/srv/irclogs.ubuntu.com/2011/11/23/#ubuntu-desktop.txt

RAOFHm.  That's not a test, but it is a nice demonstration of how the nvidia drivers can fail :)01:34
TheMusoheh01:39
=== eeejay is now known as eeejay_is_afk
robert_ancellRAOF, um, is xinerama undocumented?03:00
=== eeejay_is_afk is now known as eeejay
RAOFrobert_ancell: Xinerama is basically server-internal; it wouldn't surprise me to learn that it's undocumented.05:06
robert_ancellRAOF, yeah, I found out they were going to document it but gave up in the end now its obsolete05:06
robert_ancellit's pretty simple though05:07
robert_ancellRAOF, Xinerama = 1 screen with xinerama virtual screens right?05:07
robert_ancelland XRANDR = n screens each with n virtual screens (or monitors in GDK language)05:07
RAOFNot really; there's still only one protocol screen in XRandR.05:08
RAOFXinerama is still around for the multi-GPU case - you have a single protocol screen and rendering etc gets duplicated across all physical devices.05:08
robert_ancellRAOF, so to use Xinerama/XRANDR you get rid of the multiple screen stuff in the connect message?  i.e. there can only be :0, no :0.0, 0.1...05:09
RAOFYou won't get :0.0 :0.1, etc out of an XRandR setup.05:09
RAOFI'm less certain about a Xinerama setup.  Maybe you will?05:09
robert_ancellRAOF, looking at the protocol I don't think you can05:10
RAOFHowever, if you're only interested in compiz then you don't need to care; Xinerama implies software rendering.05:10
robert_ancellRAOF, well, we will need tests for compiz with xinerama, with randr to check it handles these cases correctly05:10
RAOFI guess there are two questions here - the Xinerama hints, which are still used with randr, and the Xinerama span-a-desktop-across-multiple-gpus thing.05:12
RAOFThe latter you shouldn't need to care about; Unity just won't load there because GL doesn't work.05:13
robert_ancellRAOF, that's the rendering side, it's not exposed on the protocol how the rendering is actually done05:13
robert_ancellRAOF, I'm going to look at how hard it will be to fake that :)05:13
robert_ancellRAOF, it's just one /dev device right?05:13
RAOFUm, I've lost context.05:14
RAOFWhat's just one /dev device?05:14
robert_ancellRAOF, to support direct rendering I just have to intercept the client trying to open a file in /dev right?05:14
RAOFYes, that'd be right.05:15
pittiGood morning05:23
=== eeejay is now known as eeejay_is_afk
TheMusoMorning pitti.05:27
pittihey TheMuso, how are you?05:30
pittiTheMuso: did you happen to try multi-arch-ification of the at-spi packages? any problems in GTK for this?05:30
TheMusopitti: There doesn't seem to be any problem so far with my testing.05:31
pittioh, you did, just saw it on -changes@; nice!05:31
pittiTheMuso: I haven't uploaded GTK to Debian yet, I want to wait until the current one goes to testing05:31
TheMusoRight.05:32
pittithen merging or even syncing sohuld be eaier05:32
pittieasier05:32
TheMusoYep.05:32
pittijbicha: hey Jeremy05:36
pittijbicha: do you have a particular attachment to meta-gnome3, i. e. do you want to merge it?05:36
pittiotherwise I'll do it now, as it's uninstallable due to libgail-3-common going away05:37
pittijbicha: (the hamster-applet diff can go away now, I think)05:38
jbichapitti: you're welcome to do the merge if you like :)06:14
didrocksgood morning06:38
pittibonjour didrocks06:39
pittididrocks: congratulations for landing the unity tarmac setup!06:39
didrockspitti: heh, thanks :)06:40
didrockspitti: now, dx has to fix make check as make check in unity is failing :)06:41
didrocksbut all others 14 components seems to pass06:41
pittinice06:43
pittididrocks: does that merger actually enforce writing tests for new code06:43
pitti?06:44
didrockspitti: not yet, but that will be a tarmac plugin I plan to write (for now, let's have the basic working), that will be fairly trivial06:45
didrockspitti: basically, I'll write a plugin forcing linking to a bug, and having new tests06:45
didrockspitti: see the "Future improvments" section06:46
pittididrocks: "fairly trivial" sounds like an understatement, though :)06:48
pittiwell, there are certainly many commits which don't need changing the tests06:49
pittia whole MP should certainly at least touch it, of course, unless it's a kind of "doc cleanup" (where you have that override tag)06:49
pittibut we also had lots of "change the design" FFEs/changes which can't be covered by tests06:50
pitti(grid layout, icons, colors, etc.)06:50
pittiso maybe something like "if the MP changes N lines of actual code (.cpp), test suite must at least change by N/5 lines" or similar heuristics?06:50
pittiah, no; design changes are implemented in code, too06:51
pittididrocks: but anyway, for now it's probably more a socially enforced thing :)06:51
pittirobert_ancell: hey Robert, how are you?06:54
pittirobert_ancell: so you landed Gunnar's language selector lightdm changes in precise?06:54
pittirobert_ancell: there's https://code.launchpad.net/~gunnarhj/ubuntu/oneiric/lightdm/lang-chooser/+merge/83031 for oneiric, I wondered how you would like to handle this06:54
pittirobert_ancell: should I do the review/merge, do you want to, do you think we should let this mature in precise for a bit?06:55
didrockspitti: sorry, was merging some branch07:00
didrockspitti: yeah, I still want to do the heuristic of "N line of changes, M lignes of tests"07:00
didrockspitti: with the override tag, but reporting the result somewhere to prevent abuses07:01
mvopitti: silly question, but who can set priorities for a blueprint? it seems like I can't even though I'm the approver https://blueprints.launchpad.net/ubuntu/+spec/consumer-p-software-center-enhancements07:42
mvo(this one is a example :)07:42
pittimvo: I'm not sure, ~ubuntu-drivers perhaps?07:42
pittiat least that _should_ be the team which can07:43
* mvo looks07:43
pittimvo: try again now?07:43
pittimvo: I added you07:43
mvo\o/07:44
mvothanks, I can now!07:44
* mvo hugs pitti07:44
pittimvo: SUPAPOWAHS!07:44
* pitti hugs mvo back07:44
mvohttp://www.supermanhomepage.com/images/superman-returns7/sr-poster2.jpg07:44
didrockswaow, so much power for mvo :)07:45
pitti*shrug* he could always upload something to precise which will silently click the buttons in my browser for him :)07:46
didrockspitti: I'm curious, if there is a promotion in main for adding a build-dep to a package, but then, the binary packages are not supposed to be installed anywhere, should I still need to put in a seed like supported? or the fact that it's a build-dep of a package in main is enough to not shown in component-mismatched?07:46
pittithe powers of being root on millions of machines07:46
didrocksyeah, that's frightening :)07:46
pittididrocks: those packages should explicitly _not_ be seeded07:47
micahgpitti: there was a drive to empty that team ( no pun intended )07:47
pittididrocks: indeed, main is closed under depends, recommends, and build-depends07:47
didrockspitti: excellent, thanks! :)07:47
pittimicahg: I thought we wanted to make the purpose of it more explicit, i. e. use it for planning (blueprints mainly)07:47
pittiin the past it had a whole lot of other privs, like bug targetting etc, but that either already went a way, or is meant to be07:48
micahgpitti: i'm not sure of all the details, but uds-organizers is a member of drivers so it seems that's a part of it07:48
pittimicahg: ah, that could be indeed07:49
rodrigo_morning08:43
didrockshey rodrigo_, how are you?08:45
rodrigo_bonjour didrocks08:46
pittihey rodrigo_, how are you?08:50
rodrigo_hi pitti08:51
pittiseb128: please ignore the current retracer crash; flaw in my new client-side duplication logic, have the fix ready now09:15
pittiseb128: it -- OMG! -- detected an actual duplicate09:15
pitti(the previous logic assumed that they would not be actually uploaded)09:16
seb128hey pitti09:16
seb128;-)09:16
pittiworking again now09:19
glatzora morning seb128 pitti mvo rodrigo_09:24
rodrigo_hi glatzor09:25
rodrigo_and seb12809:25
didrockshey seb12809:28
mvohey glatzor!09:30
seb128hey glatzor didrocks rodrigo_09:31
seb128hey mvo pitti09:31
mvohey seb12809:31
rodrigo_hey seb128, mvo :)09:31
mvohello rodrigo_09:31
chrisccoulsongood morning09:33
seb128hey chrisccoulson09:33
seb128how is everybody doing today?09:34
chrisccoulsongood, i think :)09:34
chrisccoulsonhow are you?09:34
seb128chrisccoulson, quite good thanks ;-)09:41
seb128didrocks, congrats on the dx landing infrastructure work09:47
seb128looks great by reading your detailed email09:47
seb128didrocks, thanks for the email btw, nice to see people giving details about what they are doing ;-)09:47
didrocksseb128: yeah, I think details was needed, despite the length of it :)09:47
didrocksseb128: and thanks! :)09:48
glatzorpitti, python-packagekit, libpackagekit-glib2-14 and gir1.2-packagekitglib-1.0 need to go through the full MIR process again?10:03
glatzorpitti, I would like to see that the session indicator is ported to PackageKit10:04
seb128glatzor, what's the issue with the current code? is the aptdaemon interface it's using going to be dropped?10:05
glatzorpitti, aptdaemon doesn't provide the "show me if there are updates" method natively. So they made a kind of hack by simulating a system upgrade after each (!) finished transaction10:05
seb128oh ok10:06
seb128glatzor, well, knowing dx it will probably take until after feature freeze for them to work bug fixes or that sort of cleanups10:07
chrisccoulsonlol10:07
seb128chrisccoulson, what part of that log is a lol one? ;-) (I was not really joking for dx)10:08
chrisccoulsonseb128, oh, i thought you were joking. aren't we meant to be fixing that this cycle? (ie, not breaking everything and then waiting until after feature freeze to fix everything)10:10
seb128chrisccoulson, well, their code is hackish but sort of work10:10
seb128we said we wouldn't let them land broken code yes10:10
seb128but knowing them I doubt they will rework code which "works" before feature freeze10:11
seb128though I'm happy to be proven wrong ;-)10:11
glatzorseb128, pitti by design aptdaemon was a solution which did not provide any "query" methods since we did all of those in the client application10:17
glatzorseb128, pitti but the packagekit dbus interface has got the UpdatesChanged signal to notify clients when it is a good time to query for updates. since I implemented the pk interface this was a low hanging fruit10:18
seb128glatzor, can you open a bug about that against indicator-session?10:18
seb128glatzor, we will get assigned and triaged10:19
seb128but I can't promise that dx will fix it soon10:19
seb128they tend to try to land their features first and look at bugs fixing after feature freeze by experience10:19
pittiglatzor: if those are just new binaries, no MIR necessary10:33
=== tkamppeter_ is now known as tkamppeter
glatzorpitti, packagekit moved to universe - perhaps since it isn't used by kubuntu anymore10:45
glatzorpitti, https://wiki.ubuntu.com/MainInclusionReportPackagekit10:45
pittiglatzor: right, the existing MIR is still valid10:45
pittiwe just need somethign that pulls it back in, i. e. I figure a dependency10:45
glatzorpitti, thanks. fine.10:49
chrisccoulsonhmmm, it's really bad that this is considered news: http://www.omgubuntu.co.uk/2011/11/firefox-8-finally-lands-in-ubuntu-11-10/ :(12:06
=== MacSlow is now known as MacSlow|lunch
rodrigo_hmm, any python expert -> doesn't python have sethostname? I can't find it in any module12:15
pittirodrigo__: hm, according to http://bugs.python.org/issue10866 it was added in February12:43
rodrigo__pitti, ok, guess I was finding outdated pages12:44
pittibut it doesn't seem to be in the release yet12:44
rodrigo__ok :(12:44
pittirodrigo__: subprocess.call(['hostname', '...']) ?12:45
rodrigo__yeah12:45
mvorodrigo__: http://paste.ubuntu.com/746995/plain/ <- this will work too13:05
rodrigo__mvo, right, thanks!13:05
mvoI love ctypes, its a bit like a gun with no safety catch, but very useful at times13:08
rodrigo__mvo, :)13:13
rodrigo__lunch time, bbl13:15
=== MacSlow|lunch is now known as MacSlow
seb128Sweetshark, hi, is bug #749986 a "known issue" for you? do you know if it's a theme issue or unity issue or gtk issue?15:10
ubot2Launchpad bug 749986 in libreoffice "Resizing LO window from bottom right brings up pop-up and undesirable effects" [High,Confirmed] https://launchpad.net/bugs/74998615:10
seb128it seems it has to do with the grips added to gtk and the theme from the comment15:11
seb128seems also the sort of bug we should fix in this cycle15:11
seb128Sweetshark, seems its not wm dependent, likely libreoffice is trying to do handle the gtk widgets itself and doing it wrongly15:18
seb128pitti, \o/15:21
pittiseb128: what did you do?15:22
ogra_gymnastics :)15:22
seb128pitti, me? nothing. you, "apport-request-retrace" \o/15:22
pittiaah ;)15:22
seb128pitti, there is small potential confidentiality issue in doing things that way though, not sure if that's concrete enough to be an issue though15:23
seb128pitti, i.e that will add debug retracing to a public bug without triager verifications right?15:23
pittiseb128: right, if the master bug gets public and teh dupe has confidential data in the stack trace15:23
seb128ok15:24
=== mhr3_ is now known as mhr3
=== eeejay_is_afk is now known as eeejay
Sweetsharkseb128: cant reproduce here with fluxbox and libreoffice-gtk installed15:42
Sweetsharkseb128: so I wouldnt rule out unity there.15:42
seb128Sweetshark, it happens under gnome-shell or unity-2d15:43
seb128Sweetshark, did you try clicking on the triangle inside the libreoffice dialog? i.e not exactly in the corner itself15:43
Sweetsharkseb128: heh, there is no such triangle on fluxbox.15:44
seb128Sweetshark, that has nothing to do with fluxbox15:45
seb128though it has to do with the gtk theme15:45
seb128do you use the ubuntu gtk theme under fluxbox?15:45
* Sweetshark logs into unity.15:46
Sweetsharkseb128: I can reproduce it with unity and default theming.15:48
seb128it happens the same way under unity-2d or gnome-shell15:48
seb128so it's not unity15:49
Sweetsharkseb128: yep, I didnt hit close enough to the corner. I can reproduce with unity2d, xfwm4 and fluxbox too15:53
seb128Sweetshark, great ;-)15:54
seb128Sweetshark, note that added those grips to gtk2 is an Ubuntu patch15:54
Sweetsharkseb128: also reproducable without ubuntu/debian patches on libreoffice-master (pre-3.5)15:55
Sweetsharkseb128: ah, so nobody else noticed because we tried something extra special again? ;)15:56
seb128Sweetshark, right15:59
seb128Sweetshark, well that feature is in gtk3, we just backported it to gtk216:00
Sweetsharkseb128: ah, thats nice to know to raise interest in it upstream. Thanks!16:01
seb128Sweetshark, yw16:06
didrocksSweetshark: hey, I think you should discuss with seif or mhr3 about libroffice zeitgeist datasource integration16:09
Sweetsharkdidrocks: woha! under which stone did we hide that action item ;)16:10
didrocksSweetshark: I think you should just boostrap the discussion16:11
didrocksSweetshark: we will have some datasources for gedit and totem already16:11
Sweetsharkdidrocks: but sure, if it gets stuff going ...16:11
didrocksSweetshark: I think for libro, it's just a question giving guidance to the zg team :)16:11
* Sweetshark notes a todo16:12
didrocksthanks Sweetshark :)16:12
seifdidrocks, i could work on libreoffice too16:15
didrocksI'm sure Sweetshark will help you on that! :)16:16
pittigood night everyone!16:21
pittididrocks: hah, new zeitgeist finally uploaded \o/16:22
seifpitti, u know what rocks about it16:23
didrockspitti: good night pitti! Yeah, I shot it this time so that it's uploaded16:23
didrockspitti: 3 strikes, always the 3rd works :)16:23
seifits faster, less memory and starts up nicely16:23
pittiseif: killall python?16:23
pittiseif: yeah, that's why I pestered didrocks about uploading it yesterday already :)16:23
seifpitti, well we have one python process remaining16:23
pittican't wait to try it16:23
seifbut it depends on Mikkel16:23
didrockspitti: I have the impression that the long delay to close was due to zg as well16:24
didrockspitti: my session shutdown is way faster now16:24
didrocksseb128: FYI ^16:24
pittioh, nice! I never had that, my machine just takes ages to actually power off (or doesn't at all)16:24
didrocks(well, for a week already, as I installed it the day i uploaded it the first time)16:24
pittiaaanyway, really off now, cu tomorrow!16:25
didrockswell, s/I uploaded/I first tried to upload…16:25
didrockspitti: see you :)16:25
seb128didrocks, nice16:25
Sweetsharkseif: lets chat about it tommorrow, gotta go to the gym now ...16:44
=== eeejay is now known as eeejay_is_afk
* rodrigo__ eod's, later all17:20
seb128rodrigo__, 'night, btw can you reply on that proxy bug tomorrow17:21
seb128?17:21
seb128(just mentioning it now while I think about it)17:22
rodrigo__seb128, I already commented iirc17:22
rodrigo__I'm working on a fix in the branch where I'm doing the systemd interfacesa17:22
rodrigo__seb128, but anyway, have to go now, bbl :)17:23
seb128rodrigo__, see you!17:23
didrockssee you rodrigo__17:23
seb128rodrigo__, ok, we can discuss it tomorrow ;-)17:23
=== eeejay_is_afk is now known as eeejay
=== eeejay is now known as eeejay_is_afk
mterryWho do I bug about problems with the wiki?   A page just got deleted when I tried to edit it...  I'd like to see if there's a backup somewhere18:40
seb128mterry, what page was that?18:41
seb128mterry, you should try the is channel I guess (not sure how it's named)18:42
mterryseb128, https://wiki.ubuntu.com/Debian/Usertagging18:44
mterryit's gone!18:44
* kenvandine heads out for a doctor's appointment, be back in a couple hours18:46
seb128mterry, https://wiki.ubuntu.com/Debian/Usertagging?action=info18:46
seb128mterry, https://wiki.ubuntu.com/Debian/Usertagging?action=recall&rev=1318:47
seb128mterry, you can at least do that18:47
=== eeejay_is_afk is now known as eeejay
=== seif is now known as seifstrup
=== seifstrup is now known as kamlotfy
=== kamlotfy is now known as seiflotfy
nookIs it possible that the global menu of Unity allways show the menu items? I dont want to hover the menu to see the items... That makes me crazy :(20:49
mterrynook, fyi you can also hold alt to see them20:53
nookOk thanks, that helps to find the right way with the cursor... But no way to show allways the items :(20:55
seb128ok, precise hates me it seems21:07
seb128hey mterry, is there any chance you could look at the shotwell ftbfs there? ;-)21:07
seb128nook, do you use the menus that much? you can try to uninstall indicator-appmenu if you don't like menus to be exported21:08
nooki like the global menu... i only want to see the items all the time^^21:11
mterryseb128, thought shotwell was building...21:15
seb128mterry, https://launchpad.net/ubuntu/+source/shotwell/0.11.6-0ubuntu121:15
mterryseb128, stop breaking things!21:16
seb128mterry, well I just uploaded a new version that doesn't change anything in the build system, I blame precise moving under my feet!21:16
kenvandineseb128, and we kept telling you the upgrade to precise was smooth :)21:16
seb128mterry, but but but, you told me that you would do desktop work if that was build issues! ;-)21:16
mterry:)21:17
mterryseb128, I'll look at it in a bit21:17
seb128mterry, thanks21:17
seb128well I did a first step to precise today, I set up the sources for it and grabed cdbs and a few other things21:17
seb128tomorrow I might do a partial upgrade of desktop stuff ;-)21:17
seb128I'm always scared of updating to the new xorg or kernel :p21:18
mterryseb128, so cautious!  take a page from pitti's book  :)21:18
kenvandinei used do-release-upgrade21:18
kenvandinenice and smooth :)21:18
seb128those brave men21:18
mterryseb128, higher versions means they're better!  :)21:18
seb128jumping in unstable like that21:18
kenvandinemterry, that is what i keep hearing :)21:19
seb128you can't understand, I come from Debian where dist-upgrading to unstable means something :p21:19
seb128but maybe the world changed during those 10 years21:20
mterry:)21:20
TheMusoI did a fresh install of precise, as I still keep my Oneiric install around just in case something happens to precise. But since installing precise, I haven't had any major dramas.21:20
mterryseb128, remember, this is just dist-upgrading into testing21:20
seb128mterry, not sure that's any better than unstable, testing is where when something breaks it takes a week to get it fixed ;-)21:21
seb128but yeah, gotcha, I'm going to update tomorrow21:21
mterryseb128, don't worry, us +1ers are on the case!21:22
seb128if something breaks you will keep hearing me daily until it's fixed!21:22
seb128you can ask chrisccoulson, I can do "is it fixed yet" poking every hour during a whole UDS (or almost that ;-)21:22
* mterry blocks seb12821:22
chrisccoulsonlol21:23
chrisccoulsonmterry, just remember not to buy seb128 a beer in budapest!21:24
seb128:-(21:24
chrisccoulsonheh21:24
seb128why so much hating!21:24
chrisccoulson:P21:24
seb128chrisccoulson, your twitter description say you are an happy man, not a hater :p21:24
chrisccoulsonlol21:24
chrisccoulsonit's a lie!21:24
seb128yeah, I noticed, I got fooled again it seems!21:25
chrisccoulsonheh :)21:25
seb128one day I will stop trusting what is written on the internet21:25
seb128;-)21:25
seb128chrisccoulson, btw not wanting to bother you with work but I think tb notifications behave weirdly for me21:25
seb128waiting to get some extra emails to say, I usually read the summaries in the bubble and not the title but I think the title mentioned some thousand new emails received when I had like 5 new bug emails21:26
seb128is the title showing the box count rather than the unread count?21:27
chrisccoulsonnot sure about that :)21:27
mterrySo how did the Chuck meme start?21:27
chrisccoulsonask jono :)21:28
chrisccoulsonwow, i was just watching http://www.youtube.com/watch?v=pghA1XUaxxU21:28
chrisccoulsonpretty incredible!21:28
seb128indeed21:30
TheMusoYeah, I heard about that in the news the other day, but thats the first footage I've seen.21:40
chrisccoulsonyeah, i can't believe that the pilot actually walked away from that21:42
chrisccoulsonhe's one lucky dude!21:42
TheMusoHell yeah.21:43
=== eeejay is now known as eeejay_is_afk
=== eeejay_is_afk is now known as eeejay

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