/srv/irclogs.ubuntu.com/2014/06/04/#ubuntu-desktop.txt

=== broder_ is now known as broder
pittiGood morning04:56
Mirvpitti: morning! could you add ps-jenkins to https://launchpad.net/~canonical-platform-qa/+members ? cmake-extras would like to merge back from CI Train.05:11
pittiMirv: err, that sounds weird; that team isn't supposed to hand out development rights for projects, it's a "people" team05:14
pittiMirv: but I believe some "cmake"ish team was created recently, and ~canonical-platform-qa was put into that, sec05:15
pittiMirv: so ps-jenkins should perhaps be added to https://launchpad.net/~cmake-extras/+members ?05:16
pitti(I can't do that, though)05:16
Mirvpitti: right, maybe the code branch is then wrong where it's trying to push (https://code.launchpad.net/~canonical-platform-qa/cmake-extras/trunk)05:20
pittiMirv: oh yes, this should almost certainly be owned by ~cmake-extras05:20
pittiMirv: I think https://code.launchpad.net/~cmake-extras/cmake-extras/trunk is the real trunk, and Allan just forgot to delete the ~c-p-qa one05:21
Mirvbut there's also https://code.launchpad.net/~cmake-extras/cmake-extras/trunk which looks identical, so I can just push manually to there05:21
Mirvindeed :)05:21
pittiI can delete the c-p-qa one if you want05:21
pittiMirv: I think what happened is that he started with that one first, people complained about spam (and such a team is really a bad owner for projects anyway), and then created the ~cmak-eextras one05:22
Mirvpitti: sure, go ahead. I'll ask eg. satoris to pull the temporary branch pushed by ps-jenkins to the trunk.05:23
pittiMirv: ah, there's a pending MP on the c-p-qa one; I think we should better ask allan to deal with that, he'll know for sure05:23
Mirvok, makes sense05:24
=== tsimpson_ is now known as tsimpson
* didrocks implements mock ssl server now that the http is done and found http://www.piware.de/2011/01/creating-an-https-server-in-python/ which seems to be perfect for my needs!06:28
* didrocks hugs pitti06:28
larsugood morning!06:31
* pitti hugs didrocks back, bonjour!06:31
pittihey larsu, guten Morgen06:31
didrockshey larsu, pitti! :)06:31
Laneymorning!08:04
larsuhi Laney08:04
seb128good morning desktopers!08:04
seb128hey Laney, wb! had good days off work?08:05
seb128hey larsu ;-)08:05
didrocksgood morning Laney!08:05
Laneyseb128: yep, very nice thank you08:06
larsuhi seb128 :)08:06
Laneyclimbing & bought a new suit ;-)08:06
seb128what's the occasion? ;-)08:06
Laneygoing to a wedding in a few weeks08:07
LaneyI still had the same one that I bought 10 years ago in school ...08:07
seb128cool08:07
seb128doesn't fit you anymore? ;-)08:07
pittihey Laney, bonjour seb12808:07
Laneynot sure it ever did :P08:08
seb128pitti, lut, wie gehts?08:08
Laneyhey pitti larsu et didrocks!08:08
pittiseb128: très bien, danke !08:08
pittior perhaps..08:08
pittije vais very gut, dankon !08:08
seb128lol08:09
Laneystgraber: We don't have tracker in desktop, I'm happy for it to be processed as a normal gnome bugfix SRU personally08:10
seb128Laney, what was the question?08:12
Laneytracker sru08:12
seb128I'm unsure tracker follows the GNOME releases rules08:12
LaneyI thought it was by what's in the core (or whatever the name is) moduleset for jhbuild08:14
Laneywhich tracker is afaik08:14
seb128k, I'm unsure where tracker is nowadays08:15
seb128the spirit of the standing exception was to trust things that follow the GNOME feature/UI/... freeze08:16
seb128since that defines "stable" updates08:16
seb128tracker might be part of that08:16
Laneyanyways even without exceptions I tend to like upstream bugfix releases going in08:25
Laneyoh hey, the desktop-session mp was approved08:25
seb128Laney, yeah, speaking of which we should SRU stuff to trusty when we can ... like there is a new webkitgtk point release, and a new rhythmbox one08:26
Laneyyep08:27
Laneythe last webkit only got released yesterday08:27
seb128do you want to do the webkit one?08:27
Laneyok08:27
seb128thanks08:27
seb128oh ok, I just noticed because it's red on versions08:28
Laneyinto trusty-updates i mean08:28
seb128oh, ok08:28
seb128yeah, I know, I marked it verification-done on monday08:28
Laneycool, thanks for that08:28
seb128yw!08:29
didrockspitti: would you know by hard how with nosetests to be able to print debug logging statement? I'm trying to use --nologcapture --logging-level=debug, but only logger.warn() and logger.info() are printed, not the debug one (like when a test fails)08:42
pittididrocks: sorry, I don't know; I hardly ever use nose08:44
didrocksno worry, thanks! :)08:44
pittididrocks: but I doubt that you can control logger with environment variables; usualy you have to call logging.basicConfig() yourself and set the log level08:44
didrockspitti: seems that --logging-level=INFO from nose resets the logging level to what you want (even for dynamically created logger). The only "issue" is that it doesn't display them until a test fails. For debugging, I would like them to be displayed anyway08:46
didrocks--nologcapture prevents nose to capture them, but don't display them either08:46
=== Tm_Tr is now known as Guest17709
didrocksok, seems to work well now :)09:07
* didrocks drops a nose configuration files for easier loading later as well09:07
didrockswaow, awesome to see that python-mock documentation also covers my cornercase and have an example for it: https://docs.python.org/3.5/library/unittest.mock-examples.html#coping-with-mutable-arguments :)09:36
mvodidrocks: mock is the bomb09:40
larsumvo: nice picture, thanks ;)09:40
mvolarsu: your very welcome, I like it too!09:41
didrocksmvo: I'm using it for a while already, but didn't get to that case yet. Nice to see it being covered by the documention :)09:41
didrocksdocumentation*09:41
mvodidrocks: yeah, my experience exactly, its not only great, its also very well documented09:41
didrocksindeed, enjoyed a lot. Unfortunatly, most of the time, I end up in readthedoc which is slightly outdated since mock entered the python library09:43
=== Guest17709 is now known as Tm_Tr
=== alan_g is now known as alan_g|afk
=== alan_g|afk is now known as alan_g
pittiseb128: mind having a quick look at https://code.launchpad.net/~3v1n0/gnome-session/remove-compiz-requiriment/+merge/216236 ? does that make sense?11:10
pittiseb128: (I'll merge it to the correct branch and upload, just wanted to get a second opinion)11:10
=== gatox is now known as gatox_at_dr
seb128pitti, looks good to me in principle, I didn't look at the login sequence to see if that creates flickering issues or such though ... but I guess we can fix those later if needed11:13
=== MacSlow is now known as MacSlow|lunch
seb128well, Trevinho claims compiz is already started by upstart in most cases atm, so should not have that of an impact11:13
pittithanks11:13
seb128pitti, +1 from me, thanks for the sponsoring ;-)11:13
=== gatox_at_dr is now known as gatox
=== alan_g is now known as alan_g|lunch
=== MacSlow|lunch is now known as MacSlow
om26erif system-settings is started by autopilot on the phone it seems the settings are not saved in accountservice for ringtone etc, does anyone have clues on what could be wrong ?12:26
om26erthe same thing if done on the desktop works fine12:27
popeyCould someone point me in the direction of the thing that generates the xdg folders in $HOME? The ~/Music, ~/Downloads etc folders12:36
seb128popey, xdg-user-dirs-gtk12:37
popeythank you seb128!12:37
seb128popey, on desktop at least (/etc/xdg/autostart/xdg-user-dirs-gtk-update.desktop is the thing run on login)12:37
popeyperfect! thanks.12:38
seb128yw12:39
seb128om26er, could be that the dbus env is wrong? (similar to recent issues with the split greeter)12:39
daveyesdaveI'm looking for a way to colour code nautilus when connected to a remote server, for example red background when connected to live server etc...12:41
pp20daveyesdave: funny that! I asked the same yesterday... if you find out, let me know :)12:41
om26erseb128, that could be the case do you have any workarounds/suggestions ?12:44
om26erbecause when autopilot changes the tone and get backs to the sounds panel the ringtone remains the same12:45
seb128om26er, just make sure the dbus environment for those jobs is the same as the one from the user session12:46
=== alan_g|lunch is now known as alan_g
daveyesdavepp20, great minds and all that...13:11
seb128daveyesdave, seems like a wishlist to report on https://bugzilla.gnome.org/enter_bug.cgi?product=nautilus13:13
daveyesdaveseb128, do you know an alternative file manager that might already offer this function? commander etc?13:18
seb128daveyesdave, no idea, try #ubuntu for user questions13:18
daveyesdaveseb128, I'll give them a go, thanks13:19
seb128yw!13:19
=== tedg is now known as ted
=== ted is now known as tedg
=== vrruiz_ is now known as rvr
Laneynew webkit switches to geoclue 2 in Debian13:53
Laneyshould we do this too?13:53
seb128tedg, ^ do you know?13:53
seb128Laney, do you know what changed in the new one? are the providers the same?13:54
Laneyit looks like the API has been trimmed14:01
Laneybut the two versions are parallel installable14:01
seb128+1 for following Debian on webkit14:01
seb128we can still change back later if it turns out to be problematic14:01
Laneyso we'd have to promote the new one14:01
Laneyshould be ok I guess14:01
seb128that's a new codebase I think14:02
seb128so maybe stick to the old one until we SRU the new webkit/get the MIR reviewed14:02
Laneyhttp://www.hadess.net/2013/04/geocluing-desktop-slowly.html14:02
Laneydon't know how the ubuntu geoip stuff will fit in there14:03
seb128yeah, me neither, why is why I pinged tedg ;-)14:03
tedgSorry, OTP.14:04
* tedg reads backlog14:04
tedgseb128, Laney, well my understanding is that there are no plugable providers in GeoClue 2.14:05
tedgBut that's kinda a moot point in that we'll probably move things to location service.14:05
seb128tedg, do we need one?14:05
tedgSo I don't think that it matters really.14:05
tedgIf we wanted to use our GeoIP service, we'd need something plugable. But, not sure that we need one for generic webkit.14:06
tedgWe'll probably just end up working with Oxide there.14:06
seb128tedg, what is our provider doing that the upstream one isn't?14:07
seb128tedg, we don't want to regress unity7, not sure where we use geoclue ... we probably don't care much about webkitgtk, but we do care about ubiquity and the indicator I guess?14:07
tedgseb128, I think that we can have both, no? We can stick with GeoClue 1 for those while having GeoClue 2 for webkit GTK.14:08
tedgI think the issue is the service being free to use or needing API keys and the such.14:09
seb128tedg, sure, but it means we have 2 versions supported/installed14:09
tedgFor the interim releases, but I'm sure we'll be off GeoClue by the next LTS for those.14:09
seb128k14:10
seb128tedg, thanks14:10
seb128Laney, ^ makes sense to you?14:10
Laneydo you have to explicitly ask for ubuntu geoip with geoclue 1?14:11
* xnox is pretty sure ubiquity uses geoclue 1 against our canonical hosted server14:13
* xnox goes to check14:13
xnoxactually, no we don't. We just do a GET request to http://geoname-lookup.ubuntu.com/?query=%s&release=%s14:14
LaneyI think it's alright to use both for now I guess14:16
Laneyif I was curious I'd ask about the origins of location-service14:16
Laneybut I'm alright not knowing for now actually ;-)14:16
=== alan_g is now known as alan_g|tea
xnoxLaney: when location-service first came up, i thought it was related to cellphone-tower possition, which geoclue1 did not do....14:24
seb128Laney, you don't want to ask, that's another "cpp is better than g"14:24
seb128like we could have contributed to the geoclue rewrite with the GNOME guys14:24
Laneyxnox: I think there was an opportunity to influence geoclue :)14:25
Laneyseb128: yeah let's move on :P14:25
xnoxseb128: qt5 has a hard-dep on gtk2 still, i think we know what's better =)14:25
* xnox all Qt5 on Linux is delivered to you by Gtk+2 =)14:25
xnoxsame like it used to be the case with all Airbus planes to be delivered by a Boeing cargo transporter.14:26
Laneysymbiosis14:31
=== alan_g|tea is now known as alan_g
Trevinhopitti: on that patch it's probably the case also to remove unity-settings-daemon as well14:35
Trevinhopitti: as it's using upstart14:35
=== MacSlow_ is now known as MacSlow
didrocksok, seems the rain ended outside, time for a run!14:39
seb128didrocks, enjoy!14:40
didrocksthanks, wish me luck to avoid any rain shower :)14:41
didrocks(at least, temperature is nice)14:41
sil2100kgunn: hey! Did you guys propose your blueprints for the UOS already?15:57
=== gatox is now known as gatox_lunch
=== gatox_lunch is now known as gatox
=== seb128_ is now known as seb128
=== alan_g is now known as alan_g|EOD
rickspencer3hey all, mediascanner-service is using a lot of resources on my desktop :(18:08
rickspencer3and it keeps respawning if I kill it18:09
rickspencer3seb128, still around at all?18:11
Laneyrickspencer3: Looks like it has an upstart job which will restart it18:21
Laneystop mediascanner-2.0 should do it18:21
rickspencer3Laney, yeah, I finally figured that out18:21
Laneyah18:21
rickspencer3it was not in the list of upstart jobs18:21
rickspencer3but, I am probably just super confused18:21
rickspencer3I thought service --status-all would show me what it is called18:22
rickspencer3Laney, popey hooked me up (along with tab completion) :)18:22
LaneyI don't know what that does18:22
rickspencer3Laney, I assumed it would show all the services and their current status18:22
Laneyinitctl list is that one that I know18:22
Laneyanyway I would hope that it's running at a low enough priority to not impact the usage of your system18:23
Laneydid you notice it being slow?18:23
xnoxLaney: service command only ever operates on system-init (if it doesn't / leaks user-session info, it's a bug)23:02
xnoxLaney: indeed initctl commands should have been used instead...23:02
thumperTrevinho: around?23:34
thumperTrevinho: because you work crazy hours...23:34
Trevinhothumper: hi :)23:35
thumperTrevinho: hey there23:35
thumperTrevinho: I have some unity weirdness...23:35
Trevinhothumper: tell me...23:35
thumperTrevinho: I have five emails open, so lots of pips next to thunderbird23:35
thumperif I alt-` I get the tab spread, right?23:36
Trevinhothumper: you should yes23:36
thumperhowever, the images in the spread flicker23:36
thumpergoing from shaded to not, to shaded23:36
Trevinhothumper: yes, that's a known issue...23:36
thumperok23:36
thumperTrevinho: is it an issue that is likely to be fixed soon?23:36
Trevinhothumper: I also know how to prevent it, but it's not an optimal fix, so I avoided it for now... but it's something we should avoid asap23:37
* thumper nods23:37
Trevinhothumper: basically there's one-line change that fixes it (damaging the switcher area when painting these windows), but it would cause too many redraws also when not needed... and although I didn't notice any slow-down I didn't like the idea :)23:38
thumper:)23:38
Trevinhoanyway... yeah a better solution is in my list for some time23:38
robert_ancellRAOF, is there any logical reason why in X you can have a depth=32 window with bpp=24?23:52
RAOFrobert_ancell: Hm. I thought it would be the other way around.23:53
RAOFrobert_ancell: But, sure! Drivers want 32 bits per pixel, because 4 bytes is a nice round number.23:53
RAOFrobert_ancell: So basically it's a “pixels are 32bits, of which 24 are actually significant”23:54
robert_ancellRAOF, so it should be allowed?23:54
RAOFYeah. But from memory depth is the number of significant bits and bpp is the total number of bits; are you sure you're not getting those reversed?23:54
robert_ancellRAOF, I'm chasing down an assertion that's hit in X when running under qemu - the root window is 24bpp and if you use  the depth=32 visual on the window and render to it using RENDER it doesn't render/renders wrong23:55
robert_ancellSo I'm not 100% sure where the bug is, but the fact the drawable is set that way suggests something went wrong at window creation23:56
RAOFAh.23:56
robert_ancellwhich might be the fb backend, because I think that's what qemu is using; at least on my system23:56
RAOFI'm not entirely sure if RENDER handles that case correctly. It's possible to do correctly.23:57
robert_ancellRAOF, yeah, so RENDER should be dropping some bits somewhere to make it fit?23:57
RAOFYeah. The 8 bits of alpha would be a prime candidate :)23:57
robert_ancellyep23:57
robert_ancellRAOF, and a follow up there. Is there anything that actually specifies in a depth=32 visual that the remaining 8 bits is actually alpha? Is that just a convention?23:58
RAOFAs far as I can tell, convention. :(23:58
robert_ancellyeah, the spec mentions nothing. I was wondering if it got specified in a an extension somewhere23:59
robert_ancellThere really should be a XALPHA extension that turns this on23:59

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