/srv/irclogs.ubuntu.com/2009/04/17/#ubuntu-desktop.txt

philsfI can't make the guest account work from the fusa applet in latest jaunty, how can I debug it, run from command-line?00:43
chrisccoulsonwhat's not working exactly? what happens?00:46
philsfscreen blanks, seems to prepare to open the new session, flickers then returns to the previous open session00:49
wamcveyHey, anyone here actually an Ubuntu desktop developer?01:44
wamcveyok... I'll get less picky... anyone actually here?01:45
wamcveyWell... assuming that someone will eventually see what I type into this channel... could the desktop team *please* have someone look at https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/216550  I would be happy to work with anyone who is interested in actually getting some movement forward on this bug (e.g. assigned to someone would be *great*).01:51
ubottuUbuntu bug 216550 in compiz "setting a hotkey for raise/lower window doesn't work" [Undecided,Confirmed]01:51
wamcveyjust as a bit of context... it's a bug that's been around for over a year now... and reportedly is still an issue in Jaunty.01:52
pittiGood morning08:11
celthunderpitti, good morning to you too08:15
Keybukpitti: summit gets that information from LP, yes09:21
pittiKeybuk: ah, nice09:21
Keybukfiddling in summit doesn't get merged with LP, obviously09:21
Keybukthere's no API for blueprints yet09:21
=== dpm_ is now known as dpm
=== eeejay is now known as eeejay_afk
=== eeejay_afk is now known as eeejay
chrisccoulsonpitti - just saw your comment on bug 346912, and noticed you added a jaunty task13:21
ubottuLaunchpad bug 346912 in tracker "Tracker index corruption (was Tracker does not stop indexing)" [Medium,Triaged] https://launchpad.net/bugs/34691213:21
chrisccoulsoni've been spending a fair bit of time on #tracker, and I get the impression that the corruption issues require quite big changes to tracker, so it's unlikely to be fixed in Jaunty13:22
pittichrisccoulson: it was already nominated, I just accepted it13:24
pittichrisccoulson: right, that might be; we can still "wontfix" it in jaunty13:24
chrisccoulsonpitti - ah, ok, i didn't realise13:24
pittithere might still be a smaller patch to work around it, though, like deleting the index if corruption is detected, or so?13:25
chrisccoulsonthe corruption issues are not that well understood upstream, but they believe it is down to the general limitations of qdbm, which they are completely dropping in 0.7.x13:25
chrisccoulsonpitti - the latest version of tracker detects the corruption, and gives the user the option of reindexing13:25
pittioh, we are using that? I thought it used sqlite or so13:25
pittichrisccoulson: where is that option?13:26
pittidoes it pop up, or do you have to open some menu for that?13:26
chrisccoulsonpitti - it should just happen automatically. but there is already a bug report open for that, because it uses a notification with actions13:26
kenvandine_wkis there a way to corrupt it so we can test?13:26
pittiah; so it should be a dialog box13:26
chrisccoulsonso you get the standard fallback notification with "Ok" and "Cancel" buttons, which do nothing13:27
chrisccoulsonthere's a bug report for that, because "Ok" and "Cancel" just cause the notification to reappear again straight away13:27
kenvandine_wki never saw that in the list of packages to fix13:27
pittichrisccoulson: outh13:27
chrisccoulsonkenvandine_wk - the notification was added in the upload this week13:27
pitti"outch" even13:27
kenvandine_wkchrisccoulson: oh... :/13:27
chrisccoulsoni was going to edit the notifyosd wiki to add it as an application to fix13:27
chrisccoulsonbut i didnt get time13:27
kenvandine_wkchrisccoulson: tag the bug "notification"13:28
chrisccoulsoni'll do that13:28
kenvandine_wkthx13:28
pittiso, silbs indeed got it13:28
kenvandine_wkso it shows up on our radar at least13:28
pittiis there a workaround to force reindexing without the dialog}13:28
pitti?13:28
kenvandine_wkperhaps not even use a notification... if it is corrupt, there is no other solution13:29
kenvandine_wkright?13:29
chrisccoulsonpitti - perhaps we could just patch it to reindex automatically13:29
pittirm -r .cache/tracker/ or so?13:29
chrisccoulson^^what kenvandine_wk suggested13:29
pittichrisccoulson: right; I guess there's no other choice anyway13:29
chrisccoulsononce the index is corrupt, there is no other logical way out13:29
kenvandine_wkprobably dirt easy :_13:29
kenvandine_wk:)13:29
chrisccoulsoni could perhaps look at doing that13:29
pittithat would be awesome13:30
chrisccoulsonbut i've got no easy way of triggering database corruption to test:(13:30
kenvandine_wkjust look at whatever function is called when they act on the notification... and just call that instead of the notification13:30
kenvandine_wkchrisccoulson: i guess there is no test suite :/13:30
chrisccoulsonone of the theories upstream that might be causing corruption is interrupting the indexer whilst it is updating the database13:30
chrisccoulsonthat should be easy, as tracker has no integration with the session manager13:31
chrisccoulsonso you can trigger that just by logging out13:31
kenvandine_wkeww13:31
kenvandine_wkbut... the db should be getting updated in a transaction, so it would just rollback13:32
kenvandine_wki would think13:32
chrisccoulsonkenvandine_wk - possibly. i don't know enough about databases to comment though13:32
chrisccoulsonthat might be one of the limitations of qdbm possibly13:32
chrisccoulsonright, i've got to disappear now to travel home from work. i should be around when i get home though13:34
kenvandine_wkthx13:34
seb128pitti: how busy are you today? ;-)13:46
seb128pitti: I uploaded the gvfs,nautilus change for smb ad to hardy-proposed now13:46
pitticurrently reviewing unapproved queue, then I need to prep release meeting, and then do the meeting13:46
seb128pitti: if you have some time to review the uploads that would be welcome13:46
pittiseb128: SRU was also on my list13:46
seb128the changes are the same as the ones uploaded to intrepid some months ago13:46
seb128so should be no real surprise13:47
pittithanks13:47
chrisccoulsonpitti - would you mind adding a jaunty target to bug 36120514:09
ubottuLaunchpad bug 361205 in tracker "corrupt index on tracker" [Medium,Confirmed] https://launchpad.net/bugs/36120514:09
chrisccoulsonthat's the one about the fallback notification14:10
pittidone14:10
chrisccoulsonthanks:)14:10
kenvandine_wkchrisccoulson: just dropping the index also solves the notification issue... 2 birds, 1 stone... good stuff :)14:11
kenvandine_wkchrisccoulson: don't forget to tag it though14:11
chrisccoulsonkenvandine_wk - no problem:)14:13
kenvandine_wkthx14:14
rickspencer3ArneGoetje: asac: bryce: calc: kenvandine_wk Riddell seb128 pitti: what's the word on the street regarding the RC?14:24
kenvandine_wkrickspencer3: quite positive14:24
kenvandine_wktwitter/identi.ca has been going nuts :)14:25
kenvandine_wkmostly happy14:25
seb128rickspencer3: looks good to me out of all those intel discussions but that's not a stopper and I think I read rather concern on this channel than on other places so let's see ;-)14:25
asacrickspencer3: plan to do my first post-RC bug round a bit later ... will report ;)14:25
pittihey rickspencer314:25
pittirickspencer3: as usual, tons of uploads in unapproved, trawling through ATM14:25
seb128rickspencer3: GNOME wise the amount of bugs is reasonable, nothing urgent or unexpected so far reported14:26
rickspencer3seb128: yes, I noticed that we seem to be more concerned about -intel then the bloggers and reviewers14:26
RiddellI think intel problems affect qt more than gtk for some reason14:28
asacpitti: fta asks whats the policy for bringing ia32libs in sync with RC status.14:28
pittiasac: no policy, but sounds like a good idea, if it can be tested with flash and some other usages14:28
asacis that a SRU thing or are we doing this a few days before release usually14:28
asacfta: can you upload the new package to some PPA so we can test it?14:29
pittiideally we should do it ASAP14:29
pittiand since it's universe, it's not bound to CD freezes14:29
asacfta: is there a bug we can use to track this roll out?14:30
asacotherwise i can open one for you14:30
* asac on call14:33
rickspencer3Riddell: yah, I noticed that the Kubuntu users have slightly worse problems14:34
=== sabdfl1 is now known as sabdfl
=== rickspencer3 is now known as rickspencer3-afk
pittirickspencer3-afk: FYI, https://wiki.ubuntu.com/DesktopTeam/ReleaseStatus updated14:53
chrisccoulsondoes notify-osd send out a "closed" signal when it has finished displaying a notification?14:53
wamcveyAnyone awake here yet?14:55
chrisccoulsonnobody ever sleeps here14:55
fta2asac, ?14:55
wamcveywas here last night and the place was dead quite...14:55
wamcveyI was hoping I could nudge someone to take a look at a compiz issue that's been lingering around for over a year and still doesn't have an assigned owner on launchpad.14:56
chrisccoulsonwhich issue?14:56
wamcveyapparently the issue persists into Jaunty14:56
wamcvey https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/21655014:56
chrisccoulsonhave you tested it in jaunty?14:56
ubottuUbuntu bug 216550 in compiz "setting a hotkey for raise/lower window doesn't work" [Undecided,Confirmed]14:56
=== rickspencer3-afk is now known as rickspencer3
wamcveyI personally haven't, but one of the followers (David Fox) has confirmed it still exists with Jaunty14:57
chrisccoulsoni can't help you much further with that as I don't have much knowledge of compiz, and I don't really have time to test it out yet either14:59
wamcveyChris, I understand... I asked over on #ubuntu and that place was a zoo... Was hoping this smaller audience focused on the desktop would have a more likelyhood of finding someone who can at least take ownership of the issue15:01
mvowamcvey: I think this is a side-effect of a bug already filed, give me a minute to find it15:01
mvobug #35265615:03
ubottuLaunchpad bug 352656 in compiz "Ignores GNOME preferred terminal application setting" [Unknown,Fix released] https://launchpad.net/bugs/35265615:03
mvohm, maybe not15:04
wamcveymvo: yeah... doesn't seem like the same issue.15:05
asacfta2: ok. so what we want is a ppa upload, with a bug so we can test and sign it off15:12
tedgchrisccoulson: Ask on #ayatana, but I'm pretty sure it's supposed to.15:16
chrisccoulsontedg - thanks15:16
pittidavidbarth, MacSlow: you are 100% sure that the patch in https://bugs.launchpad.net/ubuntu/+source/notify-osd/+bug/331369/comments/27 fixes the multihead problem? "focus-follow" sounds quite unrelated to that15:19
ubottuUbuntu bug 331369 in notify-osd "regression vs. notification-daemon: positioning when multiple screens are available" [Medium,Fix committed]15:19
chrisccoulsonpitti - i've got a patch for tracker to automatically reindex when the index becomes corrupt15:20
pittichrisccoulson: nice!15:20
davidbarthpitti: it does finish the job15:20
pittichrisccoulson: this should make a fine SRU (doesn't affect new installs, since we don't install tracker by default)15:20
davidbarthpitti: the patch that is already in the release does fix the main issue15:20
chrisccoulsoni could do with a guinea pig to try it out on now:)15:20
davidbarthpitti: ie the fact that we're now displaying bubbles next to the gnome-panel15:21
kenvandine_wkchrisccoulson: i can test that15:21
pittichrisccoulson: put it in the bug and ask silbs to test15:21
chrisccoulsonthanks:)15:21
pittidavidbarth: so you confirm it's the right patch? I was just wondering15:21
pittidavidbarth: (this is relative to 0.9.11, I don't upgrade to 0.9.1215:21
wamcveysorry to be repetative... but is there anything more than I can/should do to try and get movement on the bug?15:21
davidbarthso 0.9.12 contains the one-liner15:23
davidbarthso 0.9.11 is the current version that fixes most of the multihead issue15:23
davidbarththe fix is really about enabling the "advanced" focus-follow mode15:24
davidbarthstrictly speaking it is not a hard requirement for being multi-head compatible15:24
pittidavidbarth: ah, ok15:24
davidbarthit's just that the bug would be cleanly closed this way15:24
pittidavidbarth: the bug is chalked down for ubuntu-9.0415:24
pittiI'm fine with that cherrypick15:24
davidbarthwhich means? going in?15:24
davidbarthcool15:24
davidbarththanks for our users ;)15:24
* pitti uploads, thanks for confirming15:24
wamcveyChecking over on #compiz-fusion if/when this feature was added into compiz.  (it's still a bug in Ubuntu though but maybe I can solve the problem myself by building my own compiz. bleh)15:32
asacwhat was the trick get keyring unlocked automatically if i use automatic-login?16:14
asacnevermind. i made a system connection out of it ;) ... now no keyring hazzles on startup16:18
chrisccoulsonpitti - my tracker patch seems to work ok16:20
chrisccoulsoni flipped a few bits in the index to corrupt it ;)16:20
pitti\o/16:21
wamcveyFYI: regarding ticket https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/216550  , the vigo plugin to compiz (http://www.acc.umu.se/~janlert/vigo/) provides key binding for raise-lower functionality. I'll update the ticket with the info on getting it compiled and installed.16:28
ubottuUbuntu bug 216550 in compiz "setting a hotkey for raise/lower window doesn't work" [Undecided,Confirmed]16:28
pittiseb128: heh, cjwatson and you are on par on http://qa.ubuntu.com/reports/bug-fixing/jaunty-fixes-report.html; but cjwatson has some more fixes in the pipeline; quick, fix something and overtake him :)17:05
seb128pitti: we should count jaunty-updates there too ;-)17:45
pittiseb128: are you experiencing bug 359392?17:48
ubottuLaunchpad bug 359392 in xserver-xorg-video-intel "[i965] X freezes starting on April 3rd" [Critical,Triaged] https://launchpad.net/bugs/35939217:48
seb128pitti: no17:48
pittiseb128: running with any special xorg.conf options such as greedy or EXAOptimize?17:49
bryceseb128: do you run compiz on?17:49
mvoseb128: or desktop effects ;)17:49
seb128xorg is rocking stable since the fast user switching fix there17:49
seb128bryce: yes17:49
bryceok17:49
seb128no, stock xorg17:49
seb128I've no speed issue and never felt the need to play with options17:49
seb128it's on my d630, i96517:50
seb128I've virtual section in xorg.conf set from the xrandr capplet17:50
seb128but that's all17:50
seb128I'm using compiz but nothing fancy there17:51
seb128ie I'm only doing alt-tab and workspace switches17:52
seb128I'm not using the cube or any other combo17:52
seb128pitti: bug sent upstream = triaged for us18:28
pittiseb128: right, but the upstream bug is not very useful, it has about zero information18:29
pittiwe got asked in the release meeting to at least understand it, since it causes a lot of unrelated bug noise in openssh and general confusion18:30
seb128pitti: so put it to incomplete if it lacks infos18:31
seb128pitti: it's either triaged or incomplete there18:32
pittiright, makes sense18:32
seb128pitti: I'm not sure to understand the bug description or the issue though18:32
seb128are we sure that's something that ever worked18:33
mvoseb128: what is the pci id of your vga card?19:11
YokoZarQuick "color of the bikeshed" question regarding Wine icon: http://lists.freedesktop.org/archives/tango-artists/attachments/20060331/63117187/wine.svg   or http://yokozar.org/blog/content/icons/wine.svg  ?19:21
chrisccoulsonmmmmm, wine19:22
* chrisccoulson goes to grab a bottle19:22
YokoZarchrisccoulson: are you using "bordeaux"-style glasses or "burgundy"-style (thin or thick)19:27
chrisccoulsonbordeaux ;)19:28
chrisccoulsonare your icons intended for any particular theme? or just to ship with wine?19:29
nhainesYokoZar: I like the freedesktop.org one the best, but I see that the yokozar one would match some other themes better.  Maybe if the first one had bolder outlines?19:38
YokoZarnhaines: tango icons (and they're both tango-compliant) are supposed to have extra highlighting in their smaller sizes (that is, the 1 pixel border/highlight isn't scaled down, but remains at 1 pixel when the image is 22x22 or 48x48)19:40
nhainesThe first one looks more elegant and the one on your blog looks more... I donno, more like a tango icon.  :)19:41
YokoZarnhaines: ironic as the first was submitted to the tango list and the latter is pulled from Ubuntu studio ;)\19:42
nhaineshehe19:43
YokoZarbut it's interesting you think the fatter one looks more elegant19:43
nhainesI love the lines in yours, but (purely subjectively) not the wide shape.  :)19:43
YokoZaroh19:43
nhainesNo, I think the thin one looks more elegant.  But I don't like the lines.19:43
YokoZaryou're mixing them up, hah19:43
nhainesI just realized I may have mixed them up.19:43
nhaineshaha, sorry.  :)19:44
pittirickspencer3: FYI, I added desktop startup speed to the https://wiki.ubuntu.com/UDSKarmic/Prep/Desktop brainstorm list19:54
rickspencer3sweet19:54
rickspencer3I suppose folks might as well go ahead and create blueprints if they wsih19:55
seb128pitti: good luck with that19:55
pittiseb128: ugh, yeah19:56
seb128pitti: I think we spent some time on that for jaunty and will not get much better result without replacing gconf bonobo etc19:56
seb128which is GNOME3 tasks upstream already19:56
pittiseb128: at least points 3, 4, 5 should be doable, though19:56
pittiseb128: ah, gconf will be replaced by dconf in 2.30?19:56
seb128pitti: that's the plan19:57
pittinice19:57
seb128pitti: desrt is paid to work on that and will be at uds19:57
pittinot in 2.28 yet, though?19:57
pittiand, will it be any faster? :-)19:57
seb128the plan is to have be there but don't expect everything to use it19:57
seb128pitti: http://mail.gnome.org/archives/desktop-devel-list/2009-April/msg00028.html19:58
pittiok, so potentially that's something we could help with19:58
pittiseb128: thanks19:58
seb128"dconf is very efficient. The majority case in accessing settings is reading (think about desktop login: 1000s of settings read, none written)."19:58
seb128pitti: ^19:58
seb128ups19:59
seb128"dconf is very efficient. The majority case in accessing settings is reading (think about desktop login: 1000s of settings read, none written). Reading in dconf is done directly from a memory-mapped file containing the settings in an efficient tree format and doesn't require an extra service to be running. The service is only needed for writes.  Communication occurs over DBus, of course. :) "19:59
seb128rather19:59
seb128pitti: I think the main offenders in gnome-panel and nautilus are gconf and bonobo20:00
YokoZardoes dconf mean settings will become cross-desktop eventually?20:00
seb128I'm not sure we will benefit from any discussion there, nautilus upstream already said he's happy to work on speed issue if we provide datas of what is slow20:00
seb128YokoZar: dunno, depends if other desktops environment are wanting to use it20:01
YokoZarseb128: I guess my question then is should there be freedesktop standards for things moving out of gconf and into dconf20:01
seb128it will not have GNOME depends20:02
=== rickspencer3 is now known as rickspencer3-afk
seb128the question is not about a standard, it's to know if other desktop which have something working for them are wanting to do the effort to switch20:02
seb128and that's not a GNOME question so dunno20:02
YokoZarahh I see20:02
seb128pitti: your items there seem rather bugs than things which need discussion20:04
pittiseb128: well, it's not immediately clear to me how to keep a gpg agent when not starting seahorse with the sessions20:05
pittisince it currently relies on those silly env vars20:05
seb128we already start seahorse feature depending on gconf keys20:06
pittiYokoZar: with KDE4 just having introduced akonadi, I have only little hope :/20:06
seb128I'm not sure that' worth the 1 second difference20:06
seb128and anyway that's rather a thing we should submit upstream now and wait for comments20:06
pittiseb128: if your budget is really 4 seconds some day, it will20:06
seb128rather than a topic discussion for uds20:06
pitti(it's a whopping 3 seconds for me)20:06
pittibut anyway, it's just a brainstorm list20:06
pittiI'm sure there's some more things we can do to speed it up20:07
seb128well just saying that I doubt we will have better ideas in a session than right now20:07
seb128it just require somebody working on the issue20:07
YokoZarspeaking of gpg and seahorse... I would like to talk about how we have three separate places where one might look for keyring prompts20:07
pittiI just can't commit to us doing 10 manmonths of upstream work, as Keybuk proposed20:07
pittiseb128: well, if the solution to seahorse or the xrdb stuff is already entirely clear, then sure20:08
YokoZarApplications-> Accessories -> Passwords and Encryption Keys, System->Preferences->Encryption and Keyring, and System->Administratin->Authorizations20:08
seb128pitti: well, it has been discussed at one uds and one sprint already20:08
seb128pitti: we had discussions, bootcharts and bugs about those issues20:08
pittiseb128: what was the approach for seahorse then?20:09
seb128pitti: I don't think discussing it again will bring anything new, we just need to do the work20:09
pittiright, sorry then; I wasn't aware of those20:09
seb128pitti: we have some feature starting conditionally in jaunty, for the gpg agent we don't have a clever solution yet20:09
seb128I'm not sure that a round table will bring it one if nobody has a clue on how to do that right now and didn't during previous discussions20:10
seb128$ grep enable /etc/xdg/autostart/seahorse-daemon.desktop20:10
seb128AutostartCondition=GNOME /apps/seahorse/sharing/sharing_enabled20:10
YokoZarWe probably should talk about the UI of the key stuff though.  There are 3 different programs with key icons in 3 different menus20:10
seb128pitti: ^20:10
seb128YokoZar: you wrote that 2 minutes ago already, no need to repeat it again ;-)20:10
dobeyhmm20:11
YokoZarseb128: I meant as a blueprint ;)20:11
seb128YokoZar: you can do that, there is lot of discussions about merging capplets, etc, that require upstream work and we don't have so much free cycles for that20:11
pittiseb128: well, that's not for GPG, is it? the gpg agent is /etc/X11/Xsession.d/60seahorse-plugins20:11
dobeyi wonder if there will be more kde art people there this time20:11
=== rickspencer3-afk is now known as rickspencer3
YokoZarIt seems like seahorse should be a capplet itself (or merged into one)20:12
seb128pitti: right, as said "some feature", we have no clear idea about this one20:12
seb128pitti: I'm happy discussing it again but I feel it's rather a technical issue than a discussion one, ie require work and thinking rather than a discussion20:13
pittiseb128: right, that needs to be done before UDS20:13
asacwhat can we do for bug 36283620:25
ubottuLaunchpad bug 362836 in network-manager-applet ""3G networks detected" notification bubble displays wrong icon" [Low,Triaged] https://launchpad.net/bugs/36283620:25
asacdxteam wants to get that fixed ... would be UI change (even if tiny)20:26
asacimpact is https://bugs.edge.launchpad.net/ubuntu/+source/network-manager-applet/+bug/362836/comments/520:26
ubottuUbuntu bug 362836 in network-manager-applet ""3G networks detected" notification bubble displays wrong icon" [Low,Triaged]20:26
asacpitti: whats your opinion? or should i rather ask in -release?20:27
pittire20:39
pittiasac: sorry, got stuck in my mail client20:39
pittiasac: erm, now?20:41
pittiasac: let's say I'm not exactly fond of that20:41
pittiasac: technically possible, if the docteam gives their ack, but I'd rather have a sabdfl exception for that20:41
pittigood night everyone20:56
=== rickspencer3 is now known as rickspencer3-afk
kenvandine_wkgood night pitti20:58
=== asac_ is now known as asac
=== rickspencer31 is now known as rickspencer3

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