/srv/irclogs.ubuntu.com/2012/01/04/#ubuntu-desktop.txt

pittiGood morning06:51
pittiAmpelbein: seems someone gave it back, g-keyring should be fine now06:51
pittihey jasoncwarner_, happy new year!06:51
jasoncwarner_hey pitti you as well!06:52
BigWhaleGood Morning.06:57
didrocksgood morning06:59
BigWhaleafter ~4 hours of sleep I can't really call it good .. :>06:59
didrocks:/07:00
BigWhalethat's why I get for staying up late and hacking ...07:00
BigWhale:>07:00
pittibonjour didrocks07:02
didrocksguten morgen pitti. How are you?07:05
pittididrocks: cold in full progress. :/07:06
didrockspitti: urgh, take it easy today then!07:06
rickspencer3didrocks, bonjour07:50
didrocksbonjour rickspencer3, comment vas-tu?07:51
rickspencer3ça va bien, et tois?07:51
rickspencer3en fait, j'ai commencé mon script pour «unity_stress»07:52
rickspencer3didrocks, so far, the script changes workspaces every .5 seconds n times07:53
rickspencer3didrocks, what else should I make it do?07:53
didrocksrickspencer3: trigger expose for applications with multiple windows07:54
rickspencer3didrocks, that sounds hard ;)07:54
rickspencer3I'll give it a try though07:54
rickspencer3like, open a bunch of gedit windows and trigger exposé?07:54
didrocksrickspencer3: yeah, and then close expose, and trigger it again. You have a keyboard shortcut in ccsm for that, so rather, using that shortcut I woul say07:55
didrockswould07:56
rickspencer3ok, I'll look into it07:56
rickspencer3obviously minimizing windows as well ;)07:56
didrocksyeah, minimzing/restoring/maximizing07:57
didrockswe can as well use my little soft for quicklist/progress bars07:57
rickspencer3hey desktoppers, who in Ubuntu One do I assign bugs to? is there a team?08:03
pittirickspencer3: ~ubuntuone-hackers AFAIK08:08
rickspencer3thanks pitti08:08
smspillazrickspencer3: how are you writing this script btw ?08:33
smspillazrickspencer3: compiz can expose certain functionality over dbus, might make it easier for you to automate these things08:33
smspillaz(user driven things, like, initiating expo, scale, etc etc etc)08:34
rickspencer3hi smspillaz it's a bash script, if that helps08:40
rickspencer3didrocks, looks like there is a compiz update in the desktop PPA08:47
rickspencer3should I take it?08:47
rickspencer3(for Oneiric)08:47
didrocksrickspencer3: seb128 did it on december, he told me that he didn't see any isssue so far08:50
rickspencer3ok08:50
didrocksrickspencer3: so yeah, please, take it and pull the trigger if things go badly :)08:50
* rickspencer3 clicks Install Updates08:50
* didrocks crosses fingers :)08:51
rickspencer3didrocks, what command in ccsm do I need to use to activate spread by keyboard?08:54
didrocksrickspencer3: uno momento! :)08:54
rickspencer3en espanol?08:54
didrocksrickspencer3: mon espagnole est pire que mon allemand, tu ne veux pas tester :)08:55
rickspencer3didrocks windows-z spreads everything08:56
rickspencer3is that good enough?08:56
rickspencer3or do you need to one app only spread?08:56
didrocksrickspencer3: yeah, was just looking at that, it's the same code, so for stressing, it's good enough :)08:56
rickspencer3ok08:56
rickspencer3thanks man08:56
didrocksyou're really welcome :)08:56
seb128hey09:06
pittibonjour seb128, ca va?09:06
seb128hey pitti, I'm good, my cold is almost over (at least my nose stopped being blocked during nights)09:07
seb128how are you?09:07
pitticold still in full progress, but bearable in the mornings09:07
seb128:-(09:09
seb128pitti, libgda5 got accepted in Debian btw09:12
seb128efficient NEWing nowadays ;-)09:12
pittiseb128: I saw, it was really quick!09:12
pittiseb128: btw, do you have a clue about bug 911622?09:12
ubot2`Launchpad bug 911622 in gdk-pixbuf "[12.04] GdkPixbuf-WARNING **: Cannot open pixbuf loader module file '/usr/lib/i386-linux-gnu/gdk-pixbuf-2.0/2.10.0/loaders.cache': Datei oder Verzeichnis nicht gefunden (file not found)" [Undecided,New] https://launchpad.net/bugs/91162209:12
seb128not off hand, let me have a look09:12
seb128pitti, ok, I know09:21
chrisccoulsongood morning world09:23
seb128hey chrisccoulson, how are you?09:25
chrisccoulsonseb128, yeah, good thanks. how are you?09:26
seb128pitti, hum, no, in fact I don't know, I though it was maybe similar to the gtk3 I fixed yesterday and the postinst call failing when the olddir was empty09:26
seb128but that has been handled in gdk-pixbuf by using find to list the files09:26
seb128chrisccoulson, I'm good thanks09:26
chrisccoulsonwow, it's quite amazing that when you search for "chrome" in google this morning, this page ranks higher than google's browser: https://developer.mozilla.org/en/Chrome_Registration09:31
seb128chrisccoulson, open a bug!09:36
chrisccoulsonseb128, lol09:36
chrisccoulsonseb128, i think it's intentional: http://searchengineland.com/google-chrome-page-will-have-pagerank-reduced-due-to-sponsored-posts-106551 ;)09:36
chrisccoulsongoogle got caught out violating their own rules with a sponsored post campaign for chrome ;)09:37
chrisccoulsonwhich is hardly surprising tbh.....09:37
seb128chrisccoulson, "sponsored post compain"?09:40
chrisccoulsonseb128, paying sites to post bogus content with a link to the chrome website to boost page rankings ;)09:40
seb128...09:41
seb128no comment09:41
chrisccoulsonheh09:41
seb128I switch to bing, at least it lists chrome first when I type "chrome" :p09:41
chrisccoulsonlol09:41
chrisccoulsonbing must be using a similar algorithm for page rankings as chrome then. perhaps someone should contact microsoft to have them decrease the ranking for the chrome website too ;)09:42
chrisccoulson**as google09:42
chrisccoulsond'ohg09:42
seb128;-)09:42
smspillazrickspencer3: yeah, so it should be possible, if we enable the d-bus plugin by default to trigger those things using d-bus-send09:55
rickspencer3smspillaz, I'm finding it rather easy with xdotool to do what I want, in fact09:56
smspillazsure09:57
smspillazxdotool works as well, although using d-bus might make things a little less fragile in case, eg, the keybindings change09:57
pittiseb128: hm, I have an apport fix for ignoring the __memcpy_sse3 and friends stuff10:11
pittiseb128: I'll also ignore __kernel_vsyscall() while I'm at it10:11
seb128pitti, \o/10:11
pittiseb128: the problem is that committing this will break the existing duplicate db10:11
pittias a lot of the current signatures have the old stuff10:12
pittiso we'll temporarily get non-detected duplicates, but dup detection will have a higher quality in the future10:12
pittithat's a bit tricky10:12
seb128well, we will deal with it10:12
pittiseb128: hopefuly the address signatures will take care of it somewhat10:12
pittibut we'll see more of these SystemErrors in teh future10:12
pittiseb128: so there's two options:10:13
pitti- leave the assertions and dupe the bugs manually10:13
pitti- ignore the assertion and force the duplication, at the expense of false positives10:13
pittiI propose we'll roll this out, wait for a few crashes, investigate that duping is really correct in these cases10:13
pittiand once we get convinced that it's right, automatically dupe for a while10:14
pittiseb128: unfortunately it's not easy to manually hack the dupe db to just filter out the ignored stuff, as we are lacking the frames after the existing ones10:14
seb128pitti, you somewhat lost me there, why will we have an increase of those systemerrors?10:15
pittiseb128: we currently get this assertion if the following happens:10:15
pitti- a crash has the same address signature as the existing one10:15
pitti(that's the new way of client-side dupe detection)10:15
pitti- after retracing the fully symbolic trace has a different signature than the existing bug10:16
pittithis assumes that fully symbolic traces are "better"10:16
seb128which is a correct enough assumption ;-)10:16
pittii. e. all equal address sigs should also have an equal symbolic sig10:16
pittiwe got the previous SystemErrors because the symbolic sigs had different __strcpy_sse42 vs. __strcpy_sse2 or what not10:16
pittiwhich broke above assumption10:17
seb128yeah, gotcha10:17
pittinow that we filter them out, we'll break it even harder, until we get crash duplicates with the cleaned up signatures10:17
seb128<pitti> - leave the assertions and dupe the bugs manually10:17
pittiso we could say "if either teh address or the symbolic signature matches, dupe it"10:17
seb128I would try that for a while and see how it goes and review the buggy cases10:17
pittiseb128: ok, that was my proposal10:17
seb128right, I agree with you then ;-)10:18
pittiand once we get sufficiently convinced that it DTRT, we'll switch over to automatic duping10:18
seb128sounds good10:18
pittithis is still in a kind of test drive where I added several asserts that the logic is correct, to avoid false dupes10:18
pittiseb128: ok, so you have been warned :)10:18
seb128thanks a lot for the explanations ;-)10:18
pittiseb128: whenever we get such a crash, we should verify that the retraced and master IDs are really dupes, and dupe them manually, then restart10:18
seb128ok10:19
seb128will dupping manually "resolve" the issue?10:19
pittifor that particular bug, yes10:19
pittioh, hang on10:19
seb128i.e the retracer will know that it matches the signature of the "new" bug which is a duplicate of the old one?10:19
pittiwe also need to kill the "wrong" symbolic signatures from the dupe db10:19
pittiotherwise we'll keep the wrong ones around forever10:20
seb128can you made a small utility which replace a signature by the one from another bug?10:20
seb128like "replace-sig bug1 bug2"10:20
seb128so we can run that when we hit those buggy cases?10:21
pittilet me think about this for a bit first10:21
seb128ok10:21
pittiwe already know which bugs now will have "wrong" sigs10:21
seb128those which have the symbols you started filtering out?10:21
pittilet me think about/check what happens if we just remove them10:21
pittiseb128: right10:21
seb128ok10:21
pittiseb128: I'm creating fake lock files while I'm working on this, FYI10:24
pitti(i. e. stop the retracers)10:24
seb128pitti, noted10:25
pittiseb128: ok, I'll just wipe the existing wrong ones from the DB; they'll just get recreated if there's already a matching address sig10:30
seb128great10:30
pitti$ cp apport_duplicates.db apport_duplicates.db.__sse_sigs10:30
tjaaltonslomo: is there a plan to get gstreamer-vaapi packaged at some point?10:47
slomotjaalton: sure, when someone with time and hardware that supports it does it ;)10:48
tjaaltonslomo: i see the long-term plan is to integrate it with -plugins-bad, but an interim solution would be a separate package until it's merged10:49
tjaaltonor?10:49
tjaaltoni have the hw, might package it then10:49
seb128chrisccoulson, there?10:49
chrisccoulsonseb128, yeah10:49
slomotjaalton: yes10:50
seb128chrisccoulson, can you help on https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/903973 ?10:50
ubot2`Launchpad bug 903973 in gnome-settings-daemon "gnome-settings-daemon crashed with signal 5 in _XReply()" [High,Confirmed]10:50
seb128chrisccoulson, just giving hints on how to debug those10:50
seb128chrisccoulson, would it help if we disable the libxklavier handler during the unstable cycle for a bit to get the real stacktraces from g-s-d?10:50
seb128chrisccoulson, rather than getting stuff libxlavier is catching?10:51
tjaaltonslomo: ok, I might spend some cycles on it then10:54
pittiseb128: retracers back on, let the fun begin :)10:55
chrisccoulsonseb128, it probably wouldn't help in this case, as the call which actually generates the error is the one which is on the stack10:55
seb128pitti, \o/10:56
seb128chrisccoulson, how do you see if libxklavier "hijacked" the call or not? I don't remember what you told me by then about libxklavier and how it was "misleading" the stacktraces10:57
chrisccoulsonseb128, the error handler used by libxklavier exists to swallow some errors that occur normally, so if you disable that then they will just be caught by the GDK handler instead (which will abort)10:58
chrisccoulsonin this case though, the call which generates the error is the one which is in the trace (XGetWindowAttributes)11:00
chrisccoulsoncalled by xklavier :-)11:00
seb128chrisccoulson, I'm not sure to understand why we had issues with libxklavier then, if it just "traps" error it should have stopped issues, not leaded to confusing stacktraces?11:03
seb128chrisccoulson, well anyway this one seems a g-s-d keyboard bug then? would you be interested to have a look to it or should I just assign it to the team?11:03
chrisccoulsoni'll have a quick look to see if i can figure out what's going on11:04
seb128chrisccoulson, thanks11:05
chrisccoulsonok, XGetWindowAttributes throws a BadMatch error if the specified window isn't actually a window11:08
seb128chrisccoulson, do you read the source to see that? ;-)11:13
seb128the manpage states "       XGetWindowAttributes can generate BadDrawable and BadWindow errors."11:13
chrisccoulsonseb128, yeah11:13
chrisccoulsondixLookupWindow returns that error it finds some resource using the specified ID, but it isn't a window11:14
chrisccoulsonthose manpages often don't reflect reality ;)11:15
seb128I see ;-)11:17
GunnarHjseb128: FYI an attempt to bring the discussion with upstream about accountsservice's role for language/locale settings forward:11:36
GunnarHjhttps://bugs.freedesktop.org/42857#c411:36
GunnarHj(pitti and rodrigo_ are on the CC list)11:36
seb128GunnarHj, hi, great11:36
GunnarHjseb128: That's basically all that I can come up with, I think. Not sure how to proceed.11:37
seb128wait for a reply from mathias?11:37
GunnarHjseb128: Yeah, that's the obvious first step, of course. :)11:39
chrisccoulsonseb128, ok, deferred bug 903973 to RAOF ;)12:29
ubot2`Launchpad bug 903973 in gnome-settings-daemon "gnome-settings-daemon crashed with signal 5 in _XReply()" [High,Confirmed] https://launchpad.net/bugs/90397312:29
=== MacSlow is now known as MacSlow|lunch
seb128chrisccoulson, way to go! ;-)12:30
seb128it's always the xorg guys fault at the end!12:31
chrisccoulsonheh12:32
didrockswell, they display, it's their fault :)12:32
chrisccoulsonit seems that i've investigated pretty much the exact same problem at some point in the past12:32
seb128didrocks, exactly!12:32
chrisccoulsonseeing as i reported https://bugs.freedesktop.org/show_bug.cgi?id=23562 ;)12:32
ubot2`Freedesktop bug 23562 in Protocol/Core "GetProperty can return BadMatch error under certain conditions" [Normal,Resolved: fixed]12:32
seb128chrisccoulson, I've done "also affect xorg-server" and assigned that one to RAOF12:33
chrisccoulsonthanks12:33
seb128yw ;-)12:33
seb128thank you for looking at it12:33
seb128g-s-d didn't change in precise and the bug started there12:33
seb128so it might be a new issue somewhere in the xorg stack12:34
nessitahello everyone!12:34
seb128hey nessita!12:34
chrisccoulsonxorg didn't change either, the actual bug has been there all along. however, i suspect that something else on the desktop has changed which just exposes the race now12:34
seb128could be yes12:34
chrisccoulsonsomething creating and destroying a window very quickly ;)12:34
nessitahi seb128!12:34
seb128nessita, show are you?12:34
seb128chrisccoulson, is there any way to find that something? and what is g-s-d tracking,why?12:35
chrisccoulsonseb128, gsd is tracking all windows for the keyboard-layout-per-window stuff12:35
nessitaseb128: great, thanks. You?12:35
chrisccoulson(ie, xklavier adds some properties to each top-level window)12:35
seb128only when that option is activated?12:36
chrisccoulsonseb128, i'm not too sure about that. possibly12:36
seb128ok, that might explain why some people see it and others not12:36
seb128I will try to activate it just to see if I get the bug12:36
seb128nessita, I'm great thanks!12:36
seb128nessita, I had a cold but it's mostly over ;-)12:36
chrisccoulsonslangasek, do you use more than one keyboard layout? (re, bug 903973)12:37
seb128nessita, do you come to budapest next week?12:37
ubot2`Launchpad bug 903973 in gnome-settings-daemon "gnome-settings-daemon crashed with signal 5 in _XReply()" [High,Confirmed] https://launchpad.net/bugs/90397312:37
nessitaseb128: oh, I envy your weather, I guess you have a cold because *is* cold there. We're suffering the heat a lot here. And sadly no, not going to budapest, not this time. I will miss you, though :-/12:38
seb128nessita, :-( I wil miss you as well12:38
seb128well it's not real cold this year, we have a weird weather12:38
seb128cold and dry is good, we have a not-so-cold and rainy weather12:38
chrisccoulsonsame here ;)12:38
chrisccoulsonit's just dull and wet here12:39
nessitaugh12:39
chrisccoulsonsame as the rest of the year ;)12:39
seb128lol12:39
nessitaheh12:39
seb128seems about right for birmingham :p12:39
nessitaso, if anyone feels like it, I could use a sponsorship for https://code.launchpad.net/~nataliabidart/ubuntu/precise/ubuntuone-control-panel/ubuntuone-control-panel-2.99.1/+merge/87406? I checked the patch pilot calendar and is kinda empty for today, but there is no rush :-)12:39
seb128nessita, doing it12:39
seb128nessita, I guess it's empty because dholbach usually fills it but he's not back from holidays yet I think12:40
nessitaseb128: thanks! (link has an extra ? at the end)12:40
nessitaah, daniel, daniel :-P12:40
seb128nessita, launchpad seems smart enough to not break on the trailing "?" ;-)12:40
seb128nessita, uploaded12:44
nessitaseb128: awesome, thanks!12:45
seb128nessita, yw12:45
nessitayou rock12:45
nessita:-)12:45
seb128nessita, you as well ;-)12:45
seb128pitti, bug #911734, is that an apport bug?12:48
ubot2`Launchpad bug 911734 in gnome-keyring "package libgcr-3-common 3.2.2-1ubuntu2 failed to install/upgrade: trying to overwrite '/usr/share/gcr-3/ui/gcr-import-dialog.ui', which is also in package libgcr-3-1 3.2.2-0ubuntu1" [Undecided,New] https://launchpad.net/bugs/91173412:48
pittihm, I thought I fixed that12:49
seb128pitti, you fixed the gnome-keyring bug12:49
seb128but "Package: libgcr-3-common 3.2.2-1ubuntu2"12:49
seb128though " Unpacking libgcr-3-common (from .../libgcr-3-common_3.2.2-1ubuntu1_all.deb) ..."12:49
seb128 12:49
seb128pitti, i.e it picks the newest version and wrongly assume it's a regression12:49
seb128but the log shows it's 1ubuntu1 which hit the bug12:49
seb128not sure what get the version wrong12:50
pittihmm12:50
seb128i.e that "the version changed between the time the bug got collected and reported"?12:50
pittiseb128: that's one plausible explanation12:51
seb128pitti, I can dup it manually but it seems a buggy retracer case, do you want it reassigned to apport or something?12:51
pittiI'm not 100% sure which information the apt hook already adds12:51
pittiseb128: all information is in the bug, so please go ahead and dup, yes12:51
pittiseb128: right, or we just reassign it to apport, that's even better12:52
seb128doing that, it's easier that to open a new bug12:52
seb128and it has the infos needed from a buggy cases12:52
seb128pitti, done and title updated, bug #91173412:54
ubot2`Launchpad bug 911734 in apport "installation issues report sometimes the wrong package version" [Undecided,New] https://launchpad.net/bugs/91173412:54
pittiseb128: thanks12:54
seb128yw12:54
GunnarHjpitti: Hi Martin, hope you are fighting your cold successfully.13:00
pittihey GunnarHj, happy new year!13:00
GunnarHjpitti: Happy new year to you as well.13:00
GunnarHjpitti: Do you possibly have time to take a shot at the a-s MP (bug 866062) and the related l-s MP? I can be available here for instant clarifications...13:00
ubot2`Launchpad bug 866062 in accountsservice "SetLanguage(): Write ~/.pam_environment instead of ~/.profile" [Medium,In progress] https://launchpad.net/bugs/86606213:00
pittiGunnarHj: I saw your post to the upstream bug13:02
GunnarHjpitti: Yeah, that's the forwarding aspect of it. But I really think that we should get the MPs into the archive before that discussion is completed.13:03
pittiit's really tricky to review, as it contains several things in one huge MP: the new API, the migration, the (still TBD) changed meaning of LANG, and the .bashrc -> .pam_environment bits13:04
pittidoing just the latter should be rather simple really, and everything else shoudl really be done upstream13:05
pittithis is now so utterly complex and error prone that it'll get harder and harder to do even more migrations in the future if upstream does something different or sticks to the current approach13:06
GunnarHjpitti: Well, I realize that it may appear complex at first sight, but there is a rather clear logic if you take a closer look.13:08
GunnarHjpitti: I did try to adjust the code in accordance with your first review.13:09
pittiGunnarHj: it's still on my list, but again, this is not something that I can review in an hour or so13:09
pittiand I'd like to pick it apart into above four parts, so that it's easier to review and understand13:10
GunnarHjpitti: I can try to split it into more revisions. But please note that the migration part covers both ~/.profile -> ~/.pam_environment and the changed LANG (format -> language) ...13:12
pittiwell, "mentally split" is fine, too13:13
GunnarHjpitti: What do you mean by that? Expanding the comments?13:13
pittibut anyway, this requires pretty much a full day of review; in December I was on a different team, and now I have some catching up to do, that's why it took so long, sorry13:13
pittiGunnarHj: no, I mean it'll take some time to understand the current behaviour, the new one, reviewing the huge changeset, finding error cases, etc.13:14
pittiand it's again pretty spaghetti-ish13:14
pittiC code, shell code, more shell scripts, zenity, the $(ls /tmp/) ... rm loop (which might have a security problem), and so on13:14
pittiand I'm not convinced that we should land the LANG changes now13:15
pittiit's a rather large change which also impacts KDE, XFCE, etc., multiple window managers, gnome, and what not13:16
GunnarHjpitti: Hmm.. as regards LANG it's my impression that the way Ubuntu uses it is the 'odd' way. GNOME does it the other way for sure.13:18
pittiGNOME only handles LANG ATM, it doesn't have a concept of a language/region split so far?13:18
pittimeh, our g-s-d package is a mess .. 20ish patches without any patch header13:19
GunnarHjpitti: I think they have: Language -> LANG via /etc/gdm/Xsession and formats -> LC_* via gsettings g-s-d13:19
pittiah, indeed13:21
GunnarHjpitti: So that's one good reason to do the LANG change, considering the move to g-c-c. It also will take care of some cases of incorrect display language.13:22
=== MacSlow|lunch is now known as MacSlow
GunnarHjpitti: Please let me know if there is anything I can do to clarify the a-s MP. Some level of 'mental split' is hopefully in the changelog...13:25
pittiGunnarHj: how do you handle existing settings in .bashrc now?13:25
GunnarHjpitti: I didn't touch .bashrc. Reason: Since possible locale settings there has been made manuelly, it's reasonable to assume that the user has chose to ignore the UI and do it manually.13:27
pittiGunnarHj: but previous l-s wrote its settings there13:27
pittiso we need a migration path13:27
pittiotherwise you could never change the language again any more13:27
GunnarHjpitti: No, you mean ~/.profile, don't you?13:27
pittiGunnarHj: ah, yes I do13:28
GunnarHjpitti: And that's certainly taken care of.13:28
pittiah, I see the sed call there13:29
pittiI think we should only do that if ~/.pam_environment doesn't exist yet13:29
pittior rather, doesn't have the locale settings13:29
pittii. e. once, not every time13:29
seb128pitti, g-s-d> we have 16 patches and they are mostly trivial ones?13:30
GunnarHjpitti: Existence of ~/.pam_environment is exactly the criteria I'm using in the MP.13:30
pittiseb128: yes, but almost none of them have a description or bug link13:30
seb128right...13:30
seb128pitti, well most are ubuntu specific13:30
seb128pitti, I will add descriptions next time I do an upload for it13:30
GunnarHjpitti: So migration happens only once, which of course is essential.13:31
pittiseb128: right, I was more concerned about the bug fixes; would really be nice to get them upstream, etc.13:33
seb128pitti, I don't think we have any fix coming from Ubuntu not upstreamed13:33
GunnarHjpitti: Possibly the latter isn't as clear as it could be, since the test for existence of ~/.pam_environment happens in the beginning of user_migration_from_profile().13:34
pittiGunnarHj: *nod*, sounds fine (we don't want to run the shell script at all if it exists)13:35
GunnarHjpitti: Right.13:35
pittiseb128: do you happen to know how to get the g_debug() output from gnome-settings-daemon? --debug doesn't13:36
pittiIIRC --debug used to work, but not any more13:37
seb128pitti, it doesn't? did you try to kill g-s-d and run a new one with --no-daemon --debug?.13:37
pittiyes, I killed it13:37
pitti--no-daemon doesn't exist13:37
pittiit does run in the foreground13:37
seb128let me check13:37
pittiit just doesn't output anything except warnings13:37
pittibug not the g_debug13:37
seb128hum13:38
pittiseb128: don't worry, I just asked in case you knew the trick by heart13:38
seb128well I usually use --debug which worked for me :p13:38
seb128pitti, oh, I know13:40
seb128looking...13:41
pittigsd_log_default_handler() has debug==113:41
pittiso it seems g_log_default_handler eats it13:42
seb128pitti, yeah, since glib 2.31 there is a G_SOMETHING to set13:43
seb128I'm trying to find which one...13:44
BigWhaleHow would I suppress delete-event on Gtk.Window? I don't want my window deleted just hidden when I click X13:45
seb128pitti, try G_MESSAGES_DEBUG=all13:46
pittiseb128: aah, that worked; many thanks!13:46
seb128pitti,  G_MESSAGES_DEBUG=all g-s-d --debug13:46
seb128pitti, yw13:46
seb128took me a while to find it back!13:47
pittiit's helpfully not mentioned in the g_debug docs13:47
seb128indeed13:48
seb128pitti, https://bugzilla.gnome.org/show_bug.cgi?id=66192613:48
ubot2`Gnome bug 661926 in general "Improve the default logging setup in GLib" [Normal,Resolved: fixed]13:48
seb128pitti, it's documented in http://developer.gnome.org/glib/2.31/glib-Message-Logging.html#g-log-default-handler13:49
seb128but yeah, the api for g_debug etc could have a note about it13:50
seb128chrisccoulson, do you think you could send to GNOME your gsd login speed patches?13:51
seb128seems you didn't do it?13:51
BigWhaleam I the only one having trouble with hide_on_delete in python? *bursts into tears*13:54
pittiseb128: ah nice, I can do G_MESSAGES_DEBUG=print-notifications-plugin13:55
seb128right13:55
seb128that was the idea, filter on domains13:55
seb128BigWhale, just put a callback on the signal and return TRUE in the handler?14:00
BigWhaleseb128, yeah I just figured it out14:00
BigWhaleI used return self.hide_on_quit()14:01
dobeykenvandine: do you even have any "readiness tests" for u1?14:16
kenvandinedobey, no... i would need you guys to help define that14:17
kenvandineit would be something like confirming upstream tests pass, etc14:17
kenvandineand some integration stuff, perhaps14:17
dobey:-/14:17
dobeykenvandine: so i had a sort of discussion with jason last night about this; and it seems upload privs for anyone not on platform team, for canonical packages, are basically worthless now :(14:22
kenvandine:(14:23
dobeyso as you can guess, i am quite unhappy about that14:24
pittikenvandine: why's that?14:25
pittias long as there are AC, it doesn't matter much who runs them?14:25
kenvandinepitti, it was really a question of if that person has to be on the platform team or not14:25
kenvandineu1 has their own uploaders14:25
kenvandineso they don't want to block on having one of us do it14:26
kenvandinethe whole reason they wanted upload rights14:26
dobeyand we have our own qa team even14:27
kenvandinepitti, so u1 had assigned distro acceptance to one of their folks14:28
pittihm, can we (i. e. you) really do a better job at QA than U1's own QA?14:28
dobeywell, i think it just wasn't clear what was supposed to go in that column14:28
kenvandinepitti, no... i can't14:28
kenvandineespecially now that i have probably nearly 30 packages on my plate14:28
dobeypitti: so what do you think?14:33
pittidobey: I think it worked quite fine the way it was; I wouldn't change it from my POV14:35
kenvandinedobey, i'll talk to jason about it14:35
pittii. e. U1 doing their own QA (lots of tests, etc.) and uploading14:35
kenvandinei really think you guys are best to fill that role14:35
dobeyok, thanks14:36
pittidobey: ^ that is, we sohuld of course help you with packaging, and other process-y bits (MIRs, and what not)14:36
pittibut I don't see that kenvandine now handling all uploads and AC checks would improve anything14:36
dobeypitti: yes, there are obviously certain cases where we will have to bug DMB/platform/whatever14:36
pittiyes, absolutely14:36
dobeybut i want to minimize that as much as possible14:36
kenvandinedobey, of course we are here to help :)14:36
dobeyto let you guys concentrate on the other 3000 packages in ubuntu :)14:36
seb128speaking of which, do you plan to upload your fix for the libsyncdaemon segfault? ;-)14:37
dobeyseb128: yes; which is why i stayed around late last night to bug jason :)14:37
seb128great ;-)14:37
kenvandinedobey, lets get that uploaded!14:38
dobeykenvandine: yep, am about to14:38
kenvandineawesome14:39
dobeyjust have to edit debian/changelog and upload14:39
kenvandinedid you propose a merge?14:40
dobeyno14:40
kenvandineok14:40
dobeyi did bzr merge-upstream last night, then bugged jason, then gtfo the computer :)14:41
* kenvandine just discovered the GI_TYPELIB_PATH env variable, that was way to hard to find14:50
pittikenvandine: ah, I could have told you14:51
kenvandinepitti, yeah... i looked for you to ask last night :)14:51
kenvandinepitti, i've added python unit tests to gwibber now using GI14:52
kenvandineso we have test parity for vala/c and python GI14:52
kenvandinewe'll know when the GIR breaks :)14:52
kenvandinepitti, can GI_TYPELIB_PATH handle multiple paths?14:53
kenvandineGI_TYPELIB_PATH=../../libgwibber:../../libgwibber-gtk14:54
kenvandinelike that?14:54
pittiyes, it can14:54
kenvandinewoot!14:55
kenvandinethx!14:55
pittikenvandine: note that you also might need to set LD_LIBRARY_PATH accordingly to use the local .so files14:55
kenvandinei did14:55
dobeyseb128, kenvandine: ok, as soon as pbuilder finishes happily, i'll dput the new package14:58
kenvandinedobey, thx14:58
seb128dobey, thanks14:58
GunnarHjpitti: Did you leave a-s for today? Just wondering if I should keep standing by. (I'm happy to do so.)15:09
pittiGunnarHj: probably not, sorry (working on something else, and keep getting pinged)15:11
pittiI'll probably stop in an hour or two, silly cold15:11
GunnarHjpitti: Ok. Would it help if we scheduled a time to process it?15:12
pittiTBH I'd just try to squeeze it in in the next days and answer in the MP15:12
pittiI got my ubuntu box down to 4 now15:12
pittiof which 2 are your MPs :)15:12
dobeyseb128, kenvandine: uploaded :)15:13
seb128\o/15:13
GunnarHjpitti: Ok, that sounds great. :)  Hope your cold goes away soon.15:13
kenvandine:-D15:13
dobeykenvandine: are you incredibly busy? i have a bunch of other packages which i have to make merge proposals for (since i don't have upload privs for them yet). but they are all trivial no-impact changes (new release only, without any source changes)15:15
kenvandinesure15:15
kenvandinedobey, create merge proposals and request a review from me15:15
dobeykenvandine: ok; thanks15:15
seb128pitti, g-s-d> did you start on login speed? ;-)15:21
pittiseb128: yes; now that I'm back to development I desperately need to start on my WIs15:21
seb128\o/15:21
seb128I should look at my WIs as well15:21
seb128I've been dealing with desktop updates and merges mostly so far15:22
pittiyeah, similar here: there's always a couple of urgent issues which get in the way15:22
seb128pitti, btw you put the bar too high on bugs closing count compared to the rest of the team, you need to let people a change so they try to catch you ;-)15:23
pittihehe15:23
seb128change->chance15:23
pittihttp://reports.qa.ubuntu.com/reports/bug-fixing/precise-fixes-report.html15:23
pittioh, nice15:23
dobeyyeah, i need to do my one WI15:24
seb128dobey, "get a working music store back"? ;-)15:24
pittiseb128: as soon as unity gets uploads again, didrocks will skyrocket to the top again anyway15:24
seb128pitti, indeed15:24
didrocksit's a so easy win :-)15:24
dobeyseb128: no, the only WI i got at UDS was to write a script to analyze some bugs filed against Ubuntu, but without a package15:25
didrocksmeanwhile, /me simulates oneconf server error for the test suite :)15:25
dobeythe music store is something i also have to do, but i don't think there are any work items for it15:26
seb128pitti, is that ok if I NEW gtkspell3 directly to main? that's basically a new version of gtkspell renamed for gtk315:28
pittiseb128: sure15:28
seb128thanks15:28
chrisccoulsonwhy-oh-why https://launchpadlibrarian.net/89040628/buildlog_ubuntu-precise-i386.firefox-trunk_12.0~a1~hg20120103r83671-0ubuntu1~umd1_FAILEDTOBUILD.txt.gz :-(15:31
chrisccoulsonhanging tests! grrrrrrr15:31
seb128chrisccoulson, yeah, those are annoying15:33
seb128pitti, gnome-keyring hanged in the testsuit on i386 and amd64 yesterday that's why your breaks fixes got duplicates for a while, a retry worked though15:34
seb128just for info15:34
pittiseb128: right, saw from backscroll; thanks15:34
seb128kenvandine, "    - Renamed gtkspell binary libgtkspell-common" do you remember why you did that?15:34
kenvandineyes15:35
kenvandinebecause of the dual build for gtk2 and gtk315:35
seb128kenvandine, hum? how is the name revelant? did you move stuff in there?15:36
seb128that binary is not even installed on my box15:36
kenvandinehumm15:36
seb128it's empty15:36
seb128I guess it only has translations that got stripped at build time15:36
seb128kenvandine, do you mind if I revert the name back to the old one? ;-)15:37
kenvandinei seem to recall someone rejecting it from NEW because the common files ended up in gtkspell and someone said the binary would be better named libgtkspell-common15:37
kenvandinesomething like that15:37
kenvandinehold on, let me look at it and try to refresh my memory15:37
seb128kenvandine, Debian packaged upstream vcs as gtkspell315:37
seb128kenvandine, which I just synced15:37
seb128kenvandine, so gtkspell goes back to be mono build of gtk2, I basically want to revert to what it was before your changes15:38
seb128kenvandine, sounds good?15:38
seb128kenvandine, then we just need to make libgwibber build with the new lib15:38
seb128kenvandine, I can look at that if you want15:38
kenvandineoh cool15:39
kenvandinedid upstream release the gtk3 version?15:39
kenvandinethey merged my patch, but tweaked it to make two separate upstream sources15:39
kenvandinewhich seemed silly to me15:39
kenvandinebut last i checked they had a dev snapshot and no release15:40
seb128kenvandine, they didn't release, debian did a vcs snapshot15:40
seb128but that works as well15:40
kenvandinethen revert what i did15:40
kenvandinethat is fine15:40
kenvandinei'll make gwibber work with it15:40
seb128kenvandine, do you want me to look at rebuilding libgwibber with the new lib?15:41
kenvandinenah15:42
seb128ok15:42
kenvandinei'll do that15:42
seb128thanks15:42
kenvandinejust get it uploaded and then i'll make it build15:42
kenvandinei looked at what upstream did there already, it'll be trivial15:42
seb128kenvandine, the lib is named libgtkspell-3-0, yours was libgtkspell3-015:43
kenvandineyeah15:43
kenvandineand they renamed the .pc file15:43
Sweetsharkpitti: so, I have a basic 3.5.0beta2 package which is still very much incomplete wrt ubuntu patches, but does compile and test on amd64. Id love to release that one to the ppa for oneiric to get some early testing, but I guess it will still break on some platforms like ppc/armel. any hint on how to proceed.15:57
Sweetshark?15:57
slangasekchrisccoulson: 903973> I don't *usually* use more than one keyboard layout, but may have had one configured at the time16:04
seb128slangasek, the question is: run gnome-control-center, click on layout, what option is selected on the right?16:09
slangasekseb128: currently, I only have one layout enabled; English (Dvorak alternative international)16:12
slangasekoh, you said on the right16:12
seb128yes ;-)16:12
slangasek"Use the same layout for all windows"?16:12
seb128ok16:12
slangasekbut - I was playing with my keyboard layout in that timeframe16:12
seb128so not the "by win" one16:12
slangasekno16:12
seb128thanks16:12
seb128that was the question ;-)16:12
slangasekbut that could have been selected briefly at the time16:12
seb128we were trying to figure if that could be due to that option16:13
seb128I will turn it on and see if it starts bugging here16:13
seb128ups16:13
pittiSweetshark: PPAs don't build ppc/armel anyway, so that seems fine?16:45
ogra_pitti, Sweetshark, the canonical-arm-dev PPA has armel enabled, i can get one of you access in case you want to testbuild for arm at some point16:48
pittiwe can also test-build on the porter machines16:48
ogra_ah, indeed16:48
pittigood night everyone! need to rest a bit, silly cold16:50
seb128'night pitti16:51
dobeyugh17:20
=== asac_ is now known as asac
dobeykenvandine: looks like https://code.launchpad.net/~dobey/ubuntu/precise/ubuntuone-dev-tools/release-2-99-0/+merge/86756 never got uploaded :(17:21
dobeykenvandine: should i get you to do that one, or stick my new release in that same branch?17:21
kenvandinenope17:21
kenvandinejust stick another new release and we'll do one upload17:22
dobeyok17:22
didrocksRan 46 tests in 474.079s17:27
didrocks\o/17:27
seb128didrocks, well done! ;-)17:28
didrocksok, I think that all OneConf tests are ok for now. Just 2 bugs to fix (and add tests) and I'll be ok at the rally :)17:29
seb128great ;-)17:30
seb128keep something to do for the trip! :p17:30
didrockswell, I still have a LWN backlog :)17:33
didrocksbut I needed some quiet time to be effective on the remaining ones, and at home was a great opportunity to finish this up :)17:34
seiflotfydidrocks: any idea how i can fix17:34
seiflotfy/bin/sed: can't read /usr/lib/i386-linux-gnu/libpangocairo-1.0.la: No such file or directory17:34
seiflotfylibtool: link: `/usr/lib/i386-linux-gnu/libpangocairo-1.0.la' is not a valid libtool archive17:34
didrocksseiflotfy: you probably have some .la file in /usr/lib/ and /usr/lib/i386-linux-gnu/. You should look at which and remove them17:35
seiflotfyyeah but i have a lot of those17:35
seiflotfywhich one is it i have to find17:36
seiflotfydidrocks: ?17:38
didrocksseiflotfy: the one which have libpangocairo-1.0 in them17:39
seiflotfy/usr/lib/i386-linux-gnu/libpangocairo-1.0.so17:39
seiflotfy/usr/lib/i386-linux-gnu/libpangocairo-1.0.so.017:39
seiflotfy/usr/lib/i386-linux-gnu/libpangocairo-1.0.a17:39
seiflotfy/usr/lib/i386-linux-gnu/libpangocairo-1.0.so.0.2905.017:39
seiflotfy?17:39
seiflotfythose?17:39
seiflotfyi have no .la files17:40
seiflotfywhich package provies them17:40
seiflotfyok seems i have libpango installed17:41
seiflotfybut no packages for it17:41
seiflotfyas in the package is installed but no files from it17:41
seb128?17:41
seb128seiflotfy, grep libpangocairo-1.0.la *.la17:42
seb128seiflotfy, in /usr/lib and /usr/lib/i386-linux-gnu17:42
seb128and delete the ones listing it17:42
seiflotfyseb128: 100% there is nothign17:43
seiflotfygrep libpangocairo-1.0.la *.la /usr/lib/i386-linux-gnu/17:43
seiflotfy?17:43
seb128grep libpangocairo-1.0.la /usr/lib/*.la /usr/lib/i386-linux-gnu/*.la17:44
seiflotfyseif@Wumbo:~$ grep libpangocairo-1.0.la /usr/lib/*.la /usr/lib/i386-linux-gnu/*.la17:45
seiflotfyseif@Wumbo:~$17:45
seiflotfyseb128: i told you nothing17:45
seb128grep libpangocairo-1.0.la /usr/local -r17:45
seb128seiflotfy, well it's likely coming from a local install17:45
seb128check your /usr/local or /opt or jhbuild dir or whatever you use17:45
seb128you have a .la somewhere on the disk listing libpangocairo-1.0.la17:46
seb12899% of the time it's a local install, i.e /usr/local or /opt, people who use make install17:46
seiflotfyseb128: ok17:47
seiflotfyseb128: found a lot though17:48
seb128seiflotfy, just rm *.la17:48
seb128they are not useful on linux, we are getting ride of them17:48
seiflotfyon ubuntu or overall linux17:49
seb128debian and ubuntu17:49
seb128but fedora doesn't install those for a long time I think17:49
seb128dunno about other distros17:49
dobeysome are (or at least, used to be) useful17:50
dobeybut maybe glade/gtkbuilder is better about that now17:50
seb128dobey, why would glade or gtkbuilder need those?17:51
seiflotfyseb128: i am freaking out sorry17:51
seiflotfybut17:51
seiflotfyi removed the .la files in my local installs17:51
seiflotfyyet17:51
seiflotfy/bin/sed: can't read /usr/lib/i386-linux-gnu/libpangocairo-1.0.la: No such file or directory17:51
seiflotfylibtool: link: `/usr/lib/i386-linux-gnu/libpangocairo-1.0.la' is not a valid libtool archive17:51
seb128seiflotfy, find / . -name *.la17:52
seb128you probably forgot somes somewhere17:52
dobeyseb128: glade and gtkbuilder themselves don't really; but back in the day we had problems with evolution, evolution-exchange, and other plugins, where we needed the .la files to pull the -rpath from automatically when building, so that "custom widgets" in glade stuff would work correctly17:52
seb128dobey, ok, dunno if that's still an issue or not but I bet seiflotfy doesn't need .la for what he tries to build ;-)17:53
dobeyhe probably doesn't17:54
dobeyi was just providing counterpoint for the "not useful on linux" argument, which isn't entirely true17:54
dobeythey do make linking slower, but to say they are entirely useless is a bit of an overstatement :)17:55
seb128well they mostly are, and I think they are not compatible with multiarch17:56
seb128which is why we are getting ride of them ;-)17:56
seiflotfyseb128: so i can remove ALL .la files from /usr/lib?17:57
seb128yes, but that will not help since you grepped those and they don't refer to libpangocairo17:57
dobeyyeah, the multiarch thing does introduce new problems :)17:57
seb128you need to find which .la list it17:58
dobeyspeaking of multiarch17:58
dobeyis there anything special i need to do for our u1 stuff, to make it multiarch?17:58
dobeythe python stuff is fine obviously, but libs and such not so much i guess17:59
seb128dobey, http://wiki.debian.org/Multiarch/Implementation17:59
didrockstime for sport here, have a good evening everyone!18:00
seb128didrocks, 'night18:00
seiflotfyseb128: what do i do once done?18:01
seb128seiflotfy, ?18:02
seb128just find the .la listing libpangocairo and delete it18:02
seiflotfyi removed all .la files18:02
seb128then run make again18:02
didrocksseb128: thanks, you too!18:04
seiflotfyseb128: does this look right http://imgur.com/W8YbN18:05
seiflotfybecause i have it installed but its not listing any installed packages18:05
seb128?18:06
seb128that seems a bug18:06
seb128try to run "dpkg -L libpango1.0-0"18:07
seb128on a command line18:07
seb128btw the version you are using is not an official Ubuntu one, dunno where you got it18:07
seiflotfyseb128: http://pastebin.com/KWMwdtbQ18:10
seb128seiflotfy, that's the content of the package18:10
seb128seems normal18:10
seb128what are you looking for?18:10
seiflotfyi am now trying to build clutter-gtk18:11
seiflotfy/home/seif/Projects/dawati/source/clutter-gtk/examples/gtk-clutter-viewport.c:41: undefined reference to `g_thread_init'18:11
seiflotfycollect2: ld returned 1 exit status18:11
seiflotfymake[2]: *** [gtk-clutter-viewport] Error 118:11
seb128ok, better18:12
seiflotfyseb128: what seems to be the problem here though?18:12
seb128seiflotfy, drop G_DISABLE_DEPRECATED from your build flag18:12
seb128or fix clutter-gtk to not use deprecated apis18:12
ricotzseiflotfy, what clutter-gtk version is this?18:13
seiflotfyricotz: trunk i guess18:14
seiflotfysorry18:14
seiflotfy0.1218:14
ricotzseiflotfy, i see, 1.1.2 doesnt have a problem18:15
ricotz0.12?18:15
seiflotfy* clutter-gtk-0.1218:15
seiflotfyricotz: ^18:15
ricotzoh, this looks pretty old then18:16
ricotzseiflotfy, check if you are really on the master branch18:16
seiflotfyi am building it for dawati18:17
dobeykenvandine: updated https://code.launchpad.net/~dobey/ubuntu/precise/ubuntuone-dev-tools/release-2-99-0/+merge/86756 though, not entirely sure if i did the changelog right18:18
ricotzseiflotfy, sorry, i dont know dawati, but the clutter-gtk source seems pretty outdated then18:18
ricotzseb128, hi, do you mind sponsoring clutter-gtk 1.1.2?18:19
seb128ricotz, hey, I can do taht18:19
seb128where is it?18:19
seb128mterry, \o/18:19
mterryseb128, what's up?18:19
seb128mterry, I had bug #904140 on my todolist for today, I can drop it thanks to you ;-)18:20
ubot2`Launchpad bug 904140 in evince "evince crashed with SIGSEGV in child_setup()" [Medium,Confirmed] https://launchpad.net/bugs/90414018:20
mterryseb128, heh  :)  yw18:20
ricotzseb128, one se18:20
ricotzc18:20
ricotzseb128, https://launchpad.net/~ricotz/+archive/testing/+build/298584318:21
seb128ricotz, thanks18:21
ricotzseb128, i will create a new source package18:21
seb128ricotz, why?18:22
seb128or for what?18:22
seb128is that still a clutter-gtk topic or did you change topic? ;-)18:22
ricotzfor a proper version18:22
ricotzseb128, or you can grab it and repack it18:22
seb128ricotz, oh, I can download that and strip the ~... from the changelog18:22
ricotzseb128, ok18:22
seb128ricotz, I need to run debuild to build the .changes anyway18:23
seb128that's not on the ppa18:23
ricotzseb128, i would have done that too ;)18:23
ricotzseb128, sorry, i totally missed the fact the cogl/clutter isnt 1.9.2 :\18:34
seb128ricotz, lol, I was going to ask18:35
seb128should we update those?18:36
seb128you recommended to not do it yet before the holidays18:36
ricotzseb128, :\, i was hoping to see some 1.9.4 tarballs18:36
kenvandinedobey, i'll take care of it18:38
dobeykenvandine: great, thanks; i've also got the other packages proposed with you as reviewer now18:40
kenvandineok, i should get to them soonish :)18:41
dobeycool. mostly not a huge rush, since there's no changes aside from version bumps. but since ubuntuone-dev-tools didn't get uploaded for the last release, it's a higher priority than the rest so please do it first :)18:42
micahgcyphermox: that evo change looks risky unless the only nss lib that evo uses are in the subdir18:50
cyphermoxmicahg: I checked18:57
* micahg wonders if there's a more proper way to access the cert store than evo is doing18:59
dobeymicahg: would be nice if everything just used the keyring for that19:06
chrisccoulsonhmmm, beer-o-clock19:15
BigWhaleThe lack of code examples, tutorials and sometimes utterly useless help on stackoverflow is somewhat frustrating time to time ... I wish Canonical could put some effort into this.19:18
dobeyBigWhale: on stackoverflow, or askubuntu?19:42
BigWhalefirst19:43
BigWhaledobey, I didn't find anything on askubuntu19:45
dobeywhat sort of questions? and why should canonical spend time going through it to provide such aexamples and tutorials?19:45
dobeyBigWhale: i don't even know hwat you're looking for. but i don't think either of those is probably the right place to have documentation. docs belong in a more organized place19:46
BigWhaledobey, Well I think it would be great if Canonical could put some effort into basic guidelines, tutorials, documentation and code examples for developers.19:48
dobeyi think we are19:49
BigWhaleor perhaps just organize things a little19:49
BigWhaleMy "rant" is more on a global scale ... Not targeted at anyone in particular. I think it would be great to have sort of one-stop place for all the developers19:51
BigWhaleI am displeased with documentation in general with open source ... :>19:52
dobeylike http://developer.ubuntu.com/ ?19:52
BigWhaleok... wow...19:53
BigWhale*cue in the awkward moment music*19:53
BigWhalebut! :)19:54
BigWhalethis site didn't come up on any searches in google... :/19:54
dobeyBigWhale: it's the 3rd result for me for "ubuntu developer documentation"19:55
dobeyand i imagine the first two results might have links to it as well19:56
BigWhalemy searches were a bit specific19:56
dobeyexample?19:57
dobeysearching for things about specific libraries/languages/etc may very well not get you there19:57
BigWhaleI think that developer.ubuntu.com is quite awesome. Now that I clicked on it a little.19:58
BigWhaleToo bad I didn't notice it sooner.19:59
seb128chrisccoulson, "beer-o-clock"  ... training for next week? ;-)19:59
BigWhaledobey, well I was (still am, to be precise) looking for simple working example on how to use cairo to draw in gtk3 windows.20:00
BigWhaleand how to make a gtk3 window transparent :)20:00
dobeyin vala? python? c?20:01
BigWhalepython preferably20:01
dobeyBigWhale: for the transparency, window.set_opacity()20:02
dobeyBigWhale: for cairo, are you trying to draw on a GtkWindow, or on some other widget?20:03
BigWhaledobey, don't get me wrong, I wasn't trying to criticize or rant, I was merely 'thinking out loud' how nice it would be to have all the developer stuff in one place.20:04
BigWhaleI am happy as a lamb to see developer.ubuntu.com20:04
dobeywell, for ubuntu, i think developer.ubuntu.com is supposed to be that place20:04
BigWhaleset_opacity? now you're just mocking me... :>20:04
dobeyi don't know how extensive the docs are there, though20:05
dobeyfor vala, i mostly just use valadoc.org as a reference20:05
BigWhaleset+opacity actually works ...20:05
BigWhaleexcuse me ... I'll go and hide now.20:06
dobeywell, it should work, unless you're under a wm that isn't doing compositing20:06
BigWhalenow I just need to figure out on how to make a window transparent and some text in it, not so transparent. :)20:07
BigWhaleI'll just stack couple of windows on top of each other or something like that ..20:08
dobeybut i've been dooing gtk/gnome hacking for like 15 years, so i guess i just expect there to not be docs and know how to go about finding stuff i need more easily :)20:08
dobeyi'm not sure what you're trying to do exactly, but athat sounds like it will probably be a bit more work to do :)20:09
BigWhaledobey, I want a countdown in 'splash screen' that is a transparent20:10
BigWhaleand for cairo, I was trying to draw on GtkWindow yes20:11
BigWhaleand I got to the point where I connected draw signal20:11
dobeyi think set_opacity() sets a wm hint on the window, which makes the wm apply transparency20:12
BigWhalebut nothing was actually drawn on the windows then :)20:12
dobeywhich just does it on everything20:12
BigWhalethen I'll probably have to do it with cairo20:12
dobeyi think you need to embed a drawable in the window, and do everything on it20:12
chrisccoulsonseb128, yeah, i'm making the most of it before i give up drinking for a while :)20:13
dobeyBigWhale: you need a GtkDrawingArea20:14
BigWhaledobey, as I understand this: http://developer.gnome.org/gtk3/3.0/GtkWidget.html#GtkWidget-draw20:14
BigWhaleI get a cairo surface here and I can draw on it as much as I like.20:15
dobeyyou misunderstood :)20:15
BigWhaleok20:15
dobeythat's not something you should connect() to on a widget, and then attempt to draw on the widget in your callback20:16
dobeyBigWhale: in Python, you can class MyWidget(Gtk.DrawingArea): and override the draw() method though, then just embed MyWidget() in the window20:17
BigWhalehmm I was looking at some gtk2 example that connected to expose-event signal and I found in some documentation that draw signal is now used instead20:17
dobeyyes, draw replaces expose-event, but you really should only use them internally20:18
dobeythey're more like virtual class methods, than signals that an app developer should connect to20:18
dobeyBigWhale: http://developer.gnome.org/gtk3/stable/GtkDrawingArea.html is really waht you want to use20:20
dobeyBigWhale: and DrawingArea is at least sensible to connect to "draw" as you were trying to do with the window20:21
dobeybut for a window, it's not so sensible :)20:22
BigWhalecool20:23
BigWhalethanks for all the info20:23
BigWhaleI'll go and hack some now... :>20:24
kenvandinedobey, ubuntuone-dev-tools requires python-dirspec >= 3.120:28
kenvandinebut the latest uploaded is 2.9920:29
dobeyeh? doh20:30
dobeyweird20:30
dobeykenvandine: ah, must have overlooked that when i copied the control file over from nightlies to do the switch to pure dh. fixed now, sorry :)20:32
kenvandineno worries20:32
kenvandineso it didn't really need that?20:32
dobeyno20:33
kenvandinemterry, i just commented on that deja-dup merge proposal, the tests conflicted on merge21:11
kenvandinemterry, i would have tried to clean it up myself, but your change was specific there so didn't want to do the wrong thing21:12
mterrykenvandine, ah, I'll clean and re-comment.  I forgot I updated that test code after21:16
TheMusoGrrrr I hate havving to use the NVIDIA drivers. NVIDIA really needs to be clubbed with a cluestick so they use XRandr.21:37
TheMusonvidia-settings is a poor child's substitute.21:37
TheMusoEspecially since one can only save the monitor config to an xorg.conf file...21:37
TheMusoAt least afaik.21:37
broderi am still firmly of the belief that the NVControl extension *could* be shoehorned into GnomeRR21:38
TheMusoBut... but... THere is XRandr. All other drivers use it. :)21:38
dobey"My enemy's enemy is my friend."21:39
TheMusoI was using Nouveau, and that worked, but afaik it doesn't have deacent PM yet, and Unity was running a little sluggish with my 24" monitor connected.21:40
TheMusoOh yeah, and I am displeased with Lenovo for connecting the VGA port, and the monitor ports on the docking station via the NVIDIA GPU only.21:40
broderugh, that sucks. the vga port on my lenovo and its docking station do work with the intel card21:41
TheMusoI could use the DisplayPort on the machine itself, if 1) I wasn't using a docking statino, and 2) I was using the monitor directly, and not via a KVM.21:42
dobeymy laptops mostly all have intel only anyway21:42
dobeyonly ones i've ever had, that didn't, had ati cards21:43
TheMusoRight, well thats the problem with getting a 15" notebook I guess, hybrid GPUs are now the norm.21:43
dobeybut old powerbook g3, ibook that doesn't turn on, and my really old thinkpad x31 aren't really usable machines to work on any more21:43
dobeyi think the problem is more that nvidia is more the norm these days21:44
TheMusoYeah that too.21:44
TheMusoAt least Lenovo lets me choose which GPU to use in the BIOS, i.e I can turn off hybrid mode/NVIDIA Optimus.21:45
dobeyi try really hard to avoid buying laptops larger than 8" though21:46
TheMusoTo each their own.21:48
=== m_conley is now known as m_conley_away

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