 0ubuntu4 upload either
[08:12] <Laney> right, otherwise it would have failed to build
[08:12] <Laney> so something clearly changed
[08:12] <chrisccoulson> yoyoyo
[08:12] <seb128> oh, a chrisccoulson!
[08:12] <Laney> sup chrisccoulson
[08:12] <chrisccoulson> heh
[08:13] <chrisccoulson> i thought i'd pop in here and say hello ;)
[08:13] <seb128> chrisccoulson, hey, how are you? it has been a while you didn't say hello to us ;-)
[08:13]  * seb128 hugs chrisccoulson
[08:13] <chrisccoulson> seb128, yeah, i'm good thanks. busy, but good
[08:13] <chrisccoulson> how are you?
[08:13]  * chrisccoulson hugs seb128
[08:13] <seb128> I'm good thanks
[08:13] <didrocks> hey chrisccoulson! :)
[08:14] <chrisccoulson> hi didrocks :)
[08:14] <seb128> chrisccoulson, I've been told you will solve https://bugs.launchpad.net/ubuntu/+source/qtwebkit-opensource-src/+bug/1192567 for us ;-)
[08:14] <ubot2> Ubuntu bug 1192567 in qtwebkit-opensource-src (Ubuntu) "[MIR] qt5webkit " [Undecided,New]
[08:14] <chrisccoulson> hah :)
[08:38] <sil2100> ACK
[08:45] <seb128> shrug
[08:45] <seb128> how slow is that armhf porter box, takes like a minute to run configure on gtk
[08:45] <seb128> going to take a while to debug that issue :/
[08:45] <ogra_> its a panda :/
[08:46] <ogra_> still no signs of the calxeda nodes for us
[08:46] <ogra_> (by the time we get them you will be able to buy the new amd arm server machines .... )
[08:47] <Sweetshark> seb128: the mother of all hacks seems to have worked ... only to find later issues.
[08:47] <seb128> :-(
[08:47] <Sweetshark> seb128: but those seem to be solvable (just taking time with the turnaround of this monster on the buildds)
[08:48] <seb128> ok, that's something
[08:48] <seb128> keep the good work ;-)
[08:48] <Sweetshark> seb128: cant we just upload the binaries from my machine? its so much faster than those buildds :P
[08:49] <seb128> talk to the is guys to see how much they like thatr :p
[09:00] <Sweetshark> seb128: hrhr. I guess I should wear full body armor, plate for that.
[09:05] <seb128> Laney, that's not a small diff to review :p
[09:05] <Laney> hrhr
[09:05] <Laney> Just tell me my QML skills suck and to start again
[09:05] <seb128> Laney, why didn't you use background images from ubuntu-wallpapers?
[09:06] <seb128> I guess that's not installed on the touch image?
[09:06] <Laney> exactly
[09:06] <Laney> I tried to use some of the images in /usr/share/backgrounds/ on the device but they didn't load up
[09:06] <seb128> anyway demo artwork is fine until we get some artwork from design on the image ;-)
[09:07] <seb128> Laney, btw did the recent commit fix the issues you were having the other day?
[09:07] <seb128> the commit from Ken to change the resize mode seems to have made it behave for me
[09:07] <Laney> which one?
[09:07] <seb128> you asked monday if I tested the settings on a device
[09:07] <Laney> getting stuck in a panel?
[09:07] <seb128> yes
[09:07] <Laney> let me check
[09:07] <seb128> you said you couldn't go back or something
[09:08] <Laney> yes seems good now
[09:08] <seb128> great ;-)
[09:08] <seb128> Laney, btw your QML skills suck, please start again :p
[09:08]  * seb128 hides
[09:08] <Laney> haha
[09:08] <seb128> (joke aside, reviewing)
[09:09] <Laney> knew I should have become a monk instead
[09:09] <seb128> I'm pondering doing nitpicking :p (why blank lines between imports)
[09:10] <chrisccoulson> i love nitpicking
[09:10] <seb128> ;-)
[09:12] <Laney> btw the idea is to remove the UbuntuSwappableImage thing once I get CrossFadeImage and then a component on top of that into the UI toolkit
[09:12] <seb128> great
[09:13] <seb128> Laney, how annoying is it if I ask "can we get it renamed "background" rather than "appearance", since that's what the designs call it"?
[09:13] <Laney> I got appearance from https://wiki.ubuntu.com/Appearance#Phone
[09:13] <seb128> Laney, we should at least change the visible name, but I would prefer the filenames to match
[09:14] <Laney> but ok
[09:14] <seb128> thanks
[09:14] <Laney> don't we have a qmlproject thingy for the system settings?
[09:14] <seb128> qmlproject thingy?
[09:15] <Laney> what do I pick to load it in qt creator?
[09:15] <Laney> I've been working with my separate project up to now
[09:15] <seb128> I usually open system-settings.pro from the srcdir
[09:16] <seb128> but that's a bit of a pain, I didn't find how to run the app or test panels from qtcreator
[09:16] <seb128> I just do the editing in there
[09:16] <Laney> that just opend it as text
[09:16] <seb128> and then have a command line where I cp over the system version and run it
[09:17] <seb128> did you do "open file or project" and picked the .pro?
[09:17] <seb128> that works for me...
[09:17] <seb128> system-settings.pro that is
[09:18] <Laney> ap
[09:18] <Laney> ah*
[09:18] <Laney> I got it to open a configuration screen and then from there to the project
[09:22] <seb128> Laney, right, that's what it does here (the configuration screen is only the first time, it writes a .pro.user then)
[09:22] <Laney> yeah I just don't know why it insisted on opening it as text
[09:22] <Laney> oh well
[09:32] <Laney> seb128: OK, pushed the rename and some other cleanups
[09:32] <seb128> great
[10:34] <Laney> Is there a packaging branch for nautilus 3.8?
[10:44] <seb128> Laney, yes, linked on https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/1130746
[10:45] <ubot2> Ubuntu bug 1130746 in nautilus (Ubuntu) "Update to nautilus 3.8" [Wishlist,Confirmed]
[10:45] <Laney> oh cool, didn't think to look there :P
[10:57] <seb128> Laney, do you work on the update?
[10:58] <Laney> just putting a patch in
[10:58] <Laney> I figure there are already people looking at that update
[10:59] <seb128> right
[10:59] <seb128> it's almost done, it just need the template patch to be adapted
[10:59] <seb128> and some testing
[10:59] <seb128> and to figure out the background story
[10:59] <seb128> I think we should just re-enable the gsd plugin
[11:00] <Laney> the rationale for disabling it was because it was redundant, so if that stops being true then +1
[11:01] <seb128> right, nautilus 3.8 stops drawing the wallpaper
[11:01] <seb128> it just put icons on a rgba background
[11:01] <seb128> in GNOME gnome-shell draws the wallpaper
[11:02] <seb128> until unity8 does it for us I guess we can just use the g-s-d code back there
[11:02] <ricotz> hi desktopers
[11:03] <ricotz> is it possible that last bamf release silently broke abi?
[11:03] <ricotz> (0.5.0daily13.06.19-0ubuntu1)
[11:04] <seb128> Trevinho, ^
[11:05] <Laney> might help if you describe what you are seeing
[11:05] <seb128> ricotz, could be, Trevinho ported it to gdbus which is a non trivial change
[11:05]  * didrocks is watching gtk3 tests executing while staring at seb128 :)
[11:05] <seb128> didrocks, I couldn't get tests to fail on porter-armhf, tried in loop, under valgrind, ...
[11:05] <ricotz> seb128, i see, then an soname bump had to be done
[11:05] <didrocks> seb128: yeah for easy reproducer :/
[11:05] <seb128> didrocks, I think I will just add a || true to the test to unblock the archive
[11:06] <didrocks> let's cross fingers
[11:06] <seb128> ricotz, why? what dbus library is used by the code is not part of the abi/api
[11:06] <didrocks> ricotz: at least, it did still work for application matching and so on with new bamf + old unity
[11:06] <seb128> ricotz, the interfaces shouldn't have changed, the update didn't break unity/didn't require unity changes
[11:07] <didrocks> seb128: yeah, let's see *if* by chance, this passed at least for this round
[11:07] <ricotz> seb128, i am meant the abi of libbamf3
[11:07] <seb128> didrocks, right
[11:07] <ricotz> i will downgrade to the older bamf to test/confirm
[11:07] <didrocks> (/build/buildd/gtk+3.0-3.8.2/debian/build/shared/tests/a11y/.libs/lt-tree-relationships:6264): Gtk-WARNING **: Theme parsing error: Raleigh.css:394:140: Cannot animate property 'background-image'
[11:07] <didrocks> FAIL
[11:07] <seb128> ricotz, I'm not sure the abi changed, as Laney said it would help if you described your issue
[11:07] <didrocks> seb128: argh ^
[11:07] <seb128> didrocks, you get it?
[11:08] <didrocks> seb128: no, I was looking at https://launchpad.net/ubuntu/+source/gtk+3.0/3.8.2-0ubuntu5/+build/4727251
[11:08] <seb128> oh
[11:08] <seb128> let me upload with make check || true
[11:08] <didrocks> seb128: thanks, let's hope it's the only one… :)
[11:08] <seb128> I've another build running on porter but I don't want to keep the archive screwed for the day
[11:08] <ricotz> Laney, seb128, this is not a unity issue, but plank (which uses bamf) isnt behaving like it should
[11:08] <Laney> on armhf only if you're going to do that please ...
[11:09] <ricotz> i was just taking a look at bamf change which arent published yet, which are already looking like an abi break
[11:09] <ricotz> http://bazaar.launchpad.net/~unity-team/bamf/trunk/revision/557
[11:09] <ricotz> e.g. the changes to  lib/libbamf/bamf-view.h
[11:10] <ricotz> that made think there might be something sneaked in earlier too
[11:10] <Laney> try the abi compliance checker on it
[11:11] <ricotz> ok
[11:11] <ricotz> thanks
[11:12] <Trevinho> ricotz: hey
[11:15] <ricotz> Trevinho, hi
[11:16] <Trevinho> ricotz: I thought I didn't change the ABI, but that class reoreding seems to do that... I wanted to revert it back, then I forgot
[11:16] <ricotz> Trevinho, if there are some ordering changes like the one i linked then it is a break
[11:16] <Trevinho> ricotz: even if I've never recompiled unity and my tools here and everything continued to work
[11:17] <ricotz> Trevinho, ok, i confirmed the reverting to the older bamf version it works again
[11:17] <ricotz> Trevinho, were there any behaviour changes too?
[11:17] <Trevinho> ricotz: did you get crashes before?
[11:17] <ricotz> Trevinho, it doesnt crash, it seems to do the wrong things
[11:18] <ricotz> which would happen if signals are reordered
[11:18] <Trevinho> ricotz: ah, so are you sure you're running the right deaemon?
[11:18] <seb128> Laney, http://paste.ubuntu.com/5783329/
[11:18] <Trevinho> ricotz: that would happen if you'd use default handlers, but it's quite unlikely that you do that... do you?
[11:18] <seb128> Laney, does that look correct to you? I hate makefiles :p
[11:19] <ricotz> Trevinho, i recently killed the daemon, but i could be the old one was still running before
[11:20] <ricotz> Trevinho, will recheck
[11:20] <Trevinho> ricotz: I've changed a couple of things on the daemon + library: use dbus properties instead of methods + signal when it's the case (deprecating them, and they're about to be removed soon)
[11:21] <Trevinho> so, for example for titles and icons we now use them... the same for urgent/active/running events
[11:21] <ricotz> i am only using the library, not the dbus interface directly, if that is what you meant
[11:22] <Trevinho> ricotz: yes, I know.... But while the dubs api has been integrated (not changed yet), the new library only talks with newly added dbus entries
[11:22] <Trevinho> s/dubs/dbus
[11:23] <Laney> seb128: why not just ifeq ($(DEB_HOST_ARCH),armhf)?
[11:24] <seb128> Laney, because I hate makefiles and I copied what I find in another rules on my disk ;-)
[11:24] <seb128> Laney, let me try that
[11:24] <Laney> you typically use the filter functions when it's a list
[11:26] <seb128> Laney, ok, that works, let me upload that, thanks ;-)
[11:26] <ricotz> Trevinho, ok, i can confirm the behaviour changed here with the new version
[11:27] <ricotz> Trevinho, will try to figure out what it is
[11:32] <Trevinho> ricotz: ok, thanks let me know
[11:32] <Trevinho> ricotz: what's exactly wrong for now?
[11:56] <ricotz> Trevinho, when closing an applications (not in favorites) the signals-(count/ordering) might have changed, could be i relied on a buggy behaviour which you have fixed now
[11:56] <ricotz> Trevinho, so i cant point a finger on it yet
[11:57] <Trevinho> ricotz: ok, take your time... Thanks for testing that
[13:26] <Laney> has anyone noticed that all of the categories apart from "Applications" keep getting disabled in current unity?
[13:26] <Laney> and sources
[13:26] <seb128> Laney, disabled where?
[13:26] <Laney> in the dash
[13:27] <seb128> in the stuff listed in the home screen?
[13:27] <seb128> or in the filter UI on the side?
[13:27] <seb128> do you do a search or just open the dash?
[13:27] <Laney> if you expand "Filter results"
[13:27] <Laney> all of the above
[13:28] <Laney> I've definitely enabled them all a couple of times but they are all now disabled apart from applications
[13:28] <seb128> I don't think they are meant to keep state
[13:28] <seb128> check with mhr3 or pstolowski
[13:28] <seb128> when you do a query the server send back what scopes to use
[13:29] <Laney> indeed, if I search for a filename it is returned
[13:29] <Laney> but after a 5s-ish delay
[13:30] <seb128> I can confirm that yes
[13:30] <seb128> the whole server returning scope to activate seems flaky/subject to latency to me
[13:30] <pstolowski> Laney: so it's when you enter a query?
[13:30] <Laney> pstolowski: well, it means that when I open the dash I only see applications listed straight away
[13:30] <mhr3> Laney, by design
[13:30] <Laney> but the returned query results indeed seem to ignore that filter
[13:31] <mhr3> Laney, filtering in home is "post-search", not pre- as elsewhere
[13:32] <pstolowski> Laney: as soon as you start typing, your filter settings are ignored and reset with recommendations. if you want to apply filters for the results, you have to apply them after the search. it's by design
[13:32] <Laney> hrm
[13:33] <Laney> I suppose I expected it to override the smart scope results
[13:33] <pstolowski> Laney: but there's also another concept - Dash is no search query is a "default" view. you can customize it by selecting filters, and it should remember the state
[13:33] <pstolowski> s/Dash is/Dash with/
[13:33] <Laney> well that seemed to be reset too
[13:33] <Laney> unless I searched and then changed it
[13:34] <pstolowski> Laney: hmm then this would be a bug
[13:34] <Laney> I might have done that though, because I didn't know about this concept
[13:34] <mhr3> pstolowski, aah, actually my weird bug was caused by more-suggestions being active for the default view... :/
[13:35] <pstolowski> mhr3: ahh, then it's a problem that it's not ignored I think
[13:35] <mhr3> pstolowski, yea, it seems to retain results from the last search
[13:35] <pstolowski> mhr3: tough I think we wanted to support more suggestions in surfacing mode anyway
[13:37] <Laney> I also never seem to get Music results
[13:38] <Laney> guess that's a server thing
[13:40] <seb128> Laney, didrocks: I can reproduce the arm test issue
[13:40] <seb128> I lean toward blaming desrt for it
[13:40] <didrocks> seb128: \o/
[13:40] <seb128> but it could be doko
[13:40] <didrocks> seb128: what box?
[13:40] <seb128> porter-armhf.c.c
[13:40] <pstolowski> mhr3, Laney: pls open bugs if you feel something's wrong
[13:40] <seb128> doing a package build
[13:40] <Laney> pstolowski: OK, where's the right place for server bugs?
[13:41] <Laney> or just all against unity and they'll be triaged
[13:41] <Laney> ?
[13:41] <seb128> didrocks, it seems to be an issue with the css embedded in the binary, that's coming out of gresources
[13:41] <seb128> didrocks, it only happens with the new libgtk, not with the system one, where the content shouldn't have changed
[13:41] <seb128> I wonder if that's a new glib issue
[13:41] <didrocks> seb128: interesting, but it seemed to be more a racy issue as you couldn't reproduce reliably on porter earlier?
[13:42] <seb128> but we previous archive built was already using the new glib
[13:42] <didrocks> ah…
[13:42] <mhr3> Laney, music where? most music is in More suggestions
[13:42] <Laney> mhr3: I get file results
[13:42] <seb128> didrocks, I don't know why I couldn't reproduce a source build, I did a ./configure && make earlier to avoid the triple build from the package
[13:42] <Laney> mhr3: but if I go to filter results and turn on Music then I also see the proper entries
[13:42] <seb128> didrocks, but I can reproduce with a package build tree
[13:43] <pstolowski> Laney: yup, just against Unity. do you have any local music (indexed with rhythmbox)?
[13:43] <seb128> didrocks, maybe a builddir!=srcdir issue
[13:43] <Laney> pstolowski: maybe not, I use banshee
[13:43]  * Laney just imported it, let's test
[13:43] <didrocks> seb128: yeah, that's possible
[13:44] <didrocks> at least, great to have a reproducer on a box :)
[13:44] <pstolowski> Laney: banshee should work too
[13:44] <Laney> seems the same
[13:44] <Laney> does it work for you?
[13:44] <Laney> e.g. Open dash, search for "Dream Theater", see files and some random applications but not the proper music results
[13:45] <mhr3> any idea why bzr doesn't want to do SSH?
[13:45] <mhr3> bzr: ERROR: Don't know how to handle SSH connections. Please set BZR_SSH environment variable.
[13:45] <Laney> Filter results -> Music, happy days
[13:46] <Laney> anyway, supposed to be sponsoring
[13:47] <mhr3> Laney, so it's not selected when you do the search?
[13:47] <Laney> no
[13:49] <pstolowski> Laney: what does 'gsettings get com.canonical.Unity.Lenses always-search' say?
[13:50] <Laney> pstolowski: ['applications.scope', 'music.scope', 'videos.scope', 'files.scope']
[13:50] <pstolowski> hmm
[13:52] <pstolowski> Laney: ok, please open a bug (quote this key value in the report)
[13:52] <Laney> wilco
[13:55] <sil2100> didrocks: quick status update:
[13:56] <Laney> cyphermox: !!!
[13:56] <Laney> What's up with https://code.launchpad.net/~gunnarhj/indicator-datetime/days-months/+merge/159214 ?
[13:57] <cyphermox> well it's approved
[13:57] <cyphermox> it would be nice is others could chime in given the number of reviewers
[13:57] <sil2100> didrocks: so, I still see some issues with indicator-client sadly
[13:57] <sil2100> cyphermox: hi
[13:57] <cyphermox> charles: ^ wanna take a look?
[13:57] <Laney> I guess they should look at it too
[13:58] <Laney> pitti: you got requested there ^ also
[13:58] <cyphermox> just a review to see if it all makes sense, then top-approve?
[13:58] <sil2100> cyphermox: regarding indicator-client - I wanted to build it locally but I get a failure on one of the tests all the time
[13:58] <cyphermox> Laney: it does pretty much make sense though, then you got to step over the knee-jerk reaction of rejecting because it "breaks POSIX" :)
[13:58] <sil2100> cyphermox: and also, I see some strange dependencies in the indicator-client debian/control
[13:59] <pitti> Laney: hm, I have an odd déjà-vu effect; I'm fairly sure I commented on this like 5 times already
[13:59] <cyphermox> like?
[13:59] <Laney> cyphermox: I deliberately haven't read it because I don't want to have any kind of reaction
[13:59] <Laney> pitti: Don't see you on the MP :P
[13:59] <cyphermox> sil2100: it built fine here
[13:59] <sil2100> cyphermox: since what are indicator-battery and indicator-time? Are they by any chance older versions of indicator-power and indicator-datetime? Since they have no LP projects and no sources
[13:59] <cyphermox> err
[13:59] <Laney> pitti: But if you don't care then just abstain?
[13:59] <sil2100> cyphermox: there are no releases for those ;/
[13:59] <cyphermox> you shouldn't get indicator-battery and indicator-time anymore
[13:59] <sil2100> cyphermox: but the packaging deps on them
[13:59] <sil2100> So duh
[14:00] <cyphermox> sil2100: not afaik
[14:00] <pitti> Laney: ah, my awesome bar has two similar MPs which got deleted
[14:00] <seb128> Laney, I think nobody, out of Gunnar, cares enough about those details to add hacks in different places
[14:00] <marga> I need someone that really understands compositing, xdamage and all that kind of stuff... Does that person exist?
[14:00] <cyphermox> I was just looking for that
[14:00] <sil2100> I just branched it
[14:00] <sil2100> lp:indicator-client
[14:00] <sil2100> Let me pull
[14:00] <Laney> seb128: I'm just doing the pilot thing here!
[14:01] <seb128> Laney, want to join the settings meeting?
[14:01] <Laney> "annoy people until they make things go away"
[14:01]  * Laney cackles
[14:01] <Laney> oh yeah, that
[14:01] <Laney> yes, yes I do!
[14:01] <sil2100> cyphermox: http://bazaar.launchpad.net/~phablet-team/indicators-client/trunk/view/head:/debian/control#L63
[14:01] <sil2100> Here is indicator-battery
[14:01] <cyphermox> sil2100: definitely nothing of the sort here, although it does look for com.canonical.indicator.battery in power.json, but that shouldn't translate to a depends
[14:02] <sil2100> cyphermox: but it's hard-coded into debian/control
[14:02] <pitti> Laney: done
[14:02] <Laney> pitti: merci!
[14:02] <pitti> de rien
[14:02] <sil2100> cyphermox: indicator-time is a Depends: of indicators-client-plugin-datetime in trunk
[14:02] <sil2100> cyphermox: so not sure what trunk you are looking at?
[14:02] <sil2100> (line 73)
[14:03] <sil2100> Maybe I'm looking at the wrong thing?
[14:03] <cyphermox> ah, nevermind, I have local changes I didn't commit
[14:03] <cyphermox> anyway, it just needs to be ripped out
[14:03] <sil2100> ACK, let's do that then and merge it
[14:03] <sil2100> didrocks: ^
[14:04] <sil2100> Then we can try re-adding it, maybe the unit test failure is only on my local machine
[14:04] <sil2100> Who knows if some packages aren't broken, but I doubt it
[14:04] <sil2100> The following tests FAILED: 3 - dynamicmenutest (Failed)
[14:05] <didrocks> sil2100: catching up!
[14:05] <sil2100> didrocks: from other news:
[14:06] <sil2100> The armhf failures we've been getting in indicators (besides the libusermetrics thing) - rerunning wasn't helping, as it seems (which me and Francis tried looking into) it's an issue with -proposed and gtk+3.0 for armhf
[14:06] <sil2100> A recent push made by seb128 is probably still buuilding for armhf
[14:07] <sil2100> I hope once it's finished, it might unblock everything again for armhf in the indicators sstack
[14:07] <cyphermox> sil2100: I get the same test failing
[14:07] <cyphermox> better ask tedg
[14:08] <tedg> sil2100, Link?
[14:08] <cyphermox> tedg: http://paste.ubuntu.com/5783768/
[14:09] <sil2100> tedg: uh oh! No link, since I was running it locally
[14:09] <sil2100> tedg: one unit test in indicators-client is failing
[14:09] <sil2100> dynamicmenutest
[14:09] <tedg> dednick, Do we need indicators-client anymore?
[14:09] <tedg> I think it's been merged into Unity8
[14:09] <sil2100> Now that's a good question
[14:11] <didrocks> sil2100: yeah, it's still gtk+3.0 is still building as you can see:
[14:12] <didrocks> sil2100: just finished btw: https://launchpad.net/ubuntu/+source/gtk+3.0/3.8.2-0ubuntu6/+build/4730684
[14:12] <didrocks> sil2100: we need to wait for the next publisher and can relaunch
[14:12] <sil2100> \o/
[14:13] <sil2100> This would unblock the indicator stack, since the libusermetrics 'base' fix is in already
[14:14] <didrocks> great :)
[14:15] <didrocks> then restarting all the stacks?
[14:16] <dednick> tedg: in the pipeline
[14:18] <tedg> dednick, Do we need to fix the test suite for it?  Could we just remove the test?
[14:18]  * tedg tells didrocks to close his IRC client
[14:18] <seb128> larsu, kenvandine: where is the work in progress for gsettings?
[14:18]  * kenvandine is anxiously awaiting to find that out 
[14:18] <tedg> dednick, Just thinking if it's code we don't care about, we shouldn't spend time fixing the tests.
[14:18] <dednick> sil2100, tedg: there is a regression in indicators-client. there was a change to qmenumodel awhile ago.
[14:18] <dednick> tedg: remove it
[14:19] <tedg> dednick, We can remove all of indicators-client or just the test.
[14:19] <dednick> tedg: just the test
[14:19] <didrocks> tedg: ok, I *fully* trust you :p
[14:19] <tedg> K, cool, thanks dednick
[14:19] <larsu> seb128, kenvandine: I just started it last night, will upload it today
[14:20] <kenvandine> larsu, we're all waiting for you :)
[14:20] <sil2100> tedg: will you create the merge for that or busy?
[14:20] <tedg> didrocks, sil2100, how do you guys want to do that?
[14:20] <larsu> kenvandine: you're not the only ones ...
[14:20] <tedg> I'm prone to just remove the autotest target
[14:20] <seb128> larsu, ok
[14:21] <larsu> kenvandine, seb128: should we stick with the qconf name even though it's a new project?
[14:21] <seb128> larsu, I can see lot of free icecream in your futur, if that helps ;-)
[14:21] <didrocks> tedg: I'll let sil2100 handling it as he followed that up more closely than I :)
[14:21] <kenvandine> dednick, i see you're assigned the bug on the messaging menu not showing missed SMS, is that by chance what you're working on?
[14:21] <larsu> seb128: \o/
[14:21] <kenvandine> dednick, that bug is really bugging me :)
[14:21] <larsu> seb128: ice creame *always* helps
[14:21] <seb128> ;-)
[14:21] <dednick> kenvandine: i was just about to start work on it
[14:21] <seb128> larsu, I've no strong opinion on the name
[14:21] <kenvandine> dednick, great... thanks!
[14:21] <seb128> larsu, you can as well call it gsettings-qt
[14:22] <seb128> or whatever works for you
[14:22] <sil2100> tedg: if it's a regression in the actual code, then I would guess that the best way is to just comment out the autotest target
[14:22] <kenvandine> seb128, larsu: i'd prefer that to dconf-qt
[14:22] <kenvandine> which seemed weird to me
[14:22] <sil2100> tedg: since this would mean that this test will anyway fail everywhere - not just when 'we' are building the package
[14:22] <sil2100> tedg: so I'm +1 for the autotest-removal/commenting-out
[14:22] <larsu> kenvandine: ya, and this is not using dconf (directly) anymore :)
[14:23] <sil2100> tedg: with some comment on top to know what's up
[14:23] <larsu> seb128: yeah I thought about that, but I don't know how the sdk people feel about having a GSettingsQt module
[14:23] <sil2100> dednick: is there a bug for that regression filled in?
[14:23] <kenvandine> larsu, is it providing both a Qt library and QML bindings?
[14:23] <larsu> kenvandine: only qml for now. Do you need a qt library?
[14:23] <kenvandine> nope
[14:24] <kenvandine> just checking
[14:24] <larsu> ya, I figured I can add it when the need arises
[14:24] <kenvandine> you could name it gsettings-qml then... unless we want to use gsettings-qt with the thought that someday we'll add a lib too
[14:24] <dednick> sil2100: yes
[14:24] <kenvandine> and the binary package will be something like qtdeclarative-gsettings1.0
[14:25] <larsu> and the import line just "import GSettings 1.0"?
[14:25] <kenvandine> larsu, i'd vote for gsettings-qt in case we add a lib later
[14:25] <kenvandine> larsu, yup
[14:25] <larsu> okay
[14:25]  * kenvandine doesn't mind seeing the "G" there
[14:25]  * larsu whips out sed
[14:25] <kenvandine> and there is already QSettings i think
[14:26] <tedg> "Say hello to my little string parser!"
[14:26] <dednick> sil2100: i think it's the same issue anyway. https://bugs.launchpad.net/touch-preview-images/+bug/1191822
[14:26] <ubot2> Ubuntu bug 1191822 in touch-preview-images "After a wireless network is selected the list of networks says "Empty!"" [High,In progress]
[14:27] <larsu> dednick: can you please add the indicator projects on these bugs? I don't get notified about any messages bugs because they're always filed against touch-preview-images only
[14:27] <tedg> sil2100, https://code.launchpad.net/~ted/indicators-client/tests/+merge/170633
[14:27] <dednick> larsu: sure
[14:28] <larsu> thanks :)
[14:40] <sil2100> tedg: commented!
[14:41] <sil2100> tedg: by 'I guess they shouldn't fail' I mean that they'er testing the functionality, and even if they fail that's a different issue, and we need integration tests being ran if possible - so I would like them to be installed
[14:42] <sil2100> tedg: since they don't block package builds
[14:42] <tedg> sil2100, But if we're going to ignore the results, why waste the CPU time?
[14:56]  * xnox is in favor of running yet ignoring results from known_to_fail testsuites. Expected Fail, is there for a reason.
[15:01] <sil2100> tedg: ok, so at least leave the autopilot tests in
[15:02] <sil2100> tedg: so if you could re-add the indicators-client-autopilot.install it would be all-ok
[15:02] <tedg> sil2100, K
[15:11] <didrocks> sil2100: rmadison says that gtk+3.0 3.8.2-0ubuntu6 is now published in proposed
[15:11] <didrocks> sil2100: so rebuilding should be fine
[15:12] <sil2100> o>
[15:13] <sil2100> Should I rebuild everything, or just the ones that failed in the PPA?
[15:15] <tedg> sil2100, Updated: https://code.launchpad.net/~ted/indicators-client/tests/+merge/170633
[15:16] <sil2100> tedg: thanks! Approved that already
[15:16] <tedg> sil2100, Fast!  :-)
[15:17] <sil2100> ;p
[15:17] <seb128> jbicha, hey, did you notice that abiword is stucked in proposed for 3 weeks? do you plan to rebuild pyabiword for the library rename? (or ask for it to be dropped since it seems that's what Debian did)
[15:19] <didrocks> seb128: I would say just the one which failed
[15:19] <didrocks> sil2100:
[15:19] <didrocks> ^
[15:19] <didrocks> sil2100: there is no ABI breakage, nothing :)
[15:19] <sil2100> I guess ;)
[15:22] <jbicha> seb128: it is possible to get pyabiword to build with the new abiword but the bigger problem is that pyabiword doesn't work & is unmaintained so I filed bug 1187435
[15:22] <ubot2> Launchpad bug 1187435 in pyabiword (Ubuntu) "Please remove pyabiword" [Undecided,New] https://launchpad.net/bugs/1187435
[15:23] <sil2100> didrocks: just out of curosity - when we're publishing things for daily-release to saucy, the source packages are pushed to the archive and rebuilt - yes?
[15:23] <didrocks> sil2100: hum, not sure to get you :)
[15:24] <didrocks> ah, the question is what publication does? :)
[15:24] <seb128> jbicha, can you get it to build?  the sugar stuff was useful, I'm not sure I'm comfortable deciding to just delete those, I would prefer to let whoever care fixes those or drop if needed
[15:27] <jbicha> except that pyabiword has been broken for quite a while and I haven't found a distro that's managed to get it to work
[15:27] <jbicha> that sugar stuff doesn't work because pyabiword is quite broken
[15:28] <jbicha> bug 1020074
[15:28] <ubot2> Launchpad bug 1020074 in pyabiword (Ubuntu) "Segfault when importing abiword" [Undecided,New] https://launchpad.net/bugs/1020074
[15:29] <jbicha> it's still in unstable so if anyone does figure it out, the packages can be easily re-added to Ubuntu
[15:30] <sil2100> didrocks: yes ;)
[15:30] <didrocks> sil2100: so, no rebuild involved: it's a binary copy from the ppa to distro
[15:31] <didrocks> sil2100: we have to do that, otherwise, we can't ensure that what we tested is what the user will get (in case it's rebuilt in a different order for instance)
[15:31] <didrocks> with different build-deps
[15:31] <didrocks> sil2100: the last paragraph in https://wiki.ubuntu.com/DailyRelease/StackPublish#Copy_to_distro for more details
[15:31] <sil2100> ACK, thanks!
[15:31] <didrocks> yw :)
[15:31] <didrocks> sil2100: the only thing that the archives will try to rebuild is if we didn't copy all archs
[15:32] <didrocks> (like armhf still building, but we force the publication meanwhile)
[15:32] <didrocks> or no powerpc binary because of some build-deps: it will try to build them and go into build-dep waiting mode ;)
[15:36] <sil2100> didrocks: hmm, strange thing I see
[15:41] <sil2100> didrocks: indicator-power failed because of python not getting installed, even though it's in Recommends for libglib2.0-dev
[15:41] <sil2100> didrocks: aren't we're installing recommends by default?
[15:42] <sil2100> didrocks: also, I think it's not a good idea to serve an /usr/bin binary with a crucial dependency in recommends
[15:42] <Laney> not in package builds
[15:42] <sil2100> Since /usr/bin/gdbus-codegen fails
[15:42] <sil2100> So hm, basically by installing libglib2.0-dev we can't use all the tools?
[15:43] <Laney> got a link?
[15:43] <sil2100> Because indicator-power uses libglib2.0-dev to get its /usr/bin/gdbus-codegen, but can't use it becaue normally python is not installed
[15:43] <sil2100> Laney: yep, here:
[15:43] <sil2100> https://launchpadlibrarian.net/142892077/buildlog_ubuntu-saucy-armhf.indicator-power_12.10.6daily13.06.20-0ubuntu1_FAILEDTOBUILD.txt.gz
[15:44] <sil2100> (all archs had the same thing)
[15:44] <sil2100> python is in recommends but is not installed
[15:44] <sil2100> (libglib2.0-dev recommends python)
[15:44] <Laney> I doubt that is your problem
[15:44] <Laney> laney@iota> dpkg -S /usr/bin/python                                                                                                        ~
[15:44] <Laney> python-minimal: /usr/bin/python
[15:45] <didrocks> sil2100: we do, let me have a look
[15:45] <didrocks> ah Laney is looking :)
[15:45] <sil2100> hmmm
[15:45] <sil2100> Ok, so I wonder what's up
[15:45] <didrocks> (sorry, was in a discussion)
[15:46] <sil2100> didrocks: ;)
[15:47] <sil2100> Laney: do you have any ideas on what can be wrong then?
[15:47] <Laney> sec
[15:50] <Laney> sil2100: I guess the advice is to build-dep on python, indeed
[15:52] <sil2100> Laney, didrocks: so we don't fetch those recommends in builders indeed? Ok, so let's do that, wonder how this was passing before
[15:52] <Laney> something must have changed its Depends
[15:52] <didrocks> sil2100: that's possible, I'm sure quite sure if we do it. But anyway, if we rely on python, we should dep on it
[15:54] <Laney> package builds are supposed to be minimal so they don't install stuff they don't need to (recommends) :-)
[15:54] <sil2100> didrocks: I can quick fix it and add the python dep to the indicator-power...
[15:54] <Laney> that's right
[15:55] <sil2100> Laney: well, if gdbus-codegen needs python then I guess it should depend instead of recommends :D
[15:55] <didrocks> sil2100: sounds good to me!
[15:56] <Laney> It's not unheard of for less used features to be moved to recommends
[15:56] <sil2100> Laney: oh, ok, then that makes sense
[15:56] <Laney> A lot of glib-using packages will not need python
[15:57] <Laney> But it's a bit unfortunate that you are quite likely to get python installed in some other way and so not notice this
[15:57] <sil2100> Indeed
[15:57] <Laney> oh well, the fix is easy
[15:57] <sil2100> Laney: thanks for the check!
[15:57] <Laney> np
[15:59] <sil2100> https://code.launchpad.net/~sil2100/indicator-power/add_python_dep/+merge/170655
[15:59] <sil2100> Anyone ^
[16:00]  * sil2100 looks at didrocks and Laney with teary eyes
[16:01] <Laney> i don't have powers for that :(
[16:01] <didrocks> sil2100: how can I resist then? approved! :)
[16:01] <sil2100> ;)
[16:02] <jbicha> stgraber would like to have users who have been using the gnome-session-fallback session to automatically be transitioned to the new gnome-session-flashback but that looks pretty tricky
[16:03] <jbicha> session-migration won't help since we want to change user's configurations *before* they log in
[16:04] <Laney> does the fallback session go away?
[16:04] <jbicha> it was just renamed; it's still powered by gnome-panel 3
[16:05] <Laney> symlinks?
[16:07] <jbicha> that might work but I'm thinking the system would store that it was running gnome-session-fallback instead of -flashback so we'd have to keep those symlinks forever?
[16:07] <Laney> probably
[16:07] <Laney> do you know where it's stored?
[16:08] <seb128> have it migrated after login then you can drop it after the lts
[16:08] <seb128> didrocks did that when they renamed the gnome/ubuntu/unity sessions
[16:09] <jbicha> not exactly, there's ~/.dmrc; there's /var/lib/AccountsService/users/; there's gsettings org.gnome.session session-name; and even lightdm.conf has something too
[16:09] <seb128> just keep using the old filename
[16:09] <seb128> that's easier
[16:09] <seb128> who cares, that's just a name on disc
[16:09] <seb128> you can change the description
[16:10] <Laney> I think that might be easiest
[16:10] <didrocks> and safest
[16:11] <didrocks> we had a 3 year-plan transition with lightdm and did it :)
[16:11] <mlankhorst> Laney: http://buienradar.nl/bliksem-1-uur .. shall I bike today?
[16:11] <mlankhorst> lightning stuff is near me atm
[16:11] <Laney> seems to be moving away :P
[16:12] <mlankhorst> yeah.. but not sure if nothing else is coming or not..
[16:12] <Laney> bike to your local climbing centre and do some fun bouldering
[16:12] <mlankhorst> >:D
[16:12] <mlankhorst> bbs
[16:12] <mlankhorst> food
[16:41] <mlankhorst> and the rain is gone, I'm off to bike, muahah
[16:48] <desrt> seb128: what do i get blamed for today?
[16:59]  * didrocks waves good evening
[17:00] <Laney> desrt: He was searching for people to blame for https://launchpadlibrarian.net/142386450/buildlog_ubuntu-saucy-armhf.gtk+3.0_3.8.2-0ubuntu4_FAILEDTOBUILD.txt.gz
[17:25] <larsu> kenvandine: let's try this again: https://code.launchpad.net/~larsu/+junk/gsettings-qt
[17:25] <larsu> kenvandine: I'll make a proper project out of it after lunch
[17:26] <larsu> compile with qmake for now, use like so: http://paste.ubuntu.com/5784344/  (run from $top_srcdir with `qmlscene -I . test.qml`)
[17:26] <larsu> seb128: ^^
[17:26] <larsu> kenvandine: most importantly: won't work without the qt5 edgers ppa because of lp #1187746
[17:26] <ubot2> Launchpad bug 1187746 in qtdeclarative-opensource-src (Ubuntu) "Please provide fix for QTBUG-31226" [Undecided,Fix committed] https://launchpad.net/bugs/1187746
[17:35] <seb128> larsu, danke
[17:37] <seb128> desrt, gtk tests fail on arm, it looked at first that it was getting corrupted resources out of gresources ... but that was only a first, I extracted the css and it's valid
[17:45] <pitti> seb128: hm, the new libgpgme11 pulls in gnupg-agent gnupg2 libassuan0 libksba8 pinentry-gtk2
[17:45] <pitti> seb128: do we really want all that? it's quite sizable, and pulls in old gtk2 stuff
[17:46] <seb128> pitti, I saw yes, that's coming from Debian ... we should probably demote some recommends to suggests
[17:46] <seb128> well, at least pinentry-gtk2
[17:46] <seb128> the other ones seem fine to me
[17:47] <pitti> --no-install-recommends still pulls these in
[17:47] <Laney> it's Depends: gnupg2 probably
[17:47] <seb128> gnupg-agent depends on pinentry-gtk2 | pinentry-curses | pinentry
[17:48] <seb128> gnupg2 depends on gnupg-agent
[17:50] <Laney> they switched from gnupg to gnupg2
[17:50] <Laney> afaict
[17:50] <Laney> the former not having as many dependencies
[17:50] <pitti> perhaps our gnome-keyring bits should Provide: gnupg-agent?
[17:50] <Laney> (or features)
[17:51] <Laney> it seems to call the binary
[17:51] <pitti> meh; not good
[17:52] <Laney> sorry, I mean gpg not gpg-agent
[17:52] <Laney> so maybe Provides would work to cut off that chain
[17:52] <Laney> ah, but the dep is versioned from gnupg2
[17:56] <seb128> Laney, pitti: we should maybe demote the depends on gnupg-agent to a suggest
[17:56] <seb128> or | gnome-keyring
[17:56] <Laney> Don't know what would be correct there
[17:57] <Laney> Easiest might be to go back to gnupg from gpgme
[17:57] <Laney> maybe defining a virtual package would be a good idea
[17:58] <Laney> anyway, i'm off, let's think about that tomorrow
[17:58] <Laney> night
[18:02] <seb128> Laney, 'night
[18:02] <seb128> it seems backward to stay on an outdated gnupg version though
[18:09] <seb128> desrt, I guess you don't have a saucy/armhf device with you?
[18:19] <Saviq> larsu, ooh, so slim was enough?
[18:21] <larsu> Saviq: hm?
[18:21] <larsu> oh, gsettings-qt. ya :)
[18:21] <Saviq> larsu, lp:~larsu/+junk/gsettings-qt
[18:21] <larsu> yep, qqmlpropertymap ftw
[18:21] <kenvandine> larsu, cool, i'm on it
[18:21] <larsu> it does what 90% of dconf-qt codebase used to do
[18:22] <larsu> the remaining 10% is now gsettings
[18:22] <larsu> so there wasn't that much left :)
[18:23] <Saviq> larsu, one thing I was recently thinking (although this can be a wrapper around gsettings-qt)
[18:23] <Saviq> larsu, is that we might need to have different defaults per-form-factor
[18:23] <larsu> Saviq: note that this only contains qml bindings right now (not really usable from qt)
[18:24] <Saviq> larsu, yeah, who cares about qt these days ;P
[18:24] <Saviq> larsu, and not only different defaults, but different values
[18:24] <larsu> Saviq: interesting. I don't think dconf has facilities for this kind of thing, but you'd need to talk to desrt about it
[18:25] <larsu> well, I guess we might be able to get something working with relocatable schemas
[18:25] <Saviq> larsu, yeah, I think we might be fine with a flat key/value, assuming we'll be smart at e.g. using separate setting databases for different form factors, or a prefix / suffix
[18:26] <larsu> Saviq: yeah, I'll talk to desrt about it
[18:27] <Saviq> larsu, there's some UX questions here, obviously (i.e. if you add an app to the launcher while in phone mode, should it get added in desktop mode as well?)
[18:27] <Saviq> larsu, and it's probably a per-setting question, too
[18:27] <Saviq> larsu, but yeah, just food for thought for now
[18:30] <larsu> Saviq: right, we will probably have to support both for different keys (i.e. you might want different launcher items and backgrounds on your phone, but share lists of locations that you'd like to get the time in your clock for)
[18:30] <larsu> ooh, this is a tricky subject
[18:31] <Saviq> larsu, indeed ;)
[18:32] <larsu> kenvandine, Saviq: I put lgplv3 in the headers, is that correct?
[18:32] <Saviq> larsu, AFAIK, yes
[18:32] <kenvandine> yes
[18:32] <larsu> cool, thanks
[18:33] <larsu> kenvandine: I pushed it into lp:gsettings-qt as well
[18:34] <kenvandine> cool
[18:34] <kenvandine> i'm adding the cmake magic now
[18:35] <larsu> awesome! Thanks
[18:39] <kenvandine> larsu, so what do i need to build this with?
[18:40] <tkamppeter_> larsu, hi
[18:41] <kenvandine> larsu, like the pkgconfig packages?
[18:41] <kenvandine> gio obviously
[18:41] <larsu> kenvandine: gio and qtqml
[18:41] <larsu> tkamppeter: hi
[18:42] <kenvandine> oh, you have it setup for qmake already
[18:42] <larsu> kenvandine: and it would be cool if that qt patch landed in saucy...
[18:43] <larsu> kenvandine: yep, qmake && make in the GSettings dir builds it
[18:43] <Saviq> kenvandine, larsu QtCore, I imagine, too
[18:43] <larsu> (on my machine *cough*)
[18:43] <Saviq> larsu, btw http://qt-project.org/doc/qt-5.0/qtcore/qstringlistmodel.html
[18:43] <kenvandine> do we really need to use cmake for it?
[18:44] <larsu> kenvandine: nope.
[18:44] <Saviq> larsu not sure it's the right place, but it would be nice to wrap the stringlist in a model
[18:44] <larsu> Saviq: which string list? The list of keys?
[18:44] <Saviq> larsu, no, when a key is a list of values
[18:44] <larsu> ah!
[18:45] <larsu> good point.
[18:45] <Saviq> erm, when a value of a setting is a list
[18:45] <larsu> ya
[18:45] <Saviq> larsu, so that only minimal changes are sent to the app (obviously this needs smarts to reduce the changes to a minimum)
[18:45] <larsu> so that code is almost 1:1 what dconf-qt did
[18:45] <Saviq> larsu, sure, but who's to say dconf-qt did everything right? ;)
[18:46] <larsu> Saviq: ha, I wasn't saying that, just that I haven't put much thought into that part :P
[18:46] <Saviq> larsu, sure
[18:46] <Saviq> kenvandine, well, CMake is the company policy, right?
[18:46] <larsu> kenvandine: this is why I asked for it ^^
[18:46] <Saviq> kenvandine, shouldn't be many more lines than the QMake
[18:46] <kenvandine> Saviq, i haven't seen that documented anywhere
[18:47] <larsu> Saviq: is the stringlistmodel something that we need asap?
[18:47] <Saviq> larsu, no
[18:47] <Saviq> larsu, it can even be an addition to the API
[18:47] <larsu> hm, well it would be an API break, right?
[18:47] <Saviq> larsu, i.e. you can get it as a QStringList, but if you want a model - here!
[18:47] <larsu> right, but the default would always be a stringlist?
[18:47]  * larsu is unsure which is more qmly
[18:48] <Saviq> larsu, model is definitely more qmly
[18:48] <Saviq> larsu, e.g. items in launcher, you want to move them, not replace the whole set with a new one
[18:48] <larsu> Saviq: ok. I'll try to add that soon then and make it the default for lists
[18:48] <kenvandine> i would prefer a model
[18:48] <Saviq> larsu, but because gsettings doesn't give you such data, you need to extrapolate
[18:49] <larsu> Saviq: I need to keep a copy of the valus around anyway, so that wouldn't be a problem
[18:50] <Saviq> larsu, yeah, but to actually minimize the amount of changes (i.e. grouping the items that were moved together), moving from the end instead of from the beginning
[18:50] <Saviq> larsu, and then because it's a whole new list, you need to add / remove, too
[18:50] <Saviq> larsu, but initially I'd be fine with a model that's reset() every time
[18:50] <Saviq> larsu, at least it'd be a model straight away, just the behaviour would improve
[18:51] <larsu> Saviq: that's a good first step I agree. Everything else will then be an optimization without API breakage
[18:51] <larsu> right :)
[18:52] <Saviq> kenvandine, indeed there's no mention of CMake on the preferred tech wiki
[18:52] <Saviq> kenvandine, but then qmake is being phased out, afaik
[18:52] <kenvandine> Saviq, right
[18:53] <Saviq> tvoss, you'll know - is CMake policy now?
[18:53] <kenvandine> we still have quite a bit of qmake
[18:53] <tvoss> Saviq, as much as possible, yes
[18:53] <kenvandine> and is certainly much simpler to deal with
[18:53] <tvoss> kenvandine, do you see an issue transitioning over to cmake?
[18:54] <Saviq> kenvandine, that probably depends on how complicated your project is
[18:54] <kenvandine> it's quite simple and already uses qmake
[18:54] <kenvandine> i'd prefer just leave it as is
[18:54] <Saviq> /we need CMake templates for QML plugins
[18:55] <kenvandine> what we have now in cmake is a real mess
[18:55] <kenvandine> very inconsistent and copy and paste cruft
[18:55] <kenvandine> our qmake stuff is much cleaner, imo
[18:55] <kenvandine> i'd rather us have standard templates for cmake
[18:55] <Saviq> +1, we need common CMake modules
[18:55] <kenvandine> before switching stuff
[18:55] <Saviq> kenvandine, seen http://blog.qt.digia.com/blog/2013/05/31/qbs-1-release/ btw?
[18:56] <tkamppeter> larsu, seb128 told me some days ago that bug 959195 occurs for him in Saucy.
[18:56] <ubot2> Launchpad bug 959195 in indicator-printers (Ubuntu) "65 cups notifier processes running" [Undecided,Fix released] https://launchpad.net/bugs/959195
[18:56] <kenvandine> Saviq, yup
[18:56] <tkamppeter> larsu, at lease the symptoms of it.
[18:57] <kenvandine> Saviq, it looks very interesting
[18:57] <larsu> tkamppeter: hm, I fixed that two cycles ago... thanks for pointing me to it, I'll talk to him
[18:57] <Saviq> kenvandine, yeah
[18:57] <kenvandine> yet another reason to not just up and switch projects :)
[18:57] <Saviq> ;)
[18:57] <kenvandine> don't get me wrong... i don't like qmake
[18:57] <kenvandine> but i surely prefer qmake to cmake :)
[18:57] <kenvandine> and all build systems suck
[18:59] <seb128> larsu, talk to me!
[18:59] <tkamppeter> larsu, thanks.
[19:03] <larsu> seb128: I was about to when my connection broke. Are you still seeing the issue?
[19:03] <seb128> larsu, no
[19:03] <larsu> seb128: the problem was that indicator-printers created too many cups subscriptions
[19:03] <larsu> I fixed that long time ago, so I can't imagine them being stray processes from back then on your machine
[19:03] <seb128> larsu, I apparently had a buggy /etc/cups/subscriptions.conf* in there ... I had the issues for months, I never bother asking
[19:03] <dobey> meh. Alt_L is working correctly in gnome-shell guest session. wtf.
[19:04] <seb128> larsu, it's gone since I followed your comment and did a rm /etc/cups/subscriptions.conf*
[19:04] <seb128> larsu, so I guess nothing was cleaning buggy configs left behind?
[19:04] <larsu> seb128: ah! That's probably it then. I couldn't just clean the subscriptions in an upate to indicator-printers because there might have been "real" subscriptions in there
[19:05] <larsu> seb128: it was a bit of a screw-up on my part. It should be fixed for you now that you cleaned the conf, let me know if it happens again
[19:06] <seb128> larsu, thanks, I'm pretty sure it's fixed though, it was consistent for some months and it stopped happening since yesterday
[19:06] <seb128> larsu, I just had been too lazy to ask about it, I didn't have any noticable issue due to those processes
[19:06] <seb128> ;-)
[19:07] <dobey> weird. found something that was set to "Mod2+Alt_L" :(
[19:07] <seb128> dobey, buggy user config?
[19:07] <larsu> seb128: ya, they are very lightweight (just a small C program which does nothing but wait for a job to happen in CUPS). It messes with ps output though ;)
[19:07] <dobey> seb128: what the heck is Mod2?
[19:08] <seb128> dobey, not sure, r-ctrl is "mod2+control r" for me (that's what I put the HUD on)
[19:10] <sarnold> mod2 can be set via xmodmap; run xmodmap without arguments to see what it is currently set to
[19:11] <dobey> i'm quite certain i've never changed that keybinding
[19:13] <dobey> mod2 is numlock. that doesn't even make any sense :)
[19:14] <sarnold> same here. my laptop has no numlock. That I know of.. :)
[19:14] <seb128> dobey, well it's weird, if you go to the keybinding panel and try to set e.g the hud to r-ctrl or r-shit you will get "mod2+rctrl/shift"
[19:15] <dobey> seb128: the weird thing is that this was the "Switch applications" keybinding
[19:16] <seb128> weird
[19:16] <dobey> seb128: I would never have set it to Alt_L, or Alt_L+NumLock. neither of those makes sense. so i wonder how it got changed
[19:19] <seb128> dobey, buggy software but good luck to figure which one...
[19:23] <dobey> yeah :-/
[19:30] <jbicha> attente: indicator-keyboard doesn't work for me
[19:30] <larsu> Saviq: is there a QVariantListModel as well? I'm unsure what I should do about non-string lists
[19:31] <larsu> doesn't look like it from a quick search in the docs
[19:31] <attente> jbicha, i just updated the manifest file and uploaded a new version to the ppa
[19:31] <larsu> well I guess I can always make my own
[19:32] <jbicha> attente: yeah I built it from your bzr branch; I'm on saucy with no other ppas except your i-keyboard one
[19:32] <Saviq> larsu, yeah, there isn't anything built in
[19:33] <Saviq> larsu, but the API is small enough, I'd say
[19:33] <attente> jbicha, the fix i pushed was just 5 minutes ago
[19:33] <attente> jbicha, at least i believe that's what the problem is
[19:33] <Saviq> larsu, pete-woods did a QVariantListModel by almost c&p QStringListModel
[19:34] <Saviq> larsu, on one hand that might pose licensing issues, but on the other if you'd do it yourself, you'd end up with 99% same code, probably..
[19:35] <attente> jbicha, it's at revision 57, i'm not sure if that's the one your local build is based on
[19:35] <larsu> Saviq: heh, I'll have a look, thanks!
[19:36] <jbicha> attente: ok, it works after I pulled and rebuilt
[19:36] <jbicha> I'm not sure that it should say EU for English (United States) though
[19:38] <jbicha> I don't know if it's my sbuild or if it's bzr-builddeb but it doesn't like that maintainer doesn't have a @ubuntu.com email
[19:39] <seb128> how so?
[19:39] <Laney> DEBEMAIL= bzr bd -S
[19:40] <seb128> right, I never got why it behaves differently if you DEBEMAIL is @ubuntu.com
[19:40] <jbicha> it's easy to add an Ubuntu team as maintainer
[19:40] <seb128> especially that I usually hit the issue on ubuntu specific sources, like unity stack
[19:40] <seb128> it's annoying
[19:41] <attente> what email should we use instead?
[19:41] <seb128> run update-maintainer
[19:41] <seb128> it will set the maintainer to ubuntu-devel group
[19:42] <seb128> (that's a script in ubuntu-dev-tools ... if you don't have the package install it, it has handy scripts)
[19:42] <attente> seb128, ok
[19:42] <seb128> but that warning is just annoying
[19:43] <seb128> it's a check meant to make sure we update the maintainer info when we take  a debian source and change it
[19:43] <seb128> so they don't get bugs or emails about our buggy changes ;-)
[19:44] <attente> you'd think it would be ok with a canonical email though...
[19:45] <seb128> not sure if the issue is because we use "0ubuntu<n>" revision style
[19:45] <seb128> we should just claim the namespace and use -1 ;-)
[19:47] <attente> jbicha, about the EU thing, we were trying to be smart about how it determines the abbreviation
[19:48] <attente> but if it doesn't work for one of the most popular layouts... maybe it's not so smart after all...
[19:49] <seb128> heh
[19:49] <seb128> it doesn't work for Fr either :p
[19:49] <seb128> it uses FA
[19:49] <seb128> is that because for "French Alternative"?
[19:49] <attente> seb128, that's actually a bug on my part :P
[19:49] <jbicha> https://wiki.ubuntu.com/LanguageAndText?action=AttachFile&do=get&target=gsd-indicator.png
[19:49] <seb128> ah... ;-)
[19:50] <attente> seb128, it sees the cedilla and isn't sure how to parse the format
[19:50] <attente> so it thinks that the 'a' that follows is a part of a new token
[19:50] <attente> lol
[19:50] <seb128> hehe
[19:50] <seb128> not sure how the old code and GNOME are doing
[19:51] <seb128> is that info in iso-codes?
[19:51] <seb128> I never really looked at the logic of other codebases for that...
[19:51] <jbicha> I'd expect something like En or US
[19:51] <attente> yeah, gnome is just using the iso code, iirc
[19:51] <attente> but we're trying to do some weird mish-mash of language and country
[19:52] <attente> which isn't going to work
[19:52] <seb128> hum, isn't GNOME trying to do that as well?
[19:52] <seb128> that topic is complex :/
[19:52] <seb128> I would say to just go with the simple rule that was used so far if the mix is a can of worm
[19:53] <seb128> well, if you can get it right and that makes mpt happy that's bonus ;-)
[19:53] <seb128> we can as well try to push it like that, deal with bug reports, and if that doesn't fly for users then change our strategy and go back to a more classic way...
[19:54] <jbicha> the wiki does say "first two capital letters from the display name" which does come out to EU
[23:25] <rickspencer3> hmmm, software-center is not launching for me :
[23:57] <RAOF> @pilot in
[23:57] <meetingology> RAOF: Error: "pilot" is not a valid command.