/srv/irclogs.ubuntu.com/2010/09/15/#ayatana.txt

JanCBorchardtvish, is there already a bug, blueprint or similar poking stick for Launchpad to not send out mails to group members when the bug they are subscribed to through a project has been marked as invalid for it? (tl;dr hundredpapercuts ;)00:23
JanCBorchardtvish, darn, it’s always that I _really_ think about it only when I already sent the message.00:25
ScottKJanCBorchardt: Yes.02:31
ScottK#launchpad is a better channel for such a question though.02:31
klattimerhey, I'm having a problem with updates in maverick10:03
klattimerthe ubuntu-desktop package wants to remove Xorg10:03
lucidfoxo_O10:04
klattimer:/ thinking this is a little weird10:04
lucidfoxCould you pastebin the exact output of apt-get?10:04
klattimerwell, ubuntu-desktop wants to remove itself too for some reason10:04
klattimerlucidfox sure, I'm just updating everything else and doing some installs10:04
klattimersoon as I've done that I'll provide a paste of the stuff10:04
klattimerlucidfox: nevermind, on the command line the packages are being held back10:17
klattimerI was using synaptic to update while I was doing some installs10:17
lucidfoxbug #541756 - OK to close this?10:57
ubot5Launchpad bug 541756 in liferea (Ubuntu) "liferea needs a monochrome icon (affected: 5, heat: 28)" [Wishlist,New] https://launchpad.net/bugs/54175610:57
lucidfoxMonochrome icons are basically a feature of Ubuntu indicators, and Liferea uses the messaging indicator10:57
vishlucidfox: i'd suggest you wait for symbolic icons to land.. app cannot add icons for two different themes11:27
vishlucidfox: afaik, that would be in gnome311:28
Tiibiidiihi, sorry if i ask this here, but it seems the ubuntu channel more closely related to unity12:06
Tiibiidiiwhere should i report a unity bug?12:06
Tiibiidiihere? https://bugs.launchpad.net/unity12:06
Tiibiidiior here? https://bugs.launchpad.net/ubuntu/+source/unity12:07
Tiibiidiithe first one is upstream, so it should be better12:07
Tiibiidiibut it seems that only with the second one i can automatically upload my system's information with ubuntu-bug12:07
Tiibiidiiupstream seems to have sligthly more bugs (so, it seems more used and more relevant)12:08
Tiibiidiidownstream seems to have more new bugs... but this may be a symptom that the developers pay more attention to the upstream bug tracker12:09
=== MacSlow is now known as MacSlow|lunch
lucidfoxTiibiidii> you can link the same bug to both locations12:25
ryehi all, is there anyone with Acer Aspire One with Intel Corporation Mobile 945GM/GMS/GME, 943/940GML Express Integrated Graphics Controller ? unity does not really start now, (unity:2446): ClutterGLX-CRITICAL **: Unable to make the stage window 0x1a0004f the current GLX drawable - Mesa 7.9-devel implementation error: Bad renderbuffer format: 21 - am I the only lucky one now?12:25
Tiibiidiirye, njpatel said that the latest graphics updates in maverick wreak havoc with unity12:26
Tiibiidii(with intel graphics, that is)12:26
Tiibiidiii've also an intel card and i'm refraining to update12:26
njpatelyeah, for me it's with anything gl on intel12:26
njpatelon 94512:26
njpatel965, compiz works okay, but clutter fails12:27
Tiibiidiihe wrote that... uh, here he is :) ... hopefully friday it'll be fixed (if i remeber correctly)12:27
Tiibiidiilucidfox, you're right, but it's more manual work to do :D, and actually i'm not accustomed to link bug in more projects12:29
Tiibiidiiuntil now i've almost only reported bugs to ubuntu, and tried helping with translation on other projects... but maybe it's time to get a little more experience with these things12:30
ryeTiibiidii, njpatel thanks for the info12:34
=== MacSlow|lunch is now known as MacSlow
klattimerdavidbarth, kenvandine are either of you aware of any memory leaks currently affecting libdbusmenu-glib?14:58
klattimerI think I've found some and want to avoid duplicates where possible14:58
kenvandineklattimer, i am not14:59
kenvandinetedg, ^^14:59
tedgklattimer, Hmm, okay.  I'm not aware of any.15:00
kenvandinetedg, your code never leaks, right?15:00
klattimertedg: check out the valgrind log on https://bugs.launchpad.net/indicator-application/+bug/56927315:00
tedgkenvandine, Of course.  By design.  I just free memory randomly just in case ;)15:00
ubot5Launchpad bug 569273 in Application Indicators "memory leak in gnome-power-manager (affected: 58, heat: 305)" [Medium,In progress]15:00
klattimerit really seems it's libdbusmenu-glib15:00
tedgklattimer, Yeah, so ayan has a small GTK program that creates and destroys menus and causes a leak in GTK...15:01
tedgNot that dbsumenu isn't at fault also :)15:02
klattimertedg: ah, that was my second guess15:02
tedgBut, I think that there more than one contributor there.15:02
klattimeryeah, there's lots of small ones in gtk15:02
tedgI think it's caused by the fact that no one creates and destroys menus as rapidly as gpm does with that patch.15:02
klattimeryeah, that's possibly one way to solve it15:03
klattimerthe issue is that it updates frequently15:03
klattimeralthough I've heard some noise from your branches you have a signal for menu clicked now?15:03
tedgI was talking to hughsie a while back and he was saying it depends on the battery firmware how often it updates.15:03
tedgWhich probably explains the variation of folks reporting things on the bug.15:03
klattimeryeah, that was a bit of a question up in the air15:04
klattimerso some people's battery goes like crazy updating the menu, others more leisurely but still gpm grows15:04
klattimer:/15:04
klattimermaybe we can fix the gpm patch by hooking up to your menu clicked signal, and only updating the menu then15:05
klattimerit would lessen the impact of the leak15:05
tedgOr I was thinking just having it update the label instead of rebuilding the whole menu.15:05
tedgIt'd save dbus traffic as well.15:05
klattimerthat's another option15:05
klattimerI'll look at doing it that way15:05
tedgThe valgrind log is interesting.15:05
tedgI wonder if there's something of the wrong type so that hashtable isn't getting free'd.15:06
tedgIn theory dbus-glib should do that... but I don't have a huge amount of faith in it.15:06
klattimertedg: line number?15:06
tedgklattimer, The last entry in the valgrind log.  For properties_copy15:06
tedgIt allocated a hashtable which get's passed to dbus-glib.15:07
klattimeryeah15:07
tedgNo, I think that serialze_menu item has a leak of that hashtable....15:08
tedgUhg.15:08
davidbarthklattimer: hi; thanks for checking that; do you think you could have a patch for maverick still?15:14
klattimerdavidbarth: what's the deadline15:14
davidbarthie, should i target that against the 10.10 milestone?15:14
davidbarthwell, the hard FF is this thursday15:14
klattimertomorrow?15:14
davidbarthyes15:14
klattimerOK, I'll do my best15:14
davidbarthif the leak is severe, it should be commitable after that, or if not, as an SRU or Natty15:14
davidbarthklattimer: cool, thanks for your help15:15
klattimerdavidbarth: this is probably fixable now15:17
klattimerjust looking at the code and in theory it's possible15:18
klattimeroh15:18
klattimer:/15:18
klattimerdavidbarth: nope15:18
klattimerit's not fixable quickly15:18
davidbarthdoh15:18
davidbarthwhat's the issue15:18
klattimeradding/removing devices requires a new menu to be created15:18
klattimerso it's not just a matter of when the battery level changes, change the menu item15:19
klattimerit all routes through the same portion of code for adding/removing and updating values15:19
klattimerso the proper fix is to try and fix the libdbusmenu-glib issues, and possibly an upstream bug on gtk15:19
klattimerdavidbarth: bare in mind, the most major memory leaks are already fixed15:19
davidbarthklattimer: so it's a minor leak15:21
klattimerdavidbarth: in theory it's only minor15:22
davidbarthklattimer: if a gpm triggers the leak every 10s, how long for it to eat 10M, 100M?15:22
klattimer2Mb per hour is as high as I've seen it15:22
davidbarthso 48MB a day15:22
klattimergive or take15:22
klattimerbut as tedg says, it's dependent on battery firmware15:22
davidbarthbut for HW that is really verbose15:22
davidbarthright15:22
klattimerdavidbarth: https://bugs.launchpad.net/indicator-application/+bug/569273 comment #53 has some details15:23
ubot5Launchpad bug 569273 in Application Indicators "memory leak in gnome-power-manager (affected: 58, heat: 305)" [Medium,In progress]15:23
davidbarthmakes sense for an SRU if that can't get into the release within the next 2 weeks (RC)15:23
klattimermaybe you want to reassign to libdbusmenu-glib for now?15:24
tedgklattimer, I think there's two bugs.15:25
tedgklattimer, The one there with the hashtable, but it's in a new function we only started using a lot in Maverick.15:25
klattimertedg: in libdbusmenu and gtkmenu?15:25
tedgklattimer, I think the leak above it in the valgrind log is the same one that was in Lucid, which is the gtkmenu one.15:26
tedgklattimer, So Maverick made it worse, but it's still a problem :)15:26
klattimerhas the gtkmenu bug been upstreamed yet?15:26
klattimerdo you have details/ayan's test code?15:26
tedgklattimer, I believe that ayan hadn't found the leak 100% yet.  Just had a small program to exercise it.15:26
klattimerI could produce a decent valgrind log for upstream to gtk15:27
tedgI'm not sure where ayan is on that... but certainly helping out there would be good.15:27
tedgklattimer, Can you double check this for me?  https://code.edge.launchpad.net/~ted/dbusmenu/hashtable_free/+merge/3554015:29
klattimersure15:29
tedgIt passes the test suite, so it must be perfect :)15:29
klattimerin serialize_menuitem15:31
klattimertedg: it looks good to me15:31
tedgCool.  Thanks klattimer!  One down :)15:32
tedgA million more memory leaks to go :)15:32
klattimerwell, that'll probably solve the majority of the issue15:32
klattimerheh15:32
klattimertracing memory leaks is sometimes like trying to walk on marbles15:32
klattimertedg: the good thing about fixing in libdbusmenu is that we save memory in a lot of apps15:34
klattimerfixing in gtk would be better wins15:34
=== njpatel_ is now known as njpatel
tedgYup.  The more the code's used, the more each line counts.15:36
tedgMight be an interesting metric to calculate using something like Ohloh's DB.15:37
* tedg needs interns15:37
klattimerheh15:40
Paddy_NIWill indicator-appmenu be available for lucid at all?15:41
tedgPaddy_NI, We have no plans to backport it, but someone could.  I believe their are packages in the Unity PPA, but I don't believe they're up-to-date.15:42
tedgthere are....15:43
Paddy_NIOh that would be nice15:43
* tedg apparently offended mpt with his misuse of "there" :)15:43
Paddy_NIAs far as I can see there will be no need for the gnome-globalmenu project15:43
Paddy_NIHe another question, how much of the gnome global menu code did you use?15:46
Paddy_NI*hey15:46
tedgPaddy_NI, Not any, we have a different approach.  Mostly using DBus which they weren't.  So we talked to them to learn about the issues get ideas, but the code sadly wasn't useful.15:50
klattimertedg: would be nice of have an "application" menu as part of appmenu15:51
klattimeri.e. the current application name15:51
klattimersort of bold... with window manager functions ;)15:51
klattimerlike a certain fruit company we all know and love15:51
tedgklattimer, Yeah, we were just focused on getting the menus out for now.  It'd be interesting to do something like that.15:52
tedgklattimer, Not sure.  It seems like it's a lot of duplication with the File menu in most Linux apps though.15:52
klattimerit's a massive improvement over globalmenu anyway15:52
klattimerwell, file->quit never makes sense15:53
tedgThe problem is changing applications just sucks.  I think we learned that with appindicators :)15:53
klattimeryeah, I can see the problems15:53
klattimerit's just not easy to say "look, this is a great api, and much better interaction, lets all switch"15:53
klattimerand get people on board that way15:54
tedgYeah, unfortunately there's a perception problem we have there.  Where any Canonical project is "not upstream" so therefore people sadly avoid dependencies.  Hopefully that'll change over time :(15:55
Paddy_NItedg, Well I for one look forward to trying indicator-appmenu and Unity15:56
Paddy_NI:)15:57
Paddy_NIfrom what I hear the dbus approach is much better15:57
klattimerPaddy_NI: it is and doesn't rely on deprecated gnome libs15:57
klattimerso it'll survive the gnome 3 turmoil15:57
Paddy_NIExcellent15:57
Paddy_NIWas getting QT applications to work with it very difficult?15:59
tedgagateau, ^16:03
Paddy_NII was also wondering when application indicators land will the top panel in gnome be a little crowded on the netbook edition with the window controls and appmenu?16:03
Paddy_NIIt is just I have not seen a screenshot of this as yet16:04
agateauPaddy_NI: we have a Qt implementation of dbusmenu and a Qt patch16:04
tedgPaddy_NI, It depends on how many applications you run :)16:04
=== rgreening_ is now known as rgreening
jcastrohttps://bugs.edge.launchpad.net/ubuntu/+source/tomboy/+bug/62774416:53
ubot5Launchpad bug 627744 in tomboy (Ubuntu) "note names are blank in applet menu (affected: 2, heat: 18)" [Undecided,Confirmed]16:53
jcastrois anyone available to look at this?16:53
jcastroklattimer: how's your c# these days?16:56
klattimermy c# is practically non existent atm16:57
klattimerhaven't used it in 2years16:57
jcastrok, maybe sense has time16:57
klattimerbut it's all code at the end of the day16:57
klattimerso if you're hardup assign it to me16:57
jcastroI'll assign for now but ask for help16:57
klattimerk17:00
jcastroheya vish17:00
klattimerjcastro: if it's any help, i cant reproduce the above bug17:05
klattimerworksforme17:05
jcastroit also worksforme17:05
vishjcastro: hey.17:07
jcastrovish: if any of your papercutter gang wants to look at #627744 then I won't complain. :D17:09
vishbug 62774417:10
jcastrokenvandine: how's your tomboy applet dropdown?17:10
ubot5Launchpad bug 627744 in tomboy (Ubuntu) "note names are blank in applet menu (affected: 2, heat: 20)" [Undecided,Confirmed] https://launchpad.net/bugs/62774417:10
vishjcastro: hehe , the gang got scattered after the UIF! ;)  but i think lucidfox likes those.. and sense is in hiding :D17:12
* vish tries to find someone..17:12
jcastrocool, if I see them I'll ping them, but since you're on opposite time zones if you see them ... :)17:12
vishsure..17:13
ronocCimi: you there ?17:56
=== MacSlow is now known as MacSlow|break
Cimironoc: yep18:20
ronochey Cimi, hows it going ?18:20
Cimicould be better but it's ok ;)18:20
Cimithank you18:20
ronocgood stuff, this bug is annoying me -> https://bugs.launchpad.net/indicator-sound/+bug/63277418:21
ubot5Launchpad bug 632774 in The Sound Menu "right border of transport control is different colour (affected: 1, heat: 46)" [Low,Confirmed]18:21
Cimieheh18:21
Cimime too18:21
ronocits seems as if the bg colour for ambience is slightly off in contrast to what is use to draw the border of the menuitme18:21
Cimithe problem is that the GtkDrawinArea seems filled with bg[NORMAL]18:22
ronocis that not correct ?18:22
ronocshould it not be ?18:22
CimiI mean18:22
CimiI would prefer to have it transparent18:22
Cimibratsche: ?18:22
ronocahh its not a biggie18:22
ronoci was just hoping to get a quick fix in for tonight18:23
Cimimaybe he has an idea18:23
ronoci think bratsche is at lunch18:23
ronoccool18:23
ronocbratsche, can you ping me when you are back18:24
ronocbiab18:25
bratscheI'm back.18:34
bratscheronoc, Cimi18:34
bratscheCimi: We can't really fill anything with transparent until we land rgba.  But we need an idea of how to specify rgba colors in the theme, per-widget and per-state.18:36
Cimibratsche: ok thank you18:38
bratscheCimi: I haven't been following, but is there anything new theming-wise in git master that we should be aware of?18:38
bratscheDo you know?18:38
Cimibratsche: maybe the css thing by carlos garnacho18:38
Cimiwait a second18:38
bratscheBut has that landed?18:39
Cimihttp://blogs.gnome.org/carlosg/2010/08/23/css-like-styling-for-gtk/18:39
bratscheCimi: I updated my patch to add corner resize grips today, btw.18:39
Cimicool18:40
bratscheCimi: I'll pull down garnacho's branch and look at it.  Ideally I want to be able to specify an RGBA color to background-color or foreground-color properties in the CSS.18:40
bratscheIf it doesn't support that, I'll just add it. :)18:40
ronocbratsche, so will I park this bug for now ?18:47
ronoci.e. can't be fixed until this rgba stuff lands18:48
bratscheUhh.. that doesn't look like an RGBA issue.  That just looks like there are two slightly different colors being themed somewhere.18:49
ronocyeah that is what I thought18:49
bratscheBut I dunno.18:49
ronocCimi, ^18:50
ronocwould it be possible Cimi  that the bgnormal is slightly different in the theme compared against what is actually used to draw the border of the menu18:50
Cimironoc: bratsche simply, the menu background has a horizontal gradient :)19:08
Cimiit's not a flat bg[normal]19:08
Cimiwhile the gtkdrawingarea is a flat bg[NORMAL]19:08
bratscheAh, makes sense.19:08
ronocah19:08
* Cimi had to go, leaving irc on anyway19:08
bratscheNo worries.19:08
=== MacSlow|break is now known as MacSlow
vishkenvandine: tedg: seen Bug 636693 ? it has a patch..21:05
ubot5Launchpad bug 636693 in indicator-session (Ubuntu) "Redundant lock before launching guest session (affected: 1, heat: 1728)" [Undecided,New] https://launchpad.net/bugs/63669321:05
kenvandinei haven't21:05
vishseems like we are doing it twice there.. would be nice if someone commented on the patch21:07
tedgvish, No, looking.21:08
vishthx21:09
Paddy_NIHey this is a little offtopic but have any of you read this or had to sign the agreement --> http://aseigo.blogspot.com/2010/09/copyright-assignments-gone-wild-or-why.html21:43
kklimondaPaddy_NI: it's completely offtopi but I've read it :)22:22
Paddy_NIPerhaps canonical should rethink how they treat their contributors :)22:23
kklimondaI disagree but that's my humble opinion22:24
kklimondasure, there is a question whether it's a valid copyright assignment at all as it isn't signed in any way.22:24
kklimondaI haven't had any objections to sign the agreement myself - but then I have a pragmatic view about things and what I care about is getting things fixed and not about who owns the code or what can he do in the future with it.22:27
JanCBorchardtvish, I actually just came online to ask: »WTF is up with the bug watch updater!?« :)23:29
vishJanCBorchardt: yeah.. they just turned them on.. :)23:29
nigelbJanCBorchardt: there was un update23:30
nigelbso, now it updates importance too23:30
vishJanCBorchardt: also , re: the lp unsubscribe , there is a bug about that , but no one seems to have reached a consensus how to handle it23:30
JanCBorchardtvish, nigelb: ah, thanks!23:31
JanCBorchardtvish, do you have a link?23:31
vishJanCBorchardt: not sure , but ask in #launchpad or try searching ;)23:32
JanCBorchardtvish, ok23:34
vishJanCBorchardt: oh , i havent forgotten the mail either.. :)23:36
JanCBorchardtvish, this looks like it: https://bugs.edge.launchpad.net/malone/+bug/20498023:38
ubot5Launchpad bug 204980 in Launchpad Bugs "bug contacts should be able to unsubscribe from implicit subscriptions (affected: 4, heat: 44)" [High,Triaged]23:38
vishJanCBorchardt: yup , thats the one :)23:39
devildantevish, now that's five :p23:39
vishdevildante: fail! ;p23:39
devildanteargh23:39
vishdevildante: this was already in the 3 ;)23:39

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