/srv/irclogs.ubuntu.com/2013/01/31/#ubuntu-desktop.txt

=== m_conley_away is now known as m_conley
smspillazdoes anyone understand the fine details of nux's implementation of thread local storage ?04:05
smspillazparticularly in the context of the tests?04:06
=== m_conley is now known as m_conley_away
smspillazargh04:55
smspillazhmm04:55
smspillazTrevinho: you around ?04:55
smspillazTrevinho: what's the view on GLib usage in nux? Am I allowed to use it or should I avoid it if possible?04:56
RAOFsmspillaz: Wait, nux has an implementation of thread local storage?05:54
pittiGood morning05:55
gnomefreakhow one can help me but i would lik etofind out how to use the cube and wobbly windows befor i file a bug on it, please point me in the right direction06:11
pittignomefreak: these are not officially supported, but if it's anywhere then in ccsm06:11
gnomefreakpitti: i enabled them in ccsm yesterday06:12
gnomefreakwhen i booted just now i got the compiz splash screen06:13
gnomefreakthanks for the update06:13
smspillazRAOF: it has an abstraction, but I'm not worried about that right now06:14
RAOFsmspillaz: Oh, you're still in Perth, right? Want to catch up sometime?06:15
smspillazRAOF: now I'm looking at how to (sanely) inject some kind of GSource to poll on06:15
smspillazRAOF: sure, are you here ?06:15
RAOFsmspillaz: Indeed I am.06:15
smspillazawesome06:15
smspillazhow long until ?06:15
RAOFThe 10th06:15
smspillazhmm okay06:15
smspillazshould be around until then06:15
gnomefreak.win 2006:16
RAOFsmspillaz: Can't you get a GSource to give you an fd?06:16
gnomefreakdamn'/win 2006:16
gnomefreakhyhn06:16
RAOFHEh.06:16
smspillazRAOF: well, I'm just thinking about the best way to poll on some fd, inside of a nux thread06:16
smspillazbut I think I have an idea06:16
smspillazI love it how nux make check segfaults06:18
smspillazI'm fixing that before someone screams at me for trying to merge anything else06:18
RAOFnux has an event loop, doesn't it? Can't you inject into that?06:19
smspillazI can06:21
smspillazRAOF: I'm just wondering how appropriate it would be to use GIOChannel with g_io_add_watch in something that's in Nux/Nux06:21
smspillazconsidering portability issues (eg, compiling without GLib main loop support on Fenetres)06:21
RAOFObviously you can't do that.06:21
smspillazRAOF: my sarcasm detector is broken06:22
RAOFThen again, can you poll on an fd on Windows? I'd have thought you'd be using various Handles, and win32 isn't particularly posixy.06:22
smspillazRAOF: you can, but I can't block the event thread06:23
RAOFYou can't do a non-blocking select-y thing? Work out whether there's anything to read first, then read if so?06:23
smspillazRAOF: the point is to integrate it into the event loop, don't really want to have to rely on checking it every few ms06:24
smspillaz(I wouldn't have to do this if the nux test framework used lots of threads for no apparant reason)06:25
RAOFI'm not familiar enough with the nux event loop to know how it handles waking up, but it'd be useful if it could watch fds, wouldn't it.06:27
RAOF:)06:27
smspillazit kinda does already06:27
smspillaz(geis events and x events)06:27
smspillazbut theres no way to say "watch this fd plz"06:27
smspillazcan't be too hard to implement I guess06:27
RAOFYeah, it should have one of those.06:27
RAOFAll real event loops do :)06:27
smspillazwell, it just delegates to glib06:28
smspillazwhich is making me wonder if the API even needs it, since I only need this for the tests06:28
smspillaz...06:28
smspillazI don't really think it needs it06:28
smspillazwhy is it that every time I try to fix a bug in Nux that's annoying me06:30
smspillazits always a gazillion times more complicated than I thought06:30
RAOFHeh.06:33
RAOFThat's part of what you get with a cross-platform toolkity thing.06:34
smspillazNo I think its just ....06:34
smspillaznever mind I won't go there06:34
=== Quintasan_ is now known as Quintasan
smspillazRAOF: the number of ifdefs in nux is too damn high07:34
desrtseb128: good morning08:43
seb128hey desktoper08:43
seb128desrt, hey08:43
desrtany word on if the fix took?08:43
seb128desrt, no, your fpaste link had expired by the time I wanted to try it...08:45
seb128desrt, if you can share again it would be good08:45
seb128desrt, default expiration is one hour08:45
desrtseb128: you can use a captcha...08:46
seb128doh08:46
seb128desrt, thanks ;-)08:46
seb128trying the patch...08:46
seb128(what a stupid ui)08:46
desrtit's to prevent people using it as a platform for distributing malware08:47
desrtmuch better than our system of forcing login, imho08:47
seb128yeah, it's just that like most people I stopped at the <b>expired:</b> the page you are trying to view has expired08:48
seb128users don't read :p08:48
seb128I though the captcha/entry was somewhat part of the fpaste site ui08:48
desrtpretty weak :)08:48
desrti'm just glad you didn't revert dconf meanwhile :)08:49
seb128desrt, that fixes it \o/08:58
desrtnice08:58
desrtcan you get a blame for me on who is responsible for the write that causes the crash?08:58
desrts/causes/caused/08:58
seb128desrt, trying08:58
desrtbecause this person is the cause of that other crasher too that was recently fixed08:59
desrtturns out dconf-service is not very well-protected against receiving non-sense data from clients...08:59
seb128desrt, if DCONF_BLAME is set, where is the debug output going? stdout?09:00
desrtno.  you run 'dconf blame' commandline tool09:00
seb128what am I looking for?09:01
desrtwell09:02
desrtyou said it fixes the problem09:02
desrtwhat did you do?  test logging in?09:02
seb128I've the blame log, just looking for the info09:02
desrtpaste it09:02
seb128starting a guest session09:02
desrtperfect09:02
desrtjust curious to know who is sending the change request with zero items09:02
desrtcould be a bug somewhere else still...09:03
seb128Parameters: ()09:03
seb128PID: 1702409:03
seb128ah09:03
* desrt would upgrade to R but larsu says it's a bad time :)09:03
seb12817024 pts/3    Sl+    0:00  |               \_ dconf blame09:03
seb128hum09:03
larsudesrt, it's not that bad a time if you don't need wifi09:03
seb128desrt, http://people.canonical.com/~seb128/dconf09:03
desrt16671 ?        Sl     0:00          \_ dconf load /org/compiz/profiles/Default/plugins/core/09:04
desrthuh.09:04
desrtohhh!!!09:04
desrti bet that's empty09:04
seb128iz compiz?09:04
desrtsince it's probably trying to migrate zero settings09:04
desrtso it ends up trying to commit an empty set of changes09:04
desrti should make 'dconf load' check for empty keyfiles and do nothing09:04
larsuseb128, morning!09:05
seb128larsu, hey, how are you?09:05
desrtsince i go around advising people to use it in this way :)09:05
seb128desrt, ;-)09:05
larsuseb128, good, thanks. Everybody speaks a weird language here09:05
seb128larsu, esperanto?09:05
desrtseb128: it sounds like french, but they don't know how to count properly09:05
seb128larsu, or you mean the people in the street?09:05
seb128lol09:05
seb128yeah, that's faked french, don't trust them :p09:06
desrtlike, they don't know that you talk about 99 like "4 20 10 9"09:06
desrtidiots09:06
seb128;-)09:06
larsuseb128, ha, no. Only one person speaks that here09:06
seb128desrt, how did you figure it was 16671 from this log?09:07
seb128desrt, the parameters from that call are non empty09:07
seb128desrt, btw are you happy with the patch? eg should I upload it?09:08
desrtyes. please do09:09
desrtalso... compiz... wow09:09
desrtthat's a lot of writes on login09:09
seb128desrt, thanks again for taking time to work on that one09:09
seb128talk to Didier when he's back next week :p09:09
popeyhmm.. updated my machine this morning... got a new xserver-xorg-core and now x won't start with nouveau or nvidia drivers09:12
seb128popey, raring?09:12
popeyyes09:13
seb128mlankhorst, ^09:13
seb128tjaalton, ^09:13
popeyhttp://paste.ubuntu.com/1592607/  <- xorg.0.log09:13
desrtseb128: larsu just confirmed on his R box that those writes are not on every login09:13
desrtbut 70 writes during first login is still very bad...09:13
seb128desrt, right, I was going to say that the segfault is happening only with guest sessions for me, e.g new users09:13
seb128desrt, I think the compiz profile stuff is hackish and it does write all the profiles keys on first start or something09:14
desrtya... i think so too :(09:14
mlankhorstpopey: looks fine? :S09:17
popeyhttp://paste.ubuntu.com/1592628/09:21
popey[+0.33s] DEBUG: Failed to start greeter09:21
seb128popey, where did you get the Xorg log? you probably want the one in /var/log/lightdm if that's lightdm which fails to start09:24
popeyah09:24
seb128popey, does it work if you sudo stop lightdm and startx?09:24
popeydidn't realise there was a separate xorg.log, sorry09:24
popeyalan@wopr:/var/log/lightdm$ sudo cat x-0-greeter.log09:25
popey/usr/sbin/unity-greeter: symbol lookup error: /usr/lib/x86_64-linux-gnu/libgdk-3.so.0: undefined symbol: wl_registry_interface09:25
popeythat more likely it?09:26
popeyhttp://paste.ubuntu.com/1592635/ is /var/log/x-0-greeter.log09:26
popeyhttp://paste.ubuntu.com/1592637/ is /var/log/x-0.log09:27
seb128popey, using a crack ppa?09:27
popeyheh, didn't think so..09:27
seb128popey, what version of libwayland0 do you have?09:27
seb128I saw a bug report yesterday from somebody who had a 1.0.5git09:28
seb128dunno where it's coming from, but it seems to break abi and break gtk (since it's built with wayland)09:28
popeybah!09:28
popeyhttp://ppa.launchpad.net/canonical-qt5-edgers/qt5-beta1/ubuntu/09:28
* popey pokes Mirv 09:28
popeygood spot, thanks seb12809:28
seb128popey, yw09:28
seb128I don't like much that gtk on X breaks when libwayland abi changes though09:29
seb128Laney, ^09:29
Laneyhmm?09:29
Mirvpopey: bzoltan still handles beta1, although I can help if I know what should be fixed. I know he was trying wayland related builds there yesterday09:29
Laneywhat do you want? the SONAME should be bumped and gtk rebuilt if the ABI breaks09:29
seb128Laney, just as fyi, that might make me revert the wayland backend09:29
Laneydoesn't seem like gtk's fault09:29
seb128Laney, well, I just don't like the idea that gtk-X breaks when there is a problem in the wayland backend09:30
seb128I would assume that wayland issues only impact people running that backend09:30
seb128Laney, bug yeah, you have a point, the bottom of the issue is wayland breaking abi without soname change09:32
RAOFHas anyone hooked up sbuild to autopkgtest so that builds automatically do a test run?09:32
xnoxRAOF: no but there is lp:auto-package-testing which has scripts to setup testbed (cloud image) and run adt tests against package/branch/ppa in kvm using that cloud image.09:35
RAOFxnox: Hrm. Not quite what I was after, but I guess it'll have to do :)09:37
TuxMuxhello everyone09:45
xnoxRAOF: that's what jenkins uses itself. So it's a more matching environment.09:48
TuxMuxlol09:49
=== vrruiz_ is now known as rvr_
pittiseb128: bug 160311 -- didn't we use to have a GTK3 change/patch/upstream feature to add a resize handle on the bottom right corner? that seems gone now11:20
ubot2Launchpad bug 160311 in metacity "Resizing windows by grabbing window borders is difficult" [Low,In progress] https://launchpad.net/bugs/16031111:20
seb128pitti, that's a GTK3 upstream feature11:21
seb128we used to distro patch gtk2 to provide it as well11:21
pittibut it seems gone now?11:21
seb128I think it's activable from the theme11:21
seb128it's a theme option11:21
seb128it's off by default11:21
pittiah, that's why light-themes has a task there11:21
seb128since gnome-shell/compiz both have invisible borders11:21
seb128invisible borders = an extra area you can grab11:22
pittithere was a recent flurry of comments on that11:22
seb128that fixes the bug without needing an extra ui element11:22
pittiack11:22
seb128well, not a lot we can do, users can tweak their theme...11:23
seb128pitti, http://git.gnome.org/browse/gnome-themes-standard/commit/?id=23c3146ddafe731366212b39d051fef121b29c8a11:26
seb128pitti, I guess it's a matter of changing those 2 parameters to !011:26
pittiah, that's the visible corner?11:26
pittithanks11:26
seb128yw11:26
=== jhernand1z is now known as jhernandez
=== amithkk_ is now known as amithkk
=== amithkk is now known as Guest8601
ricotzcyphermox, hello :)12:44
ricotzcyphermox, could you update network-manager and network-manager-applet to 0.9.7.995? and a look at them and nm is pretty straight forward, but nm-applet needs quite some patches refreshed12:46
=== MacSlow is now known as MacSlow|lunch
GunnarHjseb128: Hi Sebastien, do you have a minute?13:55
=== bdrung_ is now known as bdrung
ritzseb128 , hi https://bugs.launchpad.net/ubuntu/+source/unity/+bug/80624814:01
ubot2Ubuntu bug 806248 in unity (Ubuntu Precise) "unity::TimeUtil::TimeDelta returns an int value which overflows after 24 days of uptime" [High,In progress]14:01
ritzwhom do I talk to about this14:01
seb128ritz, #ubuntu-unity, sil2100  Mirv popey14:02
ritzseb128++ thank you :)14:02
seb128yw!14:02
ritzseb128 busy ?14:05
seb128ritz, yeah, there is always something to do here :p but I can reply to questions14:06
ritzseb128 how long before you log-off ?14:06
seb1284.5 hours at least14:06
ritzseb128 wrt https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/106597914:06
ubot2Ubuntu bug 1065979 in gnome-settings-daemon (Ubuntu) "external/internal monitors mirrored on boot when laptop lid is closed" [Wishlist,Confirmed]14:06
chrisccoulsonhi jibel14:19
chrisccoulsonwill jenkins be happy with http://people.canonical.com/~chrisccoulson/reftest.xml?14:20
chrisccoulsonit does validate against https://svn.jenkins-ci.org/trunk/hudson/dtkit/dtkit-format/dtkit-junit-model/src/main/resources/com/thalesgroup/dtkit/junit/model/xsd/junit-4.xsd ok14:20
jibelchrisccoulson, excellent. I'll update the jobs to collect and publish these result files. You can generate more than one, e.g one per testsuite, if you wish.14:23
chrisccoulsonjibel, cool, thanks. i've still got to implement this for one other test-suite, but it shouldn't be too difficult now14:24
cyphermoxricotz: yeah, I'll try to get to it today, just need to finish some work on bluetooth first14:41
ricotzcyphermox, thanks14:43
=== m_conley_away is now known as m_conley
ritzseb128 hi, pm. thanks14:48
=== MacSlow|lunch is now known as MacSlow
xnoxstrange but system-service-d is not running on my nexus right now.15:41
xnoxwhat do i need to use to dbus activate it?15:41
Laneyxnox: one way is to open the details panel in g-c-c15:44
mterrydesrt, got a sec to talk about a crash in dconf-service that I'm seeing on our unity jenkins autolander service?15:45
Laneymterry: got https://launchpad.net/ubuntu/+source/d-conf/0.15.2-0ubuntu5 ?15:46
* mterry confirms15:47
mterryI mean, I'm checking15:47
mterryLaney, nope...  that may be the bug.  Thanks!15:48
xnoxLaney: thanks15:48
seb128xnox, that service is supposed to go away, don't spend too much time trying to optimize/fix it ;-)15:55
desrtmterry: sure15:56
xnoxseb128: i know =) i'm tinkering to see if python -O or pypy make any difference, my current lab rats are the currently python idle processes.15:56
desrtmterry: ah.  cool.15:56
xnoxseb128: i know they are going away, but some other pieces of python code will not.15:56
seb128xnox, right, python optimization is good in any case15:57
xnoxseb128: but idle is nice to measure as I don't have to launch anything explicit =)15:57
seb128right15:58
seb128well, u1 and oneconf at least are python as well15:58
GunnarHjseb128: ping16:37
seb128GunnarHj, hey, sorry I saw your ping earlier but I was a bit busy and forgot about it ... better to just ask your question usually ;-)16:38
GunnarHjseb128: Ok, here it is (not a question, really, but anyway...)16:39
GunnarHjseb128: Some locale settings migration code in accountsservice.postinst 'disappeared' in the beginning of the Quantal cycle, probably by mistake.16:39
GunnarHjhttp://people.ubuntu.com/~gunnarhj/accountsservice-discontinuity/accountsservice-discontinuity.diff16:39
GunnarHjNot good. :( I'm not sure if it should be reinserted - pitti may have an idea.16:39
seb128GunnarHj, I probably dropped it, why is it still needed?16:40
seb128we don't support upgrades from < precise16:40
seb128and users on precise should already have migrated16:40
seb128the logic was: anyone upgrading to precise will have the migration done16:40
seb128the only users upgrades from quantal (in a supported way) are precise user16:41
seb128so migration shouldn't be needed on quantal16:41
seb128it's usual practice to keep migration code until the next LTS and then drop it16:41
GunnarHjseb128: Well, I guess it would only be useful for upgrades from pre Precice.16:41
seb128that's not supported16:41
seb128the only upgrade path supported is to upgrade to precise16:41
seb128then to quantal16:41
GunnarHjseb128: Ok, I see. Then it's not a problem.16:41
seb128good ;)16:41
GunnarHjTnx.16:42
seb128GunnarHj, thank you for asking, better to check when unsure ;-)16:42
=== tkamppeter_ is now known as tkamppeter
=== ogra_ is now known as ogra
seb128mterry, do you know offhand if unity-greeter renders its background image/wallpaper or if it uses g-s-d for that?17:38
mterryseb128, it renders it17:38
mterryspeaking of, I should upload a patch for the g-s-d thing, now that it's upstream17:38
seb128ok, so we could probably desactivate the g-s-d background plugin by default17:38
seb128it does nothing but checking out if nautilus is running and bailing out on the desktop17:38
seb128mterry, g-s-d thing?17:39
seb128mterry, you mean the spawning?17:39
mterryseb128, the dbus-activation vs spawning it17:39
seb128mterry, yeah, that would be great, that should fix screen rotation on the greeter ;-)17:39
bcurtiswxis there still plans for the unity-webapps so that there does not need to be a browser icon AND webapp icons?17:43
sarnold(I thought the webapp icon was 1/3 of the selling points?)17:44
bcurtiswxif i only have gmail and twitter open, why would i need my firefox to say it's open ?17:44
bcurtiswxthats what the twitter and gmail webapp icons are for17:44
sarnoldah, I see :) my firefox never has fewer than 20-ish tabs..17:46
bcurtiswxsarnold, you part of that team ?17:46
sarnoldbcurtiswx: no17:47
bcurtiswxwho's the head of that team ?17:47
seb128bcurtiswx, yes, there are plans to make webapps proper apps17:57
seb128e.g be in chromeless ui rather and not match the browser17:57
seb128that was the plan last cycle already, just ETOOMUCHTODO17:57
bcurtiswxseb128, we got a break from gnome 3.8 so what's taking up all your time? :P17:58
seb128bcurtiswx, https://blueprints.launchpad.net/ubuntu/+spec/desktop-r-arm-reduce-footprint ;-)17:58
bcurtiswxthe phone push 'eh17:58
seb128bcurtiswx, and https://blueprints.launchpad.net/ubuntu/+spec/desktop-r-reduced-power-ram17:58
seb128bcurtiswx, well, making Ubuntu fit for mobile form factors yes, but improved performances benefit desktop and laptop users as well ;-)17:59
bcurtiswxwait, people use desktops still? haha.  I don't have the time to fix, only complain. so I'm still going to contribute any way i can :D18:00
seb128;-)18:00
bcurtiswxnow if anyone has a tool to write my dissertation for me, i'd totally have more time to spend fixing18:01
seb128bcurtiswx, google, copy, paste? :p18:02
bcurtiswxhaha, i WISH. oh well.. back to typing.18:02
* bcurtiswx waves bye18:02
=== gatox_ is now known as gatox
qengho(No editor installed? cp and sed will do.)18:31
sarnoldnot brave enough sed -i? :)18:32
qenghosarnold: not that brave.18:32
cyphermoxqengho: sed -i you can backup before changes -- sed -i.bak 's/whatever/whateverelse/'18:42
qenghoyep.18:47
chrisccoulsondoes the number of workspaces in compiz keep resetting back to the default for anybody else?18:51
chrisccoulsoni always set mine to 6, but every couple of days i notice it has reset back to 4 again18:51
qenghochrisccoulson: I see that a lot.19:00
chrisccoulsonqengho, ah, i'm glad it's not just me :)19:00
chrisccoulsontis a bit annoying though19:00
qenghochrisccoulson: it'd be better if it were just you.  :(  Let's talk to #ubuntu-unity19:01
qenghochrisccoulson: I'm asking.19:07
chrisccoulsonthanks19:09
=== wendar is now known as allisonrandal
chrisccoulsonok, i think i'm going to have to restart. my session seems really broken (windows not appearing, randomly not painting) :/19:14
=== AlanChicken is now known as alanbell
=== alanbell is now known as AlanBell
mterrympt, the automount behavior that Ubuntu has now; do you like it?  (in the context of upstream dropping the automount tool we use now in favor of embedding it in GNOME Shell; we need to figure our plan of action)20:31
mterrympt, specifically, it automounts usb sticks and acts on the preferences in System Settings -> Details -> Removable Media, as far as I can tell20:32
mterryseb128, ^ has this been discussed before?20:33
mterryseb128, I ask because the mount helper is taking 2.1M PSS space on the nexus620:33
mterry7 even20:33
seb128mterry, I don't think the UI/interaction got revisited/discussed20:34
seb128mterry, we have a workitem to merge that helper in unity though20:34
mterryseb128, ooh, we do.  OK20:35
seb128mterry, https://blueprints.launchpad.net/ubuntu/+spec/desktop-r-gnome-fallback20:37
seb128mterry, the WIs are on this spec20:37
=== m_conley is now known as m_conley_away
robert_ancell_mterry, ping23:50
mterryrobert_ancell, hi23:50
mterryrobert_ancell_, hi23:51
robert_ancell_mterry, do you know what commit to u-g made the message box expand when large messages come from PAM?23:51
robert_ancell_i.e. the change that makes the "Multi Info Prompt" user work in u-g --test-mode23:51
mterryrobert_ancell_, I believe that was during quantal, as part of some refactoring work from the remote server support23:51
robert_ancell_yeah I thought so23:52
mterryrobert_ancell_, don't know exact commit23:52
robert_ancell_any idea how hard it would be to backport to precise?23:52
mterryrobert_ancell_, but that made box handling much more dynamic23:52
mterryrobert_ancell_, difficult I'd guess23:52
robert_ancell_:(23:53

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