[02:59] /c/c [05:12] * TheMuso notes that at-spi over dbus is a little painful to use, performance and accessibility wise. [05:14] I've seen a lot of discussion about at-spi2 as an example of how to make dbus performance a non-issue by making calls suitably asynchronous. [05:17] Yeah afaik its being worked on upstream. [05:18] But there is more than that, as the framework has often got to iterate over a list of objects. Take the pidgin buddy list. With at-spi via dbus, its rather slow to read out every item as you arrow over them, and the further to the bottom you go, the longer it takes. [05:19] Ow [05:20] Yeah. [05:20] The CORBA framework by comparrison, is rather snappy. [08:05] good morning [08:05] Good morning didrocks :) [08:07] hey nigelb :) [08:10] Hey didrocks, nigelb. [08:10] hey TheMuso, how are you? [08:10] didrocks: Not too bad thanks. Yourself? [08:11] TheMuso: I'm fine, thanks. Some snow at less than 100m (of height away) [08:11] Good morning [08:11] didrocks: Nice. [08:11] Morning pitti. [08:12] Guten Morgen pitti [08:19] HOla TheMuso [08:52] morning [08:53] morning [08:54] salut huats [08:54] hey rodrigo_ [08:54] hello didrocks [08:55] salut didrocks, huats [08:57] rodrigo_: speaking French now? :) [08:57] un peu :) [08:58] tu ne sais pas dans quel engrenage tu mets le doigt :) [08:58] (good luck to translate that) [09:05] didrocks, hehe [09:06] la mafia française va me couper le doigt? :) [09:08] rodrigo_: ahah ! non, pas à ce point là :) (cela veut plutôt dire: « tu ne sais pas dans quoi tu t'engages ») [09:09] porquoi? parler français va me poser des problèmes? :) [09:09] rodrigo_: non, tu seras juste *obligé* de nous parler français après :) [09:09] hey rodrigo_ [09:10] rodrigo_, don't worry, finger cuting is a rusian mafia thing, the frenchs have better ways ;-) [09:10] ah, pas de problème, j'aime parler français, mais je ne pratique jamais [09:10] hey seb128 [09:10] seb128, :D [09:11] seb128: that's "la classe à la française", isn't it? :) [09:11] seb128, and how is the french way? :) [09:11] morning rodrigo and seb128 [09:12] rodrigo_, clear head cuting ;-) [09:12] ah, oui :) [09:13] * rodrigo_ learns russian [09:13] oh, obviously I'm in the wrong channel, bye! [09:14] rodrigo_: Больше водки, пожалуйста! [09:14] pitti, sorry, just decided to learn russian, so will answer when I learn :) [09:15] rodrigo_: (it's "more Vodka, please!" :) ) [09:15] ah :) [09:19] hum, here we go again, it's only 10:20 and they start drinking [09:19] "what a team" [09:20] seb128, that is didrocks influence... [09:20] what??? :) [09:20] huats, I know, it has started since he feels comfortable there [09:20] huats: you should notice that seb128 always finish his beer before me at UDS, I think that means something :) [09:20] yeah, we didn't drink at all until didrocks joined the company [09:21] so easy, so easy… :) [09:21] so true at the same time ;-) [09:22] rodrigo_, and seb128 you have no idea how hard it has been to share a room with didrocks at FOSDEM where we all know beers can be found everywhere... [09:23] huats, don't go ever again to FOSDEM with him, yes :D [09:23] I can feel your pain [09:23] * seb128 hugs didrocks [09:24] but yeah, we all like didrocks [09:24] * rodrigo_ hugs didrocks also [09:24] * huats is happy to share that with you all :) [09:24] * didrocks hugs seb128 [09:24] * didrocks hugs rodrigo_ as well [09:26] (while I am giving a training :)) [09:26] how professional it is :) [09:26] hmmm, i left my laptop running 2 firefox builds last night for me to test this morning, and my laptop ran out of disk space! [09:26] :( [09:28] build it in the cloud :) [09:28] (or at least in the DC chroots) [09:28] yeah, i might start doing that. i keep running out of disk space! [09:29] i guess when i'm always running at 99%, that's going to happen ;) [09:29] lol [09:29] * chrisccoulson needs a bigger disk [09:30] hey chrisccoulson, how are you? [09:30] hey seb128, i'm good thanks. other than losing time from my overnight builds failing ;) [09:30] how are you? [09:31] I'm fine thanks [09:31] seems laptops are not fit for firefox [09:31] you should buy a solid desk config ;-) [09:31] chrisccoulson: how much do you have? [09:32] last time I built ffox it didn't take that much, but it might have grown a lot since then [09:37] oh, today is going to be a fun day [09:37] my laptop crashed about 3/4 of the way through a rebuild! [09:38] chrisccoulson: use *cough* PPA *cough* [09:39] nigelb - firefox builds in a PPA is usually slower [09:39] especially if there are build issues, i need access to the build environment to debug them [09:39] ah [09:40] looks like seb is right, time for a solid desktop :) [09:40] (and a netbook to ssh into it) [09:40] as a start, an USB HD should help a lot [09:40] yeah ;) [09:40] i might get a USB HD :) [09:41] you don't already have one? [09:41] where do you store all your music and videos? [09:41] (the backups, I mean) [09:41] pitti - i do, but it's used to mirror the contents of the disk on my desktop [09:41] (which has all my photos and music on) [09:41] and that's nearly full too ;) [09:41] although, not quite as bad, that's only running at about 90% [09:42] omg, /var/cache/pbuilder/build = 25GB! [09:42] that's crashed build environments laying around [09:42] instant win === MacSlow is now known as MacSlow|afk [09:43] excellent, i now have *loads* of disk space [09:45] 83% full now, that's the lowest it's been for ages [09:45] hah [09:45] chrisccoulson: you actually build all stuff in pbuilder? [09:45] pitti - not normally [09:45] isn't that maximizing pain, time, and disk space? [09:46] only towards the end of the cycle when i need to be sure stuff really works [09:46] h [09:46] ah [09:46] i usually just use debuild, which is what i'm doing now [09:48] my home folder is 65GB [09:50] oh, i can get rid of ia32-libs source package [09:50] i'm never going to touch that again ;) [09:51] chrisccoulson: getting more space with the rm -rf axe - always the best kind :) [09:52] pitti - yeah, i've just freed up another 9GB with that axe ;) [09:57] wow, 36GB available now [10:00] chrisccoulson: that should be enough for how many build trees? 20? [10:04] Hi All, it seems that the ubuntu-desktop ppa cocked up my compiz. I've now got it close to working, but I've got no window borders. I get the error "compiz (decoration) - Warn: Property ignored because version is 20080901 and decoration plugin version is 20080529" - does anyone know how to fix this? [10:04] pitti - that's enough for 7 build trees ;) [10:05] ouch [10:05] yeah, they're pretty big! === MacSlow|afk is now known as MacSlow === MacSlow is now known as MacSlow|lunch === asac_ is now known as asac [11:19] pitti, should somebody send a meeting reminder for today? [11:20] pitti, we didn't discuss a new time btw, we said at UDS we would no? [11:20] seb128: Jason said he'd still be available today, so we should discuss it today [11:20] seb128: right, I'll send one, thanks [11:21] pitti, thanks [11:21] * seb128 writes activity report [11:22] dear update-manager, why does typing completion is all weird with you [11:27] pitti, https://edge.launchpad.net/~canonical-desktop-team/+contactuser [11:27] pitti, just for info [11:28] pitti, it's easier than having to Cc all team members [11:28] seb128: I have a local "desktopteam" alias here [11:28] but I forgot Rodrigo, it seems [11:29] seb128: rodrigo_ isn't in https://edge.launchpad.net/~canonical-desktop-team/+members either, should he? [11:29] I guess it would make sense for the cycle [11:29] done [11:30] meeting reminder forwarded, too [11:31] pitti, thanks === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away === MacSlow|lunch is now known as MacSlow [12:21] good morning! [12:21] seb128: what is the problem with u-m ? [12:31] Riddell: is that known? [12:31] python -c 'import PyKDE4.kdecore' [12:31] python: /build/buildd/sip4-qt3-4.11.2/siplib/siplib.c:10613: sipEnumType_alloc: Assertion `(((currentType)->td_flags & 0x0007) == 0x0003)' failed. [12:31] Aborted [12:31] Riddell: (breaks apport, jockey, and presumably all other python kde apps) [12:33] pitti: nope, I'll take a look [12:34] thanks! [12:50] seb128: I've got more trouble with pygi; who could I ask about this? [12:50] pitti, #introspection on irc.gnome.org [12:50] or you can ask there [12:50] thanks [13:02] seb128: ah, my initial problem is solved by installing gir1.0-gtk-3.0 (it doesn't work with 2.0) [13:02] seb128: however, I'm quite puzzled how to tell my python code to use GTK 3 now [13:02] moch in #introspection doesn't know either [13:03] pitti, Gtk.require_version [13:04] pitti, Gtk.require_version = 3 [13:05] >>> print Gtk.get_minor_version() [13:05] 91 [13:05] hm, I add that right after the import [13:05] $ PYTHONPATH=. python gtk/jockey-gtk [13:05] /usr/lib/pymodules/python2.6/gtk-2.0/gi/module.py:119: Warning: cannot register existing type `GtkActionGroup' [13:05] [...] [13:05] still seems to import the gtk-2.0 bits [13:05] I'm not really sure but kenvandine and mterry were discussing it the other day [13:05] mterry got it working [13:05] he should be online in a bit [13:06] seb128: thanks [13:06] funnily enough, python2.7 doesn't have above GtkActionGroup problem [13:08] pitti, try Gtk.require_version("3.0") [13:08] seb128: right, that seems to work better [13:08] at least it ignores ... = 4, but properly fails on ("4.0") [13:09] now it at least doesn't crash any more, just doesn't display anything yet. that's progress! [13:09] :) [13:10] pitti, ;-) [13:21] fta: ping? [13:22] hey cyphermox [13:22] hey didrocks [13:23] cyphermox: where are you with evolution? Will be nice if we can upload it soon :) [13:23] didrocks, I can't really find something obvious to fix the migration, if that's what you mean [13:24] cyphermox: what upstream told? if you provide all the info they are normally helpful [13:24] it looks like both evo and e-d-s keep files there and try to migrate them but something fails, not sure what :/ [13:24] weird that they didn't fix this [13:24] didrocks, like I said before, mbarnes looked at the code and saw it seemed fine [13:24] cyphermox: did you look at the git commits? [13:24] not very recently [13:24] wow, thanks laptop for ruining another build [13:25] cyphermox: did you ask him as told where the code handle that and use git blame? [13:25] didrocks, didn't have to, it's pretty obvious. that's all in shell/e-shell-migrate.c for evo and in the various */libebook-*/e-*-migrate.c for e-d-s :) [13:26] *unless* [13:26] this would be fixed with getting things straight from git HEAD, I haven't tried yet (was busy with synaptic) [13:27] cyphermox: so, git blame to find if there are new commits in thse files :) [13:27] cyphermox: do you think you will have time this week? [13:27] right [13:27] yeah, I could get that done today [13:27] nice :) [13:28] Nov 16 02:20:24 chris-laptop kernel: [116946.010026] [Hardware Error]: Machine check events logged => nice! [13:33] * rodrigo_ -> lunch [13:49] pitti, i think what mterry found was that the gtk version isn't 3.0, it is 2.91 [13:49] actually, i think Gtk.require_version("3.0") worked, but it was confusing because when you would check for the major_version it returned 2 [13:49] because the real version is 2.91 [13:51] kenvandine, pitti: without the require_version, you get 2.0. So you need the require_version. But yeah, major_version will still be 2 [13:51] (even with require_version 3.0, because it gives you 2.91.x) [13:52] no confusing at all :) [13:53] mterry, kenvandine: ok, thanks for confirming; I think Gtk.require_version('3.0') does what I need [13:53] i think it does [13:53] now I run into tons of trouble with assertions/crashes, having to call encode('UTF-8') everywhere (because Gtk expects UTF-8 strings, not unicode), and into all other sorts of type errors [13:54] slowly crawling through :) [13:54] but lunch first [13:54] hmmm, the x-scheme-handler changes in glib are a pain. http links open in a new browser window rather than a new tab now, because it just uses the standard Exec line from the desktop file [13:56] seems easy to build as I did it in jhbuild without an itch [13:56] (hello) [13:56] oups not the right window [13:56] seb128, didrocks, just to make sure, RB stays supported, in natty, right ? [13:57] (bansheed doesnt work on arm and the discussion came up in the meeting) [13:57] chrisccoulson_, and I'm assuming that change is what broke ghelp: links [13:57] mterry, yes [13:58] mterry - we need to update the desktop files for all the protocol handlers, and then set the defaults in desktop-file-utils as well [13:58] when i upload firefox later, it will have part of the fix in [13:58] i'm just wondering whether to change the Exec line for firefox and add -new-tab to it [13:58] it doesn't seem to affect a normal startup [13:59] chrisccoulson_, or you could add a special .desktop just for the handler [13:59] ogra_ac, define "supported" [14:00] seb128 - i could do. the issue is that firefox needs to have knowledge of the desktop file so it can set itself as the default browser [14:00] and that's already tricky enough, as the desktop files are distro-provided [14:00] currently, it will just use "firefox.desktop" [14:00] when they've accepted my patchset ;) [14:01] ok [14:01] seb128, in "main" [14:01] i.e. getting security attention and SRU love [14:01] it doesn't need SRU love - it's already perfect! [14:02] ;) [14:02] ogra_ac, I don't know, I don't think rb ever had a security update [14:02] we have to ship it for arm [14:02] ogra_ac, we do sru things in universe as well [14:02] well, i mean the canonical supported set of apps ;) [14:02] it will stay available but help to maintain it is welcome [14:02] seb128: 4 updates in lucid, none in maverick yet [14:02] k, thanks [14:03] ogra_ac, it's not really clear what we support or not, we don't support everything in main in any case [14:03] micahg, security updates? [14:03] or rather 1 in maverick in -proposed, no security updates :) [14:03] seb128, well, unless upstream mono fixes mono on ram we dont have much choice [14:03] in Hardy there was one [14:03] *arm [14:04] ogra_ac, in any case those questions don't really make sense [14:04] it's not our default player, we still like it and will try to keep it uptodate [14:04] ok [14:04] but you know how it goes with resources [14:04] seb128: and the hardy -security update was for the FF36 transition [14:04] indeed [14:05] micahg, I was just trying to point that the question was rather theoric, rb doesn't have lot of security updates usually [14:05] right [14:05] it will get less work for sure not being in the default installation [14:05] but it doesn't mean it will not be in a shipable state [14:05] ok [14:05] well, we'll be your testbed ;) [14:06] * ogra_ac shakes up the pillows [14:06] well, feel free to take over it if you want ;-) [14:06] i'll look into bugs i find [14:06] for sure [14:06] thanks [14:12] ogra_ac: can you try dget -x http://people.canonical.com/~didrocks/netbook-meta_2.036.dsc and build it on an armel system to ensure it's still working? (as ppa doesn't have armel, I can't test it there) please? [14:13] didrocks, just upload it, if the images break we will fix it [14:13] its not that we are close to A1 yet ;) [14:13] ogra_ac: ok, I just wait to make unity entering natty and then, will upload that [14:13] ogra_ac: I'll do the same for the -settings, just building on armel for now and ask for binary removal [14:14] thanks [14:14] (not sure a manual removal is needed when getting from arch:all to arch: on !something btw) [14:14] yw :) [14:35] didrocks, didiiiiiiiier [14:35] <- just updated compiz to 0.9 [14:35] not a success [14:36] seb128: what happened? [14:36] (natty, right?) [14:36] the first issue is that you didn't update the shlibs on libdecorator0 [14:37] I still had 0.8 and the deco .so bailed out on missing symbols [14:37] (I upgraded by doing an apt-get install compiz) [14:38] shlibs on libdecorator0? there is none IIRC [14:38] let me check [14:38] that's the issue [14:38] upgrading compiz didn't upgrade libdecorator0 [14:38] ok, will fix that then (a symbols file will be better and easier) [14:39] what else? [14:39] the gnome compat option was not activated ;-) [14:39] (but you know about that) [14:39] well, you know why it's not :-) [14:40] is ther a way to set my desktop grid from ccsm? [14:40] the number of workspace, lines, columns [14:40] I don't find it in the wall options [14:40] yeah, I didn't find either, and didn't asked yet sam [14:41] ok, find it [14:41] oh? where is it? [14:41] it's in the first category options [14:41] the general compiz options [14:42] oh, that's an entry [14:42] ok, I'm only missing a way to set "go to workspace " shortcuts now [14:42] well, hard to find, I was thinking it was a plugin, didn't notice it doesn't have any checkbox [14:44] seb128: ok, then, the only "unknown issue" was on apt-get install compiz to upgrade, no shlibs/symbols file. Fixing this then :) [14:44] thanks ;-) [14:45] seb128: thanks for reporting this, it got eaten in the cdbs -> dh7 transition I thing :) [14:46] didrocks: oh, you are packaging gnome stuff with dh7 now? do we need a dh_langpacks then? [14:46] pitti, no "GNOME" but compiz ;-) [14:47] but yeah, having dh7 utility for the cdbs magic we have would be nice [14:47] pitti: Amaranth began the compiz work on dh7, so I continued with than. I'll try to have a long on dh7 vs dh_langpacks for A2 [14:47] we have a task open for that since last cycle [14:47] (I should have a WI somewhere for that) [14:47] ok, that shouldn't be too hard; that would go into gnome-pkg-tools? [14:47] didrocks, there is a debhelper bug open and assigned to our team on launchpad [14:48] pitti: right now, there is a manual hack in the rules for desktop files only [14:48] if you want to claim it [14:48] seb128: sure, looking for it [14:48] pitti, well ideally we would get a --with gnome in Debian and modify that one in Ubuntu [14:48] that would avoid having to patch the rules in ubuntu to include that [14:49] *nod* [14:49] if the Debian GNOME guys can agree to b-dep'ing on gnome-pkg-tools for packages that they convert to dh7 [14:49] claimed bug #595008 [14:49] Launchpad bug 595008 in debhelper (Ubuntu) "should port the Ubuntu cdbs custom rules to dh7 (affects: 1) (heat: 20)" [Wishlist,New] https://launchpad.net/bugs/595008 [14:51] didrocks, thanks [14:51] yw :) [14:51] the "shortcuts to go to workspace " not working is likely to drive me crazy [14:52] or all my other shortcuts :) [14:55] seb128: use the gconf backend (/!\ not tested) [14:56] I'm used to ctrl + alt + arrows with my ws layout, so not a crazy thing for me :) [14:57] didrocks, it crashed when i switched it to the gconf backend here ;) === chrisccoulson_ is now known as chrisccoulson [14:57] chrisccoulson: did you report a bug about it? [14:58] didrocks, not yet, i wasn't sure if it was just a transitional issue [14:58] will try it again later once i've updated again [14:58] chrisccoulson: ok, thanks :) [14:58] I try to not clutter my laptop before we decide on either switching the backend or not [14:58] didrocks, oh also I get "compiz not responding" when closing the session [14:58] you might want to activate the session option? [14:58] shouldn't it have a GSettings backend? :) [14:59] seb128: session management plugin? [14:59] yes [14:59] where is the backend selection option? [14:59] seb128: I'll have a try and see if it fixes it [14:59] seb128: in the general options you were talking about lately [14:59] chrisccoulson: depends, because the backend looks the GNOME keys from the same backend [15:00] chrisccoulson: so, if g-s-d still store its values in gconf and we activate the gsettings backend -> fail :) [15:01] the first pixel doesn't react to click, known issue? [15:01] known issue [15:02] seb128: if you want to workaround it, desactivate the active edge in the wall plugin [15:02] that one will annoy me as well I think [15:02] thanks [15:03] it's hard to live with it for long, I agree :) [15:03] didrocks, what option is that? [15:04] seb128: so, in the wall plugin -> Bindings -> Edge flipping -> desactive for all Flip Right/Up/Down/Left [15:05] didrocks, ok, gotcha, thanks [15:05] yw :) [15:05] didrocks, do you have bugs open to track those issues? [15:05] just to make sure they don't fall off the list [15:06] seb128: yes, and it's even fix in trunk for the compiz main plugin (for the edge issue), but last time I tried a make dist, it wasn't stable enough [15:06] and as sam is reorganizing the tree, I prefer him to finish that as cherry-pick wasn't really possible (needs new code in other plugins too) [15:06] already lost 2 hours on that, was enough :) [15:08] does someone know why gsettings had both notion of path and id? seems to be redundantly the same notion to me, so I must miss something [15:15] didrocks, ok, using gconf doesn't fix the keybinding issues [15:16] seb128: hum? smspillaz was pretty sure that without the gconf backend, it won't work. Not sure if something else is broken then. I'll have a look [15:20] cyphermox, pong [15:22] cyphermox, oh, i see. rm -rf /data/bot/upstream/network-manager.git <= done [15:23] cyphermox, next respin in 15h. if you need one faster, let me know [15:26] ok, session handling doesn't work [15:26] enough playing with compiz, let's get back to work [15:26] fta, thanks, it's good enough [15:30] seiflotfy: ok, that for the info (on session handling), will check as well [15:32] didrocks, I may have just found out (finally) what's wrong in the migration code, checking it out now :) [15:33] cyphermox: oh nice, what was it? [15:33] looks to me like at least views is being iterated over to get emptied (and it is empty in all cases) but using a different method than other things which doesn't remove the source directory in the end :) === zyga is now known as zyga-afk [15:34] didrocks, there will still be figuring out what happens to .evolution/*/local though, and everything about cache is unsure. [15:35] didrocks, I think I have added enough debug g_print's and stuff to see it properly though [15:35] cyphermox: ok, nice :) [15:36] I'm just hoping that dropping my old backup in place will be enough to test migration, I'd like to avoid setting up new accounts over and over :) [15:37] cyphermox: well, I have another setup, but also not clean at all as I have half a migration :) [15:39] didrocks, right, well, in reality, I think the migration has always worked, there's just leftovers [15:40] didrocks, I don't think cache, for example, matters much. stuff will just get re-cached.. [15:40] didrocks, any chance you could find .evolution -print and send me the output, so I can compare with the two other runs I did? [15:40] cyphermox: right, but we still need to know if removing the cache live can't damage evolution process :) [15:40] .evolution -print ? [15:41] find .evolution -print [15:41] cyphermox: http://paste.ubuntu.com/533061/ [15:41] cyphermox: should be empty at the end IIRC [15:41] didrocks, right [15:41] but look, no cache for you ;) [15:42] yep [15:43] anyway, thanks, I'm looking. I think at least in your case it's really views that isn't getting removed when it should, but there are other issues because I have more files around in both the attempts I did at home [15:44] ok :) [15:56] didrocks, so the g-s-d issue was a gcc one? [15:57] mterry, did you try to figure why lpi failed to build? [15:57] seb128: right, no --as-needed related or DSO linking though [15:57] seb128, no, didn't get to it. Let me check now [15:57] seb128: doko will look at it, I pastebin him the example and the command line, should be short enough [15:57] didrocks, weird, did it fail locally for you? [15:58] seb128: yeah, it failed locally too [15:58] not for you? [15:58] ok [15:58] I'm still on gcc-4.4 ;-) === tkamppeter_ is now known as tkamppeter [15:58] I will update gcc now [15:58] ok, it seems you cherry-pick natty upgrade? :-) [15:59] well I added natty to my sources.list and I do apt-get install "things I need" [15:59] or want [16:00] I usually take the desktop stack but let other things [16:00]