/srv/irclogs.ubuntu.com/2014/07/29/#ubuntu-desktop.txt

psusiso after upgrading to utopic, after a bit Xorg decides it is going to burn several seconds of cpu time locking up on various benign activities like closing a tab in firefix, sometimes changing tabs in gnome-terminal, opening email in thunderbird... this doesn't seem to be a problem booting the livecd... any ideas how to debug this?02:06
Sarvattpsusi: guessing nvidia or fglrx drivers on the actual install and not on the livecd? :)02:15
=== m_conley is now known as m_conley_away
pittiBonjour tout le monde04:01
seb128good morning desktopers07:55
seb128hey robert_ancell07:57
robert_ancellseb128, hello07:57
seb128robert_ancell, how are you?07:57
robert_ancellgood07:57
seb128robert_ancell, so the gnome-desktop new dbus service/api solution doesn't seem to be that easy, we are thinking we should land your work07:58
seb128robert_ancell, do you think you could continue/get it done?07:58
robert_ancellseb128, sounds good to me07:58
robert_ancellseb128, I'll sort it tomorrow07:58
seb128you were still missing some bits like idle monitor iirc?07:58
seb128thanks07:58
robert_ancellyeah, that's not too much more work07:59
robert_ancellI also have to fix unity-greeter07:59
Laneynice07:59
robert_ancellI have some branches for both of those somewhere07:59
seb128great07:59
robert_ancellseb128, so no more thoughts on my GTK+ issue? As I understand it, debuild does a clearenv(), then doesn't explicitly set XDG_RUNTIME_DIR. The tests require XDG_RUNTIME_DIR so I'm wondering how it could be set08:00
robert_ancellThe other option is the Wayland backend is disabled on your system for some reason, but again, no idea why that should occur. AFAICT it's enabled for everyone right?08:00
seb128robert_ancell, I need to try again, but it's possible that wayland is not built for me, I seem to remember .symbols diff in wayland symbols08:02
seb128but not sure why that would be08:02
LaneyXDG_RUNTIME_DIR> a few packages make this in debian/rules and then export it08:02
Laneysame for HOM08:02
LaneyE08:02
seb128but why would it be working for !robert_ancell?08:03
LaneyI don't know what problem you're talking about :P08:03
seb128Laney, https://bugs.launchpad.net/ubuntu/+source/gtk+3.0/+bug/134239808:04
ubot5Launchpad bug 1342398 in gtk+3.0 (Ubuntu) "display tests fail" [High,Incomplete]08:04
seb128Laney, read comment #308:04
robert_ancellLaney, ah, that might be worth looking at08:04
seb128robert_ancell, dpkg -l | grep libwayland ?08:04
Laneyare you using sbuild and nobody else is?08:05
robert_ancellLaney, no, just bzr-buildpackage from the directory08:05
Laneyhmm08:05
robert_ancellI know it's not a super-clean-build but it's always worked fine previously08:05
seb128robert_ancell, you have a local version of libwayland in /usr/local?08:05
robert_ancellI also haven't checked how the error callback is done in Wayland, I guess there's a possibility there's a race08:05
robert_ancellseb128, no08:05
seb128just trying to rule out difference08:05
Laneyeither way I think it's reasonable defensiveness to set this08:06
robert_ancellI'll copy the glib way it's set and try again08:06
seb128thanks08:06
Laneyhow do I bzr push a revision but not its tag?08:07
robert_ancellLaney, you want the tag locally?08:07
Laneyyeah08:08
LaneyI just want to keep it unpushed until the SRU is accepted08:08
robert_ancellI think you can only delete the tag, push, then create it again with the standard tools08:08
Laneyokay08:09
robert_ancellI'm always bitten by the way you can't delete a tag once it's in an upstream branch, only move it it seems08:09
LaneyI'm going to check in #bzr to be sure08:12
=== vrruiz_ is now known as rvr
robert_ancellLaney, did you get an answer to your bzr tag question?08:42
larsurobert_ancell: Laney says no08:42
larsurobert_ancell: in #bzr that told him the same thing08:43
robert_ancellah08:43
robert_ancellyou guys all in Starbourg?08:43
robert_ancellyou guys all in Strasbourg?08:43
larsuhaha starbourg :)08:44
larsuyep, we are08:44
robert_ancelljealous08:44
larsu:)08:44
larsuwe miss you08:44
robert_ancellhave a beer for me :)08:44
larsudefinitely!08:44
seb128robert_ancell, we are going to miss you in China as well, shame that was midway for you08:47
robert_ancellyeah, I'm really sad I'm going to miss that one08:47
robert_ancellHopefully there will be other China opportunities in the future08:48
seb128yeah08:48
seb128you are off for a week or longer?08:48
robert_ancelltwo weeks08:48
seb128k08:48
seb128enjoy!08:48
seb128starting friday?08:48
robert_ancellI will do :)08:48
robert_ancellMonday08:48
seb128well, friday after work I guess ;-)08:48
robert_ancellI have the weekend to relax before my holiday :)08:49
seb128I see ;-)08:49
happyaronseb128: I'll deal with that mp when merging ibus 1.5.7, which is still sitting in NEW08:51
seb128happyaron, ok, thanks08:51
happyaronwell, upstream just released 1.5.8...08:51
seb128happyaron, let's update then ;-)08:53
happyaronyup08:53
robert_ancellLaney, aw yeah. Your XDG_RUNTIME_DIR trick works perfectly08:58
seb128robert_ancell, nice, it still doesn't explain why it works for us and not you though08:58
seb128but I guess as long as it works..08:58
robert_ancellseb128, yeah, it's weird but the Wayland backend does require this and it could be invoked in tests so we should set XDG_RUNTIME_DIR08:59
seb128robert_ancell, so, since you are on holidays in some days, can you get gnome-desktop read08:59
seb128ready before that08:59
seb128I guess gtk/mir is going to be for after you are back?08:59
robert_ancellseb128, you mean the u-c-c changes?08:59
seb128ye08:59
seb128s08:59
robert_ancellI'll upload the gtk-mir changes tomorrow too now I can build it09:00
seb128great09:00
robert_ancellok, I'm off. Enjoy GUADEC!09:02
larsuthanks!09:03
seb128robert_ancell, thanks, have a nice evening09:03
* Sweetshark does a jump to the left.09:09
* Sweetshark does a step to the right.09:09
* Sweetshark puts the hands on his hips.09:09
Sweetsharkwell, you know the drill.09:10
seb128hey Sweetshark09:10
Sweetshark(FWIW, LibreOffice 4.3.0~rc4 successfully finished building on utopic-proposed/armhf)09:11
Sweetsharkseb128: heya!09:11
seb128nice09:11
seb128no willcooke today?09:20
* seb128 sends a no meeting email then09:20
mvo_Sweetshark: so you do a timewrap dance in order to make LO build?!?09:21
Sweetsharkmvo_: that is part of the ceremony along with sacrifcing a chicken and a sardine to a cargo plane under a palm tree and other things ...09:33
mvo_heh09:34
=== Estilanda_ is now known as Estilanda
* Sweetshark refactored some 20+ year old writer core code hard to half the LOC last week. twas fun, a bit like http://i.imgur.com/9lAkenv.gif09:39
SweetsharkIm looking forward to the regression reports.09:39
mvo_:)09:40
Sweetsharkseb128: since no meeting, here is a quick status update: We'll have LibreOffice FTBFS for a few days on utopic as there is a desparate need to transition all the stuff that has been stacking up on utopic-proposed.09:56
Sweetsharkseb128: As said: 4.3.0 builds in proposed, so fixing up 4.2.4 again just for a few days seems like a waste of time. (esp. since when armhf is finished building, it will be broken again </sarc>)09:58
seb128Sweetshark, k10:00
Sweetsharkseb128: and 4.2.6 is in the PPA and looking good so far, so could be SRUed in a week or so (well, when Im back from Uruguay) ...10:00
willcookehi folks10:08
seb128willcooke, hey10:15
seb128willcooke, not sure if you wanted to have an IRC meeting with people who are not at GUADEC?10:16
seb128I sent an no-meeting email since you were not online and a part of the team is not going to be available today10:17
ricotzSweetshark, hi :)10:53
Sweetsharkricotz: ho10:56
ricotzSweetshark, sounds about time for a libreoffice-4-3 ppa ;)10:57
Sweetsharkricotz: heh, yeah.10:57
ricotzi would prefer to keep 4.3 out of the main ppa until the obvious regressions are fixed10:58
ricotzi mean for the backports10:58
ricotzi seems those new "lgw01-*" builders are quite something11:01
Sweetsharkwell, I wouldnt bother too much with waiting. After all thats what upstream releases and if you have the main PPA you should know what you are in for. Otherwise you should use the libreoffice-4-x ppa.11:03
Sweetsharkand the best way to make regression go away fast, is to have some decent install base.11:03
Sweetsharkricotz: ^^11:03
ricotzSweetshark, of course the conflicts needs to be fixed as well first11:03
ricotzand yeah, but some don't actually realized what they are committing too11:04
Sweetsharkoh, yeah. Well, packaging conflicts are something else. It should install generally ;)11:04
ricotzmeaning the libreoffice-common <-> libreoffice-base problem11:04
ricotziirc11:04
Sweetsharkricotz: well, the PPA should do what it says it does and not what we think what people think it does. the latter is the road to madness. ;)11:05
ricotzSweetshark, i see not sure i mentioned it already, but would it be reasonable to have a libreoffice-fresh and libreoffice-stable ppa too?11:08
ricotz(Sweetshark, better just "fresh" and "stable" to avoid this verbose naming)11:11
Sweetsharkricotz: IMHO no.11:13
Sweetsharkricotz: uptsream is pointlessly discussing renaming libreoffice-stable to soemthing else (there is a new proposal each week)11:14
Sweetsharkricotz: and the main PPA is already what libreoffice-fresh is.11:14
ricotzheh, ok, having a release named stable when it reached EOL is not useful anyway11:15
Sweetsharkricotz: having two ppas doing the same thing would just create endless confusion. And "killing"/deleting the main PPA makes you just lose a lot of subscribers.11:15
Sweetsharkricotz: I think we had this discussion a few times already.11:16
ricotzcould be, but the organization is still confusing, people are just using libreoffice/ppa to avoid the need to look after things and blindly do updates imo11:18
ricotzso at least a "stable" ppa would make sense which will receive major-version updates if it is considered safe11:19
Sweetsharkricotz: actually a "stable" ppa will not work anyway for most people. example:11:20
=== MacSlow is now known as MacSlow|lunch
SweetsharkWhen trusty was released it came with 4.2.3 (fresh), so trying to install "stable" 4.1.x on it would be a pain anyway.11:21
ricotz(the 9-month support cycle makes pockets obsolete quite fast as well)11:22
ricotzhmm, ok, i guess i have more the pov of precise users11:22
ricotzthe numbers of trusty and precise ppa-users are already kind of the same11:24
ricotz(ca 11000 vs 13000)11:25
Sweetsharkricotz: well, for trusty we consider shipping a new series too (e.g. something like 4.3.6 when its there). So it would only be about precise ...11:25
ricotzSweetshark, oh, has this decided already?11:26
xnoxSweetshark: doom has migrated to release pocket. So you are good to go with 4.3.0 =)11:33
Sweetsharkxnox: yay, utopic doom follows trusty quake!11:41
ricotzwhat specs do the new ppa builders have?11:43
ricotzlooks like the libreoffice precise i386 build will finish under 3 hours11:44
ricotzSweetshark, of course dont upload it as long /usr/lib/libreoffice/presets/basic/script.xlc and /usr/lib/libreoffice/share/basic/dialog.xlc conflict ;)11:52
Sweetsharkricotz: is that on precise only? otherwise, do you have a bugref?11:54
ricotzSweetshark, this is your utopic build, there is no bug11:55
ricotzi mentioned it to you for rc1 already11:55
ricotzlibreoffice-common and libreoffice-base11:55
ricotzSweetshark, this is more something for a bug report http://people.ubuntu.com/~ricotz/libreoffice/11:59
=== alan_g is now known as alan_g|lunch
Sweetsharkricotz: well, what have those pdfs to do with those .xlc files and -common and -base?12:06
ricotzSweetshark, nothing, this is just a problem which would prevent me from using 4.3.0 on a wider scale12:08
Sweetsharkricotz: well please file an issue proper. there are 27000 others to keep track of.12:09
ricotzSweetshark, will try12:11
ricotzSweetshark, i am wondering, aren't you installing the complete libreoffice stack for testing?12:11
ricotzanother issue could be the "g++-" and "gcc-" references12:14
Sweetsharkricotz: I see no conflict on  /usr/lib/libreoffice/presets/basic/script.xlc12:14
ricotzseems the GCC_VERSION is not set on generation for you12:15
ricotzSweetshark, oh, both files are in /usr/lib/libreoffice/share/basic/12:15
=== MacSlow|lunch is now known as MacSlow
ricotzthis considers the build-deps and deps of libreoffice-subsequentcheckbase in debian/control12:16
ricotzi will be back later12:17
=== oCrazyLem is now known as CrazyLemon
=== alan_g|lunch is now known as alan_g
happyaronLaney: I'm looking for how to create a packageset for cinnamon, is a mail to devel-permission suffice?12:46
Laneyhappyaron: you'll need to apply12:46
Laneywiki page + mail to devel-permissions12:46
om26erLaney, I believe you wrote the sound panel ?12:47
happyaronLaney: go through a complete ppu procedure right?12:47
Laneyright12:47
om26erah I see a comment from you, lemme first try that12:47
Laneyom26er: erm, can't remember really, maybe parts12:47
happyaronok12:47
seb128willcooke, hey, is that one 30 min talk each you want?13:11
willcookeseb128, hey - sorry I missed you earlier.  Was travelling, and then straight in to meetings.13:11
seb128no worry13:11
willcookeseb128, so about the talk13:12
willcookeseb128, if you have a topic which you would like to talk about to them, then yes, 1 talk each13:12
seb128k13:14
om26erLaney, the bug that I reported does not happen if I execute from phone's terminal13:28
Laneyom26er: so this is intended behaviour13:29
Laneythose settings can only be changed by a present user13:29
LaneyI'm not sure how autopilot is run in this respect13:30
om26erLaney, when we login as phablet user isn't that the same ?13:30
Laneyno13:30
Laneythose are considered "inactive"13:34
willcookeattente, running a couple of mins late, sorry14:00
willcookeattente, ok, done - joining the hangout now14:01
seb128attente, he's at GUADEC with us and we are in a session14:02
seb128ups14:02
seb128willcooke, ^14:02
willcookeargh - I knew that14:03
willcookethanks seb12814:03
seb128yw!14:03
willcookesorry attente - we will catch up in person next week then :)14:03
Laneyput a shirt on for no reason? ;-)14:03
willcooke:D14:03
willcookerumbled14:03
=== m_conley_away is now known as m_conley
willcookeyay!  Just got my visa14:05
Laneynice one!14:09
seb128nice14:10
seb128what flight do you take to get there btw? same as us? or got a different one?14:11
om26erbug 134987114:42
ubot5bug 1349871 in ubuntu-system-settings (Ubuntu) "3G gets enabled on returning from Flight mode" [Undecided,New] https://launchpad.net/bugs/134987114:42
om26erI have probably reported it against the wrong package, whats the correct package for it ?14:42
=== alan_g is now known as alan_g|tea
=== alan_g|tea is now known as alan_g
willcookeerk15:06
willcookeSome of my Chromium windows have disappeared.15:06
* qengho appears!15:07
willcookee.g. I've got a hangout going in a separate window, and I can hear the hangout, and my webcam is live - but I can't "find" the window15:07
willcookeah15:07
willcookejust had a Chromium crash15:07
willcookeit's still running, but I have these missing window15:07
willcookes15:07
willcookeqengho, probably just a one off, but is there anything useful I can do?15:08
seb128oh, speaking of chromium?15:08
qenghowillcooke: I've seen some weirdness with F11 full screen and Unity and alt-tab switching. Were you doing anything like that?15:08
seb128what's the status of the 36 update?15:08
seb128it was to be uploaded during the week at our previous meeting15:08
seb128which was a week ago...15:08
qenghoseb128: I still have bugs. I think I'll have these fixed today.15:09
seb128should we upload 35 then?15:09
seb128the reason it was delayed is because 36 was almost ready15:09
seb128but if it takes almost a week the "almost" is not true anymore...15:10
qenghoseb128: I don't mind if ccc uploads 35. It costs me nothing.15:10
seb128chrisccoulson, ^15:10
seb128qengho, well, he was waiting on you to get 36 out which was supposed to happen during the week15:11
seb128that didn't happen15:11
seb128so maybe we should revisit the decision?15:11
willcookethat sounds like a good idea - as there might be more bugs in 36 which haven't been discovered yet.  If 35 is ready to go, I vote for uploading it15:14
qenghoseb128: For my part in the decision, I don't know these will be the last bugs. Every day has seemed like it would be like the last problem to fix.15:14
willcookeqengho, yeah - agree.  So let's roll with 35 then15:14
willcooke?15:14
seb128+115:14
qenghoseb128, chrisccoulson, okay, 35 now please.15:15
willcookecoolio15:15
seb128chrisccoulson, ^15:15
seb128thanks15:15
chrisccoulsonI guess the problem is that it isn't zero cost for me (it's about half a day to a day of manual tests for each of the releases)15:15
chrisccoulsonwhich is why I wanted to avoid doing it twice in the space of a couple of days15:15
willcookechrisccoulson, can they be automated?  Just trying to understand if we have any scope for reducing that time and effort15:16
seb128chrisccoulson, well, it's not a couple of days15:16
seb128it's turning into a couple of weeks15:16
seb128(again, seems to be a recurrent issues, we are often late to roll chromium updates)15:16
willcookeseb128, qengho, chrisccoulson - would it help if we target a particular schedule to release?  Like, every Tuesday or whatever?  So that if we have two versions closely following each other we don't have to duplicate work - if we miss the date then the next time a release can happen is 7 days later? Or is that just asking for trouble?15:18
qenghowillcooke: there are autopkgtests that could test, but it's probably hard to set up on earlier dists.15:19
willcookeqengho, so we're talking about a trade off between putting effort in to getting tests running on earlier dists vs the amount of effort required to do it manually?15:20
qenghowillcooke: I think that's the trade-off. Manual testing is more likely to assure nothing is broken, I think.15:21
qenghowillcooke: and it's not so often a release (not every week) that we get savings any time soon in setting up stable-release automated testing.15:22
willcookeriight - so quite a long time to see a return on investment15:22
willcookehrm15:22
willcookeno easy answer here then ;)  (Because if there was you would have already done it!)15:23
qengho:)15:23
willcookeSo - I think we should go with 35 now, just in case there are any more bugs, and then we can put some more thought in to how to improve things going forward.15:23
willcookeI dont think we need to *solve* the problem straight away, if we can make it a bit better, that's progress.15:24
qenghochrisccoulson: I'll need a day after this is done to get sign-off from webapps and qa anyway.15:24
willcookethen keep chipping away at it15:24
willcookeso, back to my original problem then :)15:24
willcookeqengho, I think what happened was something along the lines of...  super-L to lock screen, went to get a coke, came back and I could still see the desktop, but I couldnt click anything, went to vty 1, looked at syslog, libcontent.so segfaulted15:25
willcookeback to desktop15:25
willcookethe unlock screen appears, and I unlock15:25
willcookeback to my desktop15:25
willcookejoin a hangout15:25
willcookemove that tab to its own window15:25
willcookethen probably minimized it15:25
willcookethen went to look for it, but it's not there15:26
willcookelauncher icon shows only a single window15:26
qenghoHuh. libcontent is half the browser. :\15:26
willcookeyikes15:26
willcookeoohhhhhhhh15:26
willcookehold on15:26
willcookeI tell you what15:26
willcookeI was also dicking around with VPNs15:26
willcookeI wonder if a network glitch upset something15:26
willcookebah - I'll kill them and restart Chromium and see if I can reproduce it15:27
willcookeI expect I wont be able to15:27
qenghowillcooke: could be VPN, but I'd expect it to break in a different library then.15:28
willcookewooahhhh15:29
willcookeit reappear back where it was before I detached it in to its own window15:30
=== gatox is now known as gatox_lunch
willcookemaybe I dreamt the whole thing15:31
qenghoAh, good old libvoodoo.15:31
willcookejust a reminder - there is no weekly meeting this week because lots of people are away15:33
willcookeunless anyone has anything they want to discuss15:33
willcooke?15:33
seb128nothing specific from GUADEC afaik15:34
=== charles_ is now known as charles
willcookeseb128, did you by any chance bump in the guy about the Gnome Session bug15:36
willcooke?15:36
willcookeseb128 and did you meet happyaron yet?15:37
seb128willcooke, no, the gnome-session maintainer is not here afaik15:37
willcookedarn15:37
willcookenever mind15:38
seb128willcooke, yes, happyaron is there, we spent some time discussing input methods with attente and him earlier15:38
willcookeawesome!15:38
willcookeso qengho, chrisccoulson, seb128  - are we all set to upload Chromium 35 ?  Not sure if we reached consensus there.15:57
=== gatox_lunch is now known as gatox
=== alan_g is now known as alan_g|EOD
=== Ursinha-afk is now known as Ursinha
=== Ursinha is now known as Ursinha-afk
=== Ursinha-afk is now known as Ursinha
=== eeejay` is now known as eeejay
=== Ursinha-afk is now known as Ursinha
=== m_conley is now known as m_conley_away

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