/srv/irclogs.ubuntu.com/2012/11/19/#ubuntu-unity.txt

=== fenris is now known as Guest70212
=== Guest70212 is now known as ejat
=== fenris is now known as Guest50405
=== mmrazik is now known as mmrazik|otp
=== mmrazik is now known as mmrazik|otp
didrockshey mmrazik|otp, I've a small question :)09:56
=== mmrazik|otp is now known as mmrazik
mmrazikdidrocks: whats up?>09:56
didrockswhat was the fix from going from ~100 tests failing like https://jenkins.qa.ubuntu.com/job/ps-unity-autopilot-trunk/label=master,machine_name=dx-autopilot-intel/382/testReport/09:56
didrocksto ~3009:56
didrocksin the 383 run09:56
mmrazikdidrocks: we are running the autopilot tests from Utah/script. We need to export DISPLAY etc there as it is not executed directly from the session09:57
mmrazikdidrocks: it turned out that in such environment autopilot is unable to use compiz-config09:57
mmrazikand there were tests which are e.g. changing the HUD invcation key via compiz-config09:57
didrocksmmrazik: hum, I'm afraid thta if it's not running from the session, you will have more test failing because of that09:57
didrockslike you miss the real dbus session config09:57
didrocksand other small env issues09:58
mmrazikdidrocks: might be the case. I was briefly talking with thomi about it yesterday. He has an idea to run it from  /etc/xdg/autostart09:58
didrocksmmrazik: yeah, that would be way better to test on a real config09:58
mmrazikdidrocks: but so far we seems to be fine with regard to environment variables09:59
didrocksmaybe that's why you don't have the migration of keys in compiz09:59
=== mmrazik is now known as mmrazik|lunch
mvoquick question: where does unity store the launcher items ? is that dconf and com.canonical.unity.launcher.favorites?10:54
mhr3yep10:59
=== _salem is now known as salem_
mvothanks! I11:05
mvomy next question would be: is it somehow possilbe (e.g. via setting of the plugin load path or something) to run unity from the bzr build dir? i.e. (cd build ; ./bin/unity) ? it runs for me, but appears to be not picking up my changes in e.g. launcher11:06
seb128mvo, looks like http://unity.ubuntu.com/getinvolved/development/unity/ has some wrappers/details on how to do that11:12
seb128mvo, but unity is a compiz plugin, you might need to copy the .so in the .compiz dir or something11:13
mhr3mvo, i suggest running the component you're interested in by itself (./launcher/launcher etc)11:24
seb128mhr3, hey11:28
seb128mhr3, https://bugzilla.gnome.org/show_bug.cgi?id=68605911:29
ubot5Gnome bug 686059 in general "run-dialog: Avoid double forking; it breaks "pkexec"" [Normal,Resolved: fixed]11:29
seb128mhr3, is unity double forking as well?11:29
mhr3seb128, we're just calling glib, which doesn't use DO_NOT_REAP_CHILD11:31
didrocksMirv: I guess you did notice that the tests don't pass anymore on arm* making the build failing11:31
* didrocks prays for ci having finally an arm build11:32
seb128mhr3, that's the fix for gnome-panel: http://git.gnome.org/browse/gnome-panel/commit/?id=76acc5b955b214420a32c827db433ff5ab136c6f11:32
mhr3seb128, see the comment:11:32
mhr3It still an incomplete fix for gnome-panel though, as you might have .desktop11:32
mhr3files using "Exec=pkexec foo" and starting such an application via the menu11:32
mhr3still fails.11:32
seb128mhr3, see the further comments :p11:33
seb128mhr3, or the dummy_child_watch hacks in the patch11:33
mhr3can't say i like it really11:34
mhr3it they don't want to reap, do it directly in glib11:34
mhr3seb128, and iirc now if the parent process does something bad it can burn the child as well11:37
seb128mhr3, hum, k11:38
seb128mhr3, you should open a glib bug about it ;-)11:38
mhr3i'm not really sure what's the correct behaviour, reaping is fine imo, pkexec is being stupid11:39
mhr3but then again, i'm sure they have reason why they require non-init ppid11:40
seb128well, we need to address it one way or another, as you say I guess they have reason and will not change pkexec11:42
seb128so we either need to do the workaround GNOME did (they did the same in shell btw) or to get glib "fixed"11:42
seb128we can't just sit on "ok, users can't run those under unity"11:42
mhr3seb128, if the issue was easy to fix it'd be fixed long time ago :)11:44
mhr3afaict the thing panel/shell are doing aren't exactly safe11:44
seb128mhr3, which is why I pointed the workaround :p11:44
Mirvdidrocks: yes, that was noticed, although no solution yet. it's good indeed that those are now run for arm11:44
seb128mhr3, what are you concerned about? is that a theorical issue or a practical one?11:44
=== mmrazik|lunch is now known as mmrazik
mhr3as i said the parent dieing might affect the child when it's not reaped11:45
mhr3so for shell it's not really an issue, if shell dies you're screwed eitherway11:46
mhr3but if apps lens dies, it's no biggie, you wouldn't like if that killed all the apps you started11:46
didrocksMirv: it's in the ppa, so in trunk11:50
didrocksMirv: it's not run for arm on jenkins :/11:50
seb128mhr3, hum, right11:51
mhr3seb128, then again, it's not like every signal was propagated, it might be just sigterm11:52
mhr3still, doing `killall unity-applica...` and that killing all your apps isn't exactly nice11:53
seb128mhr3, not acceptable indded11:53
mhr3seb128, i guess gnomies expect that people don't killall the panel or something :P11:54
ricotzmhr3, hi, interesting i should remember that ;) since plank suffers from this problem too11:55
mhr3seb128, i think talking to davidz will be best, seems he'd know about pkexec :)11:57
seb128mhr3, yeah11:57
=== dandrader is now known as dandrader|afk
=== mmrazik is now known as mmrazik|otp
=== dandrader|afk is now known as dandrader
=== mmrazik|otp is now known as mmrazik
larsume4oslav, hi :)14:19
me4oslavso me and larsu we're just talking about system settings > printing. So, what we have now officially as a design spec is this by mpt: https://wiki.ubuntu.com/Printing14:20
me4oslavI expanded it a bit: https://picasaweb.google.com/100530892038948253747/SystemsSettingsPrinting14:21
larsume4oslav, I really like the default/shared thing in the list box14:21
larsuI wonder how much room it would take away from the printer names, though14:22
larsuin your mockup, it takes 50% of the width14:22
me4oslavyeah, my changes include "Sharing" in the combobox and the "default" and "share" in the list box. And how much space it take is relevant to the number of printer the suer has14:23
me4oslavbut generally that "default|share|printer" box is ~ 1/2 the size of the +- bar below14:23
larsuoh, I mean horizontal space14:24
larsuprinter names can be quite long14:24
larsuI don't mind the header at all :)14:24
larsumpt, I guess that's the reason you put "Set as Default" at the bottom?14:25
me4oslavwell, the listbox vary can be longer if needed :)14:25
larsutrue, and maybe that's not even that much of a problem, now that we have a pretty wide system settings window14:26
mterrydidrocks, just go ahead and reject g-c-c-unity, I'll fix and upload14:26
didrocksmterry: ok, thanks!14:26
larsume4oslav, in general, I much prefer radio buttons over the check item on the right side14:26
larsume4oslav, what do you mean by "sharing type"?14:27
me4oslavlarsu - me too (and that's how the Bootloader design spec uses (the "header")14:27
larsuah, right!14:27
me4oslavabout the sharing - I litterally have no clue about sharing14:28
me4oslavif anyone cantell me how ways of sharing a printer I will update the design14:28
me4oslavI only know network printing sharing and that's it14:28
larsuit's generally just one binary decision: expose this printer to the network or not14:29
me4oslavso, nothing else, but network sharing?14:30
me4oslavperfect, that will make the design much easier14:30
me4oslavso, the user checks "shared"14:31
me4oslavthan select to witch network to share it14:31
me4oslav(that will make the combobox jump from "Printer Status" to "Sharing")14:32
larsuI've never seen that anywhere, but it's an interesting idea14:32
larsui.e. share this printer in my home network but not at work?14:32
mpthi me4oslav14:32
me4oslavoh, hi mpt14:32
me4oslavlarsu - so, a printer can be shared in two or more networks?14:33
larsume4oslav, no, not right now. I just found the idea interesting :)14:34
me4oslavthat would just mean that the radio buttons on my second sketch should become checkboxes and that'll be enough :)14:34
larsuright - but CUPS doesn't support that at all afaik14:34
me4oslavcups doesn't support multiple networks sharing?14:35
larsuno, a printer is either shared, or not14:35
mptlarsu, yes, I put "Set as Default" at the bottom because using radio buttons would eat valuable space, and changing the default is a rare thing14:35
mterrydidrocks, back in NEW14:35
didrocksmterry: will have a look soon. Thanks :)14:35
mptlarsu, or at least, something that usually isn't done more than once a month or so14:36
me4oslavI thought the same, but then again I changed it, because I have "Shared" there too and though - would look weird with just "Shared"14:37
larsumpt, makes sense. It leads to the same weird situation that we have in the session menu though: I don't know what that check mark means :P14:37
larsuwell, not exactly the same: the session menu has both the check mark and a radio button14:37
mptlarsu, true, that's a cost.14:38
me4oslavwell, so we have one compromise or another14:39
me4oslavtaking bit of space or having that checkmark14:39
me4oslavthe question I think should be which one is better when it comes in terms of consistency14:41
me4oslavis the checkmark thing-y used anywhere else?14:41
larsume4oslav, yes, in the session menu14:41
me4oslavright - we have that in the session menu and we have the radio in Bootloader design specs14:42
me4oslavanother way to look at this is we have just "shared|printer" and we have "set as default" + checkbox14:43
me4oslavmpt - hmm, the checkmark overlaps the printer name on the right side?14:44
mptme4oslav, yes, but only for the one that's the default at the moment14:45
me4oslavyeah, so we either eat space with the radio or overlap the default printer name, right?14:45
me4oslavI'm not sure I like that overlapping thing-y, I would prefer having a bit longer printing pane and eating bit of space14:48
me4oslavanyway, apart from default thing-y, mpt larsu - your thoughts on "shared"?14:52
didrocksmterry: do you know if anything changed recently? I see when bzr bd on unity dpkg-source: info: use the '3.0 (quilt)' format to have separate and documented changes to upstream files, see dpkg-source(1)14:52
mptme4oslav, I don't know. On the one hand, it seems to me that turning sharing on/off would be even less common than changing the default. On the other hand, seeing whether sharing *is* on may be valuable for troubleshooting.14:53
didrocksI guess it's because we use the 6.12 tarball from the last SRU14:53
larsume4oslav, I think it should only be the checkbox for now. The same horizontal-space argument applies: I like it in the list box because it gives a good overview over which printers are shared, but it eats horizontal space14:53
didrocksI wonder how the merger is working14:53
didrocksI guess it's pushing different tarballs and not fetching latest one :)14:54
larsumpt, do you think sharing on an per-printer basis makes sense even?14:54
mterrydidrocks, yeah I think that's harmless.  I've seen it too14:54
didrocksyep14:55
mptlarsu, is that not the case currently?15:03
larsumpt, right now you can do both: turn on sharing globally or per-printer15:03
larsuI don't know the exact semantics right now15:03
larsulike, what happens when no printer is marked as "shared" but the global sharing switch is turned on15:04
mpthmm, that's a challenge15:04
larsuI think we should either do one or the other, but not both15:04
=== om26er_ is now known as om26er
mptlarsu, most of the time -- not for hobbyists, but for vast corporate installs -- a printer listed will be one that's shared with you, and you don't have the ability to choose whether it's shared or not, right?15:06
larsumpt, right.15:06
me4oslavwait a sec guys - how does one turn sharing globally?15:06
larsume4oslav, in system-config-printer (the window that starts right now when you click "Printers" in system settings): Server (menu item) / Settings / "Publish shared printers connected to this system"15:08
me4oslavright-y and do we have any of that stuff in the new design specs?15:09
larsunope15:09
me4oslavright and do we need it? What I have in my "sharing pane" is just where to share a printer (in which network)15:11
larsuI don't know -- I just asked mpt about it :)15:13
me4oslavsame story, if we need it, than we have a challenge15:14
me4oslavbut I have no clue if we actually need it15:14
=== dandrader is now known as dandrader|lunch
=== mmrazik is now known as mmrazik|otp
=== mmrazik|otp is now known as mmrazik
=== dandrader|lunch is now known as dandrader
=== francisco is now known as Guest78941
didrocksmterry: g-c-c-u is now accepted in proposed (in universe)17:35
didrocksnot sure if we really need a MIR, as it's code already in the distro, I think we can promote it directly17:35
mterrydidrocks, sure17:36
didrocksmterry: I'm waiting it to migrate to the release pocket (probably tomorrow), not sure what bad things can happen if I promote it if it's copied meanwhile in the release pocket :)17:37
mterrydidrocks, it doesn't conflict with any of the files in g-c-c, but with both installed at same time, you'll have two appearance capplets17:37
mterrydidrocks, so it would be nice to coordinate the seeding of this with an update of g-c-c17:38
didrocksmterry: yeah, I read the code and how you renamed it :)17:38
didrocksmterry: agreed, let's do that tomorrow?17:38
mterrysure17:38
mterrydidrocks, is the 3.6 update ready?17:38
didrocksgreat17:38
seb128didrocks, mterry: not yet but I'm aiming at having it this week17:38
didrocksmterry: hum, I'm not sure, seb128 should know17:38
seb128just keep it in universe until then17:38
mterryseb128, k17:39
didrocksso let's do with that update17:39
Trevinhomterry: updated this https://code.launchpad.net/~3v1n0/bamf/remove-gtk2/+merge/134537 ;)18:47
mterryTrevinho, nice, looking18:48
=== dandrader is now known as dandrader|afk
Trevinhomterry: thanks19:14
=== dandrader|afk is now known as dandrader
=== salem_ is now known as _salem
balloonsthomi, whenever your around ping me if you would :-)22:20
thomiballoons: I'm around now :)22:21
balloonssweet :-)22:21
balloonsSo I'm going to try and put together some resources towards using autopilot to automate some of our manual testcases, and then ubiquity22:21
thomicool :)22:22
balloonsI've been messing around with autopilot and have a couple questions I guess  I can ask you now22:22
thomifire at will :)22:22
balloonsok, so I tried to put together a non-unity testcase and get it to run22:23
balloonsIn digging around, I had success instantiating  UnityTestCase -- but I'm curious if there's something better I can use22:24
balloonslet me give you a little snippet22:26
balloonsthis was the simplest example I could put together: http://paste.ubuntu.com/1371255/22:26
thomione second...22:28
balloonsno worries..22:30
thomiballoons: OK, so if you're trying to create a non-unity test case, you should use autopilot.testcase.AutopilotTestCase22:30
thomiit contains 'self.keyboard' and 'self.mouse' and none of the Unity-specific stuff22:30
balloonsthomi, ok. I figured there was a generic one.. I didn't see it :-)22:31
balloonsso my first goal was to show how to use the keyboard and mouse commands to automate stuff.. then move into introspection22:31
thomiyeah - I need to get the docs re-built and uploaded.22:31
thomiyeah, sounds good22:31
balloonsexcellent. So have you spoken with the xpresser folks post-UDS yet?22:32
balloonsor should I just push some patches to X11.py?22:32
balloons:-p22:32
thomiballoons: what patches? The conclusion was that we should integrate the image matching parts of xpresser into autopilot.... but so far it's not on anyone's immeadiate TODO list...22:33
balloonsthomi, I could add the stuff I want / need.. I already wrote a proof of concept python app to do the X11 stuff I wanted22:34
thomiballoons: I'm not aware that anything is missing - what specifically do you need?22:35
thomiit might already be there :)22:35
balloonsbasically performing checksums, reading pixels, etc.. that's the biggest missing piece from autopilot if I'm remembering correctly22:35
thomiahhh22:35
thomiright, sounds like we need to integrate xpresser sooner rather than later22:35
balloonsyea, if you can't introspect at all, your stuff screenreading22:35
thomiI assumed that for most of the stuff you were doing you'd be able to use the Gtk or Qt supporet22:36
balloonswell, I mean, I think a couple small commits could add everything.. if your open to adding it22:36
thomi*support22:36
balloonsthomi, yes I think most everything can be introspected, but options are good22:36
thomiI'm happy to add things, as long as we don't end up re-implementing xpresser - I'd rather integrate that than write it ourselves22:36
thomi(less code to maintain etc)22:37
balloonsyes yes.. ok, so while I have you, let's talk introspection a bit.. I'm interested in gtk apps to start22:38
balloonsdo you have any examples of this -- what's the best place to start digging?22:38
thomihmmmm, some documentation would be good huh :)22:41
thomialesage: do you have your gedit test code somewhere pubvlic?22:42
thomiballoons: alesage is the Gtk support person - if there's Gtk-specific issues, he's probably the best person to talk to.22:42
alesagethomi hi :) let me see . . . hi balloons22:42
balloonshi alesage :-)22:42
alesageballoons let me test this before sending it on to you22:45
balloonsalesage, sure thing.. I appreciate it22:45
balloonsthomi, duh! found the base class finally (class AutopilotTestCase(VideoCapturedTestCase, KeybindingsHelper):)22:46
balloonsnot sure how I missed that . . .22:46
thomiballoons: probably because the documentation needs to be much better :)22:46
balloonsit is very unity centric.. confusing to an outsider.. but we'll change that22:48
balloonsohh.. are there windowing functions.. do things like assert the focused window, move or resize a window, activate a window22:53
balloonsnvm nvm.. found it found it22:53
alesageballoons, here's a branch with a pretty minimal test lp:~allanlesage/+junk/UDS_AP_session/22:55
thomiballoons: yes, if you see anything that's unity-specific (even documentation), please file a bug against autopilot22:56
cariveribschaefer: hey there :) I dont how to draw the sub_LauncherIcons! please help.23:19
bschaefercariveri, what are you stuck on?23:19
cariveriwell, did new SimpleLauncherIcon() (it will be a list of those later) and setEmblem(iconpath). But I dont know how to draw it.23:20
cariveriId like to draw it just besides the this LauncherIcon.23:22
bschaeferhmm well im not 100% how they are being drawn atm...23:23
bschaeferdig through LauncherController, Launcher, AbstractLauncherIcon, and LauncherIcon23:24
bschaeferit has to get the icon texture somehow23:24
bschaeferit looks like they get drawn to the Launcher in Launcher.cpp , DrawContent23:25
bschaeferalso look under unity-shared/IconRenderer.cpp23:28
bschaeferas you'll use you window you make to draw the icons onto (I would think...)23:28
cariveriI did not find yet the LauncherIcon that gets drawed in the Launcher::DrawContent .23:36
bschaefercariveri, im pretty sure its1928...23:38
bschaeferbut that function is large and I haven't ever really looked at it23:38
bschaeferstart commenting things out and playing around with it :)23:38
cariveriOh my god. why could things be as easy as Launcher->Draw(){ ... Rander(_icon); +Render(_icon->subIcons);   };23:43
bschaeferwell...23:44
bschaeferthat im  not sure about :(...23:44
thumpercariveri: because the launcher is "special" :P23:55
cariverilol. trashed unity. will need to reboot it.23:55
bschaefercariveri, that is the fun part of messing with unity :)23:56

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!