[00:12] apachelogger: :P [06:42] apachelogger: since 4.7 SC rc is available: where can I find it? [06:49] Mamarok: Muon package manager? [06:50] karmo: not really more usefull than apt-get, what I need is the PPA [06:53] Mamarok: in oneiric you don't need ppa. unfortunately i don't know if there is ppa for stable kubuntu versions... [06:54] ah, so not available for 11.04 then [06:55] * Mamarok is a sad panda :( [09:53] Mamarok: we are barely done with oneiric, so I doubt there will be 11.04 packages before final release [09:54] apachelogger: OK, thanks anyway, I can survive another 2 weeks :) [09:55] Mamarok: http://www.youtube.com/watch?v=QH2-TGUlwu4 [09:56] and should that not help with making the time pass by fast.... http://www.youtube.com/watch?v=8AOfbnGkuGc this one will [10:05] could someone fix phonon to include the flipping qt designer plugin [10:10] ScottK: in case you have some time, could you investigate why python-kde4 depends on libpython2.6? we can't afford having 3 python versions on the cd [10:11] oh right [10:12] debfx: how much oversize do we have right now? [10:12] I noticed last week that all images were too fat [10:12] or was it this week [10:12] apachelogger: 21 MB [10:12] some week anyway ^^ [10:12] debfx: eek, that is rather a lot [10:16] yep [10:16] kde-workspace-data +2871 kB [10:16] kdm +1226 kB [10:16] those should be investigated as well [10:17] wallpapers perhaps? [10:17] jr made everything use one wallpaper file [10:17] at one resolution [10:17] which is brr anyway IMHO :S [10:19] we probably need to drop libreoffice-help-en-us [10:20] libreoffice opens the online documentation in a web browser when it's not there locally so it's not that bad [10:20] have we already decided which wallpaper to ship by default? [10:25] hm [10:25] debfx: what kde ships by default [10:26] appmenu-gtk3 appmenu-gtk [10:26] why do we need those? [10:26] and why do we need both? [10:26] to make the global menu work with gtk2 and gtk3 apps [10:26] Depends: libatk1.0-0 (>= 1.12.4), libc6 (>= 2.2.5), libcairo2 (>= 1.2.4), libdbusmenu-glib3 (>= 0.4.2), libdbusmenu-gtk3 (>= 0.4.2), libfontconfig1 (>= 2.8.0), libfreetype6 (>= 2.2.1), libgdk-pixbuf2.0-0 (>= 2.22.0), libglib2.0-0 (>= 2.26.0), libgtk2.0-0 (>= 2.21.2-0ubuntu6), libpango1.0-0 (>= 1.14.0 [10:26] both are on the cd anyway so appmenu only adds a few kb [10:27] debfx: why do we have gtk3 on the cd? [10:28] we could ditch ksystemlog ^^ [10:28] gconf2 recommends it [10:28] * apachelogger finds the application horrible in anyways anyway [10:28] debfx: why do we have gconf2 [10:29] iirc gstreamer depends on it [10:29] gstreamer0.10-plugins-good [10:30] oh and jon the taco of course needed to use qtgstreamer in qapt [10:31] 1 mib [10:32] ah, nvm, more like 100kb [10:32] virtuoso is 3 though [10:32] zeitgeist-core 1.1 [10:34] debfx: I suppose we also have some random growth due to docs in app packages [10:35] apachelogger: https://felix.fobos.de/kubuntu/kubuntu-oneiric-cd-amd64-diff.htm [10:36] everything is getting fater all the time [10:37] debfx: how did you make that handy page? [10:37] * apachelogger wonders why plasma-widget-networkmanagement grew [10:38] apachelogger: I parse the .manifest files of the images and get the package size with python-apt [10:38] groovy [10:39] debfx: you should totally put that in kubuntu-dev-tools ^^ [10:39] apachelogger: all the networkmangement packages collapsed into plasma-widget-n [10:39] or ubuntu-dev-tools [10:39] debfx: what other packages were there [10:40] Or ubuntu-qa-tools, as those are the folk most likely to be looking at this sort of thing. [10:40] knm-runtime was only 535kb the page sez [10:40] Or create a branch that automatically does this for all the flavours, and have ubuntuwire host it as a regular report. [10:40] * apachelogger likes that last option [10:40] also it reminds me of automated package QA for some reason [10:41] for which I have a blog post lingering I noticed yesterday [10:41] apachelogger: it contains a copy of the nm solid plugin which works with nm0.9 [10:41] and that is fater than 0.8 or whatever we had before? [10:42] the ugly part of the script is a manual mapping of packages that have been renamed [10:42] make it a config [10:42] the nm0.8 solid plugin is in kde-workspace [10:42] * apachelogger also made kde-l10n-common config driven :D [10:42] debfx: *is* it still there? [10:42] if so we should like shoot it dead [10:42] probably [10:43] it builddeps network-manager-dev (>= 0.7.0) [10:43] a config would be an improvement but it should automatically detect library and kernel abi bumps [10:44] oh right, we need to do kwingles [10:44] lol [10:44] debfx: well, that should be a simple enough regex [10:44] anyhow [10:45] even though the good ol kde-workspace builddeps on nm-dev, cmake does not detect it ^^ [10:45] debfx: no nm plugin in workspace anymore [10:45] just fake and wicd [10:46] which makes me wonder why fake is installed actually [10:46] right, it would have failed the build since nm 0.8 isn't in the archive anymore [10:47] is that plunder not api compatible? [10:47] not at all [10:47] in fact we can't migrate nm0.8 connections at the moment [10:49] hm [10:49] lol [10:49] fun software [10:49] debfx: so, I do not see why workspace-data grew [10:49] of course diffing two files with different name in different revisions is not all easy to do with bzr [10:50] perhaps there is a recursive match rule in the install though [10:50] * apachelogger compares debs [10:53] debfx: http://paste.ubuntu.com/644703/ [10:55] just as I suspected [10:55] \o [10:57] yeah we need to get rid of those [11:01] Nokia N950 1 Device Sent to Customer [11:01] :D [11:02] shadeslayer: you should consider yourself lucky that you do not have to mess with video drawing and opengl [11:02] * apachelogger is twisting his brain dry [11:03] hahaha [11:03] apachelogger: el neato, whatcha going to do with the phone? [11:03] tennis [11:03] apachelogger: i still have to mess with the dbus server which has all sorts of weird shit in it [11:05] right now i just have one last problem to take care of and everything should be good to go [11:06] debfx: was the kate packaging any good? [11:06] shadeslayer: well, I'd picky a shitty dbus server over gl anyway [11:06] anytime [11:06] as long as it has stable API [11:06] muhahaha [11:07] the API is stable, the part where you have figure out key value pairs to set a proper config, not so much [11:09] debfx: you could debdiff stuff you list on the diff page [11:09] would make investigation easier I supose :) [11:18] apachelogger: free for a while? [11:24] shadeslayer: I had to change a few things (see bzr branch) [11:25] katepart still needs to be split into an own package but I'm busy fighting with qtwebkit [11:25] oh [11:25] o/ [11:25] shadeslayer: that question will be answered with no until september 2018 [11:26] I guess waiting for that to happen would be inefficient [11:26] whats so special about September 2018 [11:29] debfx: thanks :) [11:29] apachelogger: debdiff doesn't show file size changes though, right? [11:30] nope [11:31] debfx: but you get the file size changes, if they are noticably different that they would show up in the tables, you get the old deb and the new and run debdiff on it, then link the resulting diff from the table [11:31] that way one can easily find out why the size changes [11:31] if there were file changes [11:32] if no files were changed but the size is different something grew, hard to get that checked [11:32] well, you could dpkg-deb -c and parse the size, but IMHO that is a bit unreliable and complicated [11:32] shadeslayer: so whats up? [11:33] apachelogger: i have a QComboBox that lists the friendly names of a akonadi resource [11:34] now i map that to the akonadi source name using a QMap, the config stores the source name, when i read the config i get the source name back and i was wondering if there was a method to set the index of the combobox to match the friendly name to the sourcename [11:35] hrm, anyone know if I can get Kate on windows? [11:35] jussi: kde windows isntaller [11:35] sigh [11:35] why does gmail keep changing my inbox look [11:35] shadeslayer: ehm [11:35] apachelogger: yeah i know, i worded it poorly [11:35] apachelogger: I mean kate only, is that possible? [11:36] shadeslayer: why not use a model? [11:36] QComboBox and a model? [11:36] sure [11:36] have you had any problem compiling soprano ? [11:36] seems much more reliable [11:36] i didn't know that it was possib;e [11:36] afiestas: nope [11:36] I'm having some issues with raptor1 raptor2 raptorXXXXXX xd [11:36] shadeslayer: sure setModel() [11:36] *possible [11:37] i seriously need to devote time to MVC [11:37] anyhow [11:37] shadeslayer: setCurrentIndex [11:37] (index) [11:37] apachelogger: thats what i was using [11:37] but its not working the way its supposed to :P [11:37] why not? [11:38] how do i map a index to friendly name or a source name? [11:38] you iter over all items in the qcb and get the index of the item that matches [11:38] then you set that [11:38] shadeslayer: you cannot, point being that your approach is ugly as it needs to rely on strings [11:38] thats what i thought [11:38] yes [11:38] which is why i want to avoid it [11:38] you need to string compare all items and then set the index to the one that matches [11:39] with a model you'd be able to find the index within the model [11:39] right, but i don't want to do that because thats very sub optimal [11:39] reading up on models then [11:39] and inside the model you can ensure persitant indexes etc. [11:39] also [11:39] oooh [11:39] shadeslayer: can the friendly names be translated? [11:40] afiestas: we have soprano 2.6.51 which carries a raptor2 patch [11:40] you probably want that [11:40] apachelogger: like i18n calls? [11:40] shadeslayer: if so you want to avoid string compares all and entirely [11:40] yofel: I do! I do [11:40] shadeslayer: for example [11:40] shadeslayer: I dunno what a friendly name is :P [11:40] nor where it comes from [11:40] though trueg told me that raptor2 support is not finished and fully working yet [11:40] apachelogger: like "Personal Contacts" [11:40] there is one bug left or something like that [11:40] but since it is friendly and UI visible it supposedly should be translated [11:40] so you cannot rely on strings [11:40] and the source name would be akonadi:?collection=41 [11:40] if the user changes his language all will fall apart [11:40] afiestas: well, we can't really choose here either :S [11:41] apachelogger: yeah, thats something that hit me last night as well [11:41] afiestas: you know where to get the patch from? [11:41] shadeslayer: collection number is unique I guess? [11:41] shadeslayer: there surely is a way to uniquely identify an akonadi resource instance/collection [11:41] and that is the value you want to look up/store [11:42] yofel: nope, but wait seems that trunk is compiling well against extra/soprano 2.6.50 which is the one ship with the distro I'm using righ now [11:42] if it compiles with the distro one, is good enough :p [11:42] well, sounds new enough ^^ [11:42] apachelogger: yeah, collection number is unique [11:43] apachelogger: thats not the problem, i can easily make a QStringMap which maps everything easily during initial ui setup [11:43] so you'd have a model of at least an internal index mapping to your uuid and the name of that item [11:43] the problem is when i need to read the config [11:43] and set appropriate values [11:43] that is why you only store the uuid [11:43] i.e. the collection number [11:44] oh, then i can read the index of the combobox [11:44] and use the appropriate source name to use in the config [11:44] something like that, yes [11:45] sounds about right [11:45] * shadeslayer goes implementing [11:47] shadeslayer: did you break kdevelop? [11:47] is kdevelop broken because you broke kate? [11:47] is kate fixed yet? [11:49] what's broken with kdevelop? [11:49] well [11:49] kdevelop uses the katepart [11:49] so if katepart is still kaput, kdevelop will be too [11:50] well, if you're not using ninja kate it'll probably be broken [11:50] shadeslayer: put katepart into seperate package? [11:51] sure can do [11:51] debfx: ^^ is that alright on your end? [11:51] would be good IMO since a ton of apps seems to use it [11:51] *seem [11:52] * yofel checks if anything else was moved there from kdelibs5-plugins [11:53] i'll do in a bit, have some other things i need to take care of right now [11:53] *do it in a ... [11:54] there's also a ton of katepart related stuff in -data [11:54] which was in kdelibs5-data before [11:54] omg [11:54] amarok++ [11:54] http://www.youtube.com/watch?v=NbTEVbQLC8s [11:54] markey: remember that? [11:55] wow, them good ol' times, how I miss them [11:55] shadeslayer: usr/lib/kde4/ktexteditor_* was also in kdelibs5-plugins, so that probably belongs to katepart too (maybe...) [11:56] LOL @ video ^^ [11:56] yeah ^^ :P [11:57] thats more like the Qt 5 Dance xD [12:00] whow so? [12:00] *how [12:15] apachelogger: yeah, I always hated it :p [12:15] awww, you have no taste whatsoever [12:30] apachelogger: this is weird, mapping the QComboBox index to source name gives me : 0 "akonadi:?collection=16" [12:30] 0 "akonadi:?collection=41" [12:31] i wonder why that value is not incremented [12:33] I do not understand [12:38] !botsnack [12:38] Yum! Err, I mean, APT! [12:39] hm, thinking about katepart, we'll have to make kdelibs5-plugins depend on that package for transitional reasons, or stuff like kdevelop or kile won't work since they don't depend on it [12:47] can someone upload mobipocket so that get's reviewed? [12:47] *gets [12:53] yofel: have you fixed the problem that it installs a okular desktop file with mimetypes but doesn't depend on okular? [12:53] apachelogger said adding TryExec=okular would fix it [12:55] hm, nope, but pretty much everything in the package except the strigi plugin has something todo with okular so I would rather make it depend on okular [12:56] for some reason I didn't get a lintian warning though [12:56] yofel: depend is the wrong relationship [12:57] if anything it should be recommends [12:57] + tryexec is in order anyway [12:57] (upstream) [12:57] yofel: then you should fix your lintian :) [12:57] I currently made it enhance okular [12:57] it should emit W: mobipocket: desktop-command-not-in-package usr/share/applications/kde4/okularApplication_mobi.desktop okular [12:59] I'll make it recommend it and file a bug upstream [12:59] and add a patch for tryexec [13:01] Nightrose: do you know if eean submitted the mid term eval yet? I cant seem to reach him [13:01] * yofel goes reading desktop file spec [13:02] apachelogger: if I understand this right you need both Exec and TryExec? [13:03] yofel: yes, TryExec is nothing but a prerequisite [13:03] ok [13:03] if tryexec fails the desktop file will not be displayed [13:03] sorta like conditional NoDisplay [13:04] apachelogger: he did [13:04] also TryExec does not have to be the same as Exec, essentially you could have an inline shell expression there [13:04] Nightrose: kthx [13:14] what should I be blogging about this weekend? [13:14] automated package QA thoughts or more phonon qml stuff? === ximion2 is now known as ximion [13:29] [lp:~kubuntu-packagers/kubuntu-packaging/mobipocket] Philip Muškovac * 11 * debian/ (5 files in 2 dirs) recommend okular and use TryExec in okularApplication_mobi.desktop [13:29] debfx: ^ [13:30] what do we do with plasma-widget-kubuntu-feedback? === Earthwings_ is now known as Earthwings [13:30] yofel: hm, you only renamed the binary package? [13:31] * debfx doesn't like those generic names [13:31] apachelogger: rm -rf [13:31] feel free to change it again, but I don't see why we should split it up since most of that is for okular [13:32] debfx: wanna do that now or I should make a work item? [13:32] yofel: I mean renaming the source package to kdegraphics-mobipocket [13:33] apachelogger: looks like it's already removed [13:33] you'll need to rename the tarball then, which I wouldn't do without talking to upstream first, or that'll be quite error prone [13:33] apachelogger: what do you think? ^ [13:33] debfx: from archive too? [13:33] yofel: why is that error prone? [13:34] yofel: FWIW IMHO upstream should also rename the tar to kdegraphics-mobipocket [13:34] mobipocket is no good name for that tar [13:34] yes, from the archive, where else do we need to remove it? [13:34] hm, nvm that, I'll send a mail to one of the maintainers [13:34] debfx: seed [13:35] debfx: FWIW, I'm not against renaming it, so if you want to rename the source too go ahead [13:36] yofel: ok, I'll upload it later today or tomorrow [13:36] yofel, debfx: renaming the binary because of generic name but not the source seems like bougs application of pita policies :P [13:37] true [13:37] hm, who decided the tar names? the maintainers or dirk? [13:38] tar == repo name [13:38] usually [13:38] but dirk should be able to override that [13:38] yofel: just mail kde-release [13:38] or whatever the list of the release team is [13:38] I keep forgetting ^^ [13:39] release-team@kde.org I think [13:39] I'll do that then [13:41] [lp:~kubuntu-packagers/kubuntu-packaging/meta-kde] Philip Muškovac * 8 * debian/ (changelog control) mobipocket was renamed to kdegraphics-mobipocket [13:52] apachelogger: something like this? http://paste.kde.org/97123 [13:52] yeah [13:53] sending [13:53] ah [13:53] hold on [13:53] ? [13:53] yofel: perhaps make it more about how mobipocket is not reflecting the content at all [13:53] as it would suggest it contains an app called mobipocket [13:53] which is not the case [13:54] ok, "... since it's not the only source that deals with mobipocket files and doesn't contain a "mobipocket" application either as you would think seeing how the other tarballs are named." [13:54] better? [13:54] yus [13:55] k [13:55] * apachelogger had too much coffee [13:55] way too much [13:55] heh [13:56] sent === zkriesse_ is now known as Guest31717 === ximion1 is now known as ximion === shadeslayer_ is now known as shadeslayer [15:10] apachelogger: shouldn't the sekrit chan be retired [15:10] seeing how there are only a few of us === ximion is now known as ximion1 [15:15] which one? [15:16] the bunker [15:17] wasn't that from pre-ktown-ssh-access times? [15:17] yus [15:18] rm -f [15:18] and now there are just 3 of us in there, just close it down i say [15:20] Nightrose: doesn't the appu up event eat up into DS timings? [15:20] shadeslayer: a bit yes [15:20] hmm ... better to attend the one on the 21st then [15:21] Nightrose: btw has Dinesh submitted my evaluation, he has a bad internet connection and i can't get to him [15:22] shadeslayer: yes - all fine - he's not on my poking list [15:22] w00t [15:23] oh you're done? fix kate then :P [15:23] i'm ~done [15:23] * yofel makes some natty backports for kdevelop [15:24] just a small piece of code left that for some reason isn't working even tho it's supposed to work [15:24] yofel: i'll have a look at kate after dinner, my brain is kaput after looking at code all day [15:25] harald has coffee to spare, get some from him :P [15:25] no coffee for me, i'll be up all night and then i'll be back to my old sleep schedule [15:33] anyone an idea why we have no bzr branches for the kdevelop stuff? [15:35] yofel: i think we have branches for packages that involve alot of effort in maintaining [15:35] ah, well, not important [15:36] kdevelop probably doesn't need alot of people working on it, pretty much the same reason rekonq doesn't have a branch [15:36] you kdevplatfrom package uses boost 1.42 btw. I'll fix it [15:36] *your [15:36] feel free to, i need to make a release by tuesday [15:36] or i'll get some whooping [15:36] sure [15:37] weird how updating works but adding doesn't :/ [16:18] Nightrose: here's something for you http://nyan.cat/ [16:19] shadeslayer: OMMG! [16:19] :D [16:20] make someone listen to that for a day nonstop and he'll have a cat phobia ^^ [16:20] or he'll see cats pooping rainbows everywhere [16:20] s/he'll/they'll/ [16:20] shadeslayer meant: "or they'll see cats pooping rainbows everywhere" [16:26] yofel: have you seen this? http://wiki.desktopsummit.org/Workshops_%26_BoFs/2011/KDE_Buildsystem_BoF [16:27] nope [16:32] well, now you have [16:33] what's that supposed to be? some cmake-addon-like-thing so the build stuff is in one place instead of every component having to care about their own scripts? [16:34] super builds? [16:35] uh, you mean it's supposed to create batched up release tars? [16:35] afaik its a CMakeLists.txt file that will pull in git repos and build them like the earlier KDE versions [16:35] yofel: https://projects.kde.org/projects/kde/superbuild [16:36] didn't someone write that already? [16:36] yus [16:37] i didn't know CMake could init git repos and pull sources until i saw that project [16:37] me neither, but I heard of that on -packagers already [16:40] ffffffffuuuuuuuuuuuuuuuuu [16:43] i'm the lord of borkage [16:43] i just made one thing work, and broke 3 other things [16:44] new person to blame when something wroks: GET [16:44] yay, now the server works [16:44] weird shit === mgraesslin_ is now known as mgraesslin [17:09] 3000s, if I hear this cat any longer I'll go crazy [17:29] yofel: lol [17:29] yofel: just mute it? :P [17:29] closed it [19:10] my rkonq crashes quite frequently.i guess its after i installd flash-plugin. [19:10] how can i confirm? === yofel_ is now known as yofel [21:46] evening