[03:48] apachelogger: ping [04:45] mmm [04:45] yofel: kde-sc-dev-latest still seems a bit broken [04:46] http://paste.kde.org/pa2dc615a/ [04:48] hm rather [04:48] libical-dev is old [04:48] :( [04:55] yofel: I don't think they backport libglu [04:55] atleast the wiki page doesn't say anything [05:56] what does the "pi" mean? "pi nepomuk-core - Nepomuk Semantik Desktop core libraries - transitional package" [06:11] good morning [06:31] which boost version are the precise backports meant to build against? [06:32] currently akonadi has 1.48 while pimlibs build-depends on unversioned (= 1.46) boost [07:30] Riddell: hey, regarding kde-wireless vs kde-wireless2: kde-wireless was my first attempt but I didn't like it. It should be ignored. I am going to trash it [07:30] Riddell: regarding the Spinner: it's a new file which I introduced in r5979 http://bazaar.launchpad.net/~agateau/ubiquity/kde-wireless2/revision/5979 maybe you haven't copied it there? [07:44] markey: pi? I don't see a pi [07:45] Mamarok: can you explain that? [07:45] and Semantik seems to be spelled incorrectly? [07:45] shadeslayer_: that's the listing inf you run aptitude search [07:46] I don't have aptitude here :S [07:46] markey: already explained it to you: nepomuk-core is a transitional package, should not be a hard dependency [07:46] not to me [07:46] nepomuk-core-runtime is the one I presume [07:46] to shadeslayer_ [07:46] ah I think those letters are some sort of status or sth [07:47] that's how transitional packages appear in aptitude search, those are marked pi [07:48] i is probably 'installed' [07:48] v is probably virtual [07:49] shadeslayer_: man aptitude explains those [07:49] but that is not the problem, markey [07:49] righto [07:49] the problem is that our cmake is looking for a package marked as transitional [07:50] which might not be installed, as it is not needed technically [07:50] hm? The package name doesn't dictate CMake files AFAIK [07:51] though difficult to say without actually knowing the real problem :) [07:52] shadeslayer_: I talk about Amarok's cmake [07:53] tbh I'm not even sure why I'm on IRC right now o_o [07:54] Mamarok: and you're just trying to build amarok on kubuntu? [07:54] or well, markey is ? [07:55] shadeslayer_: this was a warning error triggered when markey compiled this morning [07:55] I see [07:57] yep [08:20] \o [08:21] hi Quintasan] [08:26] that should fix kdepimlibs [08:53] Mamarok, markey: please note that our packages cannot know how you name a dependency in cmake (nor do they match the names of cmake config files if there are any). [08:53] so if nepomuk-core is missing, you probably want to install nepomuk-core-dev, for nepomuk-widgets you would want libnepomukwidgets-dev, etc. [08:53] nevermind, it was all a big misunderstanding [08:54] source of the confusion was, I used to have this package installed, and after some upgrades it wasn't installed anymore [08:54] for some very obscure reason [08:54] it wasn't installed here anymore either [08:54] markey uses KDE 4.10.5, I use 4.10.95 [08:55] both Raring [08:55] that's not true [08:55] ubiquity/frontend/kde_ui.py:569: undefined name 'QWidget' [08:55] oh right: We named the package nepomuk-core, debian named it nepomu-core-runtime so it was removed in the last merge and uninstalled. I then later added a transitional package to make it less confusing [08:55] markey: no? [08:55] no# [08:55] told you before [08:55] * Mamarok is confused [08:56] markey: you didn't have nepomuk-core-dev installed anymore, didn't you? [08:56] no [08:56] neither did I [08:56] well yofel just explained it [09:01] shadeslayer_: yeah, libglu was a mistake on my side. (I thought that I fixed all the wrong packages :/ ) [09:17] agateau: I copied spinner to /usr/lib/ubiquity/ubiquity/frontend/kde_components/ [09:17] agateau: it still needed the import changed [09:53] Riddell: oh ok, it is quite possible I only tested it with the test program :/ [09:53] Riddell: how did you change it? [09:53] rather, what did you change the import into? [11:02] Riddell: how does one branch from bzr using an existing local history? [11:13] apachelogger: you mean creating a branch from another local branch? [11:18] agateau: no, branching a remote branch that is derived from another remote branch from which I already have a local branch [11:18] like git remote add [11:18] so it will not download the entire history even though I have like 99% of it already locally [11:20] apachelogger: oh. my bzr foo is not strong enough for that [11:23] ::workspace-bugs:: [1206199] /etc/localtime incorrect link created by dateandtime kcm @ https://bugs.launchpad.net/bugs/1206199 (by Leszek Lesner) [11:43] agateau: I actually fear one cannot do that :/ [11:44] apachelogger: you mean mighty bzr is not as powerful as you thought?! ,) [12:04] agateau: ... :P [12:22] yofel__: fixed startneon5 a bit toget a more sensible session out of it [12:23] oh? what does it run now? [12:34] apachelogger: sorry, changed the import to from ubiquity.frontend.kde_components.Spinner import Spinner [12:34] agateau: sorry, changed the import to from ubiquity.frontend.kde_components.Spinner import Spinner [12:34] agateau: how do I run the test programme again? [12:35] Riddell: as root, go into your ubiquity source dir, then run PYTHONPATH=$PWD python3 ubiquity/frontend/kde_components/nmwidgets.py [12:35] vi !$ [12:36] yofel__, shadeslayer_: did someone retry qtwebkit? Oo [12:37] oups, wrong window :/ [12:38] Riddell: just pushed your fix in [12:39] apachelogger: I retried the one that failed [12:40] yofel__: please don't [12:40] it's completely kaput [12:41] will get stuck forever or fail right away [12:41] I also deactivated the daily build consequently [12:41] no clue why it fails though... well, I do, I just don't know what to do about it :P [12:41] qmake is being recursively invoked when checking for dependencies... recu;rsively as in checking for fontconfig will invoke the check for fontconfig... [12:42] urgh [12:42] new qt builds incoming though [12:43] yofel__: qtwebkit isn't required for anything important right now anyway, so not that important for now [12:43] FWIW, it's exactly the issue I had when building qtwebkit inside the qt5 meta [13:00] apachelogger: this neon5 session isn't running for me [13:00] apachelogger: http://paste.kde.org/p58785a9b/ [13:00] apachelogger: line 16 looks important [13:00] ah [13:00] not published yet [13:00] apachelogger: what's not? [13:00] project-neon5-session [13:01] the previous startneon5 didn't fire up kdeinit5 making it rather not doing anything ^^ [13:02] hm, actually should be published already [13:02] Riddell: make sure you have session 282~saucy1 [13:02] apachelogger: I have 281~saucy1 installed [13:02] right, needs upgrade [13:02] no 282~saucy1 available [13:02] ooh there it is [13:03] just appeared [13:03] ok.. wish me luck [13:04] well something works :) [13:04] weeh [13:06] sweet http://starsky.19inch.net/~jr/tmp/plasma2.png [13:06] presumably it's known that it's not full screen? [13:07] apachelogger: can I blog about this? [13:08] cos it's pure blogworthy [13:08] I'd do it when/if workspace starts building on raring [13:08] triggered a build on possible fix === cmagina_ is now known as cmagina [13:19] agateau: I think "wi-fi" should be "Wi-Fi" in the text [13:20] Riddell: I am using the generic Ubiquity strings there [13:20] Riddell: so it's not really in my control [13:20] agateau: oh fair enough then [13:21] Riddell: but I agree with you :) [13:21] yofel__, shadeslayer_: kdeexamples test build on saucy triggered [13:21] daily builds off for now [13:21] (didn't test locally :P) [13:22] I'll reboot again into a live system see if I can turn on debugging or anything to get it working [13:34] agateau: whole new backtrace now http://paste.kde.org/p482c1fe5/ [13:34] damn [13:35] should really have tested it with the real Ubiquity [13:35] ach who needs testing :) [13:35] Riddell: this is because it fails to load the "process-working" icon [13:35] which is provided by the Oxygen icon theme [13:35] mm, that should have been fixed by your previous fix [13:36] I hand copied the files needed by this branch into my live system, wonder if I missed something for that [13:36] I agree it should have worked [13:36] there's your whole widget relayout branch as well that might be causing confusion [13:37] I haven't been using that branch for this work [13:37] Hi folks [13:43] agateau: ok fixed by your icon theme fix from before, that hasn't made it into this branch http://bazaar.launchpad.net/~ubuntu-installer/ubiquity/trunk/revision/5955#ubiquity/frontend/kde_ui.py [13:43] Riddell: oh right, good catch! [13:43] Riddell: going to merge it in [13:45] agateau: so now I get same as yesterday, it has a gap in the steps list on the left but it skips over it and goes straight to partitioning [13:46] agateau: nothing in syslog or installer/debug about wireless [13:46] setting export UBIQUITY_DEBUG=1 doesn't help [13:46] Riddell: ok so. The gap is because we are missing a translation: I added the text to debian/ubiquity.templates but I am not sure how to get that compiled/installed [13:47] Riddell: the wireless step is skipped if you have network at install time (for example if you are in a vbox or if you have a wired connection) [13:47] Riddell: Ubuntu installer behaves the same [13:48] so you need to have a not-connected wireless card and no wired connection [13:48] you can somehow force the step to show using by passing "--wireless" to ubiquity, but it will behave strangely when the connection has already been established [13:49] Riddell: what are you testing it on? [13:49] agateau: my laptop,live system, alpha 2 [13:50] Riddell: and do you have a wired connection? [13:50] no wireless [13:50] do I need to disconnect that? [13:50] "no wireless" ? [13:51] you need 1) a wireless card, but not already connected to an access point [13:51] 2) no established wired connection [13:51] Riddell: @ [13:51] Riddell: ^ [13:52] sweet,it works! [13:52] \o/ [13:54] agateau: so most important thing, do you want to blog this or shall I? [13:54] Riddell: shouldn't we wait until it's in saucy to blog about it? [13:55] bah, waiting, who's got the patience for that? [13:55] xnox: do you trust me to review the nm.py code split? [13:55] Riddell: feel free to blog about it then, just mention it's young code :) [13:56] Riddell: as long as it merges without conflicts, is pyflakes/pep8 clean, and does crash the installer, that's fine. Added bonus if you did test that gtk-wifi page comes up. Otherwise, I would have tested that with next daily. [13:56] s/does/doesn't/ [13:56] xnox meant: "Riddell: as long as it merges without conflicts, is pyflakes/pep8 clean, and doesn't crash the installer, that's fine. Added bonus if you did test that gtk-wifi page comes up. Otherwise, I would have tested that with next daily." [13:57] damn, crashing the installer is much easier than not crashing it! [13:57] Riddell: just merged trunk in kde-wireless2 to fix the missing icon [13:58] gtk stide [13:59] ahem [13:59] agateau: and the widget refactoring? isn't there a clash in style.qss? [13:59] gtk wifi page still working fine [14:00] Riddell: I fixed the merge conflict before pushing [14:00] Riddell: I am not that clumsy :) [14:41] hmm ubiquity seems to need me to run kdesudo manually now, it didn't a few versions ago [14:46] xnox: is that your autopilot changes? [14:47] Riddell: there is no kdesudo anymore. [14:47] Riddell: pkexec is used on all flavours now. [14:48] cos, policykit fancyness [14:48] ooh policykit fancyness [14:49] * apachelogger notes that this can have unwanted sideeffects as the envrionment is not carried over [14:53] eek [14:53] workspace continues to fail [14:54] naughty workspace [14:58] probably because they can't make up their mind on the branchname [14:59] lol? [15:00] oh there was a fix for dbusmenuqt [15:03] yofel__, shadeslayer_: triggered another workspace build, saucy only in case it fails again [15:03] off for dinner now [15:05] apachelogger: I'll just answer here [15:05] xrandr and xrender can probably be dropped [15:05] mgraesslin: so that's just transitional? [15:06] I guess so, it's probably still linked [15:06] but shouldn't be used [15:06] at least we ported away from xrender [15:06] and the only xrandr code got ifdefed [15:06] xcursor is needed and I don' [15:06] t [15:06] understand why it's missing [15:06] no cmake checks or somesuch? [15:07] it doesn't complain on my system [15:07] yofel__, shadeslayer_: ^ I reckon the build will fail on xcursor missing [15:07] oh well [15:08] I will look into that tomorrow morning [15:08] and fix it [15:08] mgraesslin: any news on that kwin build patch btw.? [15:08] mgraesslin: perhaps you are find_packaging it but don't reflect it via macro_log_feature or what it's called? [15:08] yofel__: I can give you the link for the 4.10 version, but I did not yet rebase to 4.11 [15:08] mgraesslin: would at least explain why there is no indication of it up to the warning that it is set to notfound [15:08] oh that could be - macro_log_feature got dropped [15:09] well, something replaced it judging from the output of cmake, I just do not know what ^^ [15:09] (reminds me that I need to redo the entire cmake for phonon5 :@) [15:09] mgraesslin: would that be a lot of work? If not I could give it a try [15:09] mgraesslin: anyway, we'll see where the build gets today and if it fails we can prod it into building tomorrow [15:10] I'm not sure, I think the cmake stuff in base directory changed, so it won't work [15:10] -> dinner [15:10] ah [15:10] yofel__: the patch for 4.10 is http://quickgit.kde.org/?p=clones%2Fkde-workspace%2Fgraesslin%2Fkde-workspace.git&a=commit&h=2f23192fe2658b9ecde005db40b2ec366b3e18d4 [15:11] thanks! [15:11] looking at the patch I doubt it will compile on 4.11 [15:18] agateau: gosh you have code to draw icons with qpainter in here, isn't it simpler just to edit the oxygen icons? [15:19] Riddell: wanted to avoid the mess of having to get new images to install correctly, but you are probably right [15:19] Riddell: I can install real images if you prefer [15:21] agateau: oh it's fine, just surprising [15:27] agateau: yep looks good, I think I'll commit it [15:27] Riddell: have you already merged the nm-split branch in? [15:30] agateau: that's part of the kde-wireless2 branch surely? [15:31] Riddell: it is, indeed [15:31] Riddell: will LP notice? [15:31] moving bits from one file nm.py to a gtk component file makes good sense, gui bits belong better there anyway [15:47] agateau: hmm a couple of possible errors? [15:47] ubiquity/plugins/ubi-wireless.py:21: 'syslog' imported but unused [15:47] ubiquity/plugins/ubi-wireless.py:65: 'nmwidgets' imported but unused [15:48] mmm [15:48] I need to update to Saucy (no pyflakes3 here) [15:48] syslog => can probably be removed [15:48] nmwidgets => looking into it [15:48] ah but you have a comment about needing toimport it [15:48] # NOTE: Import 'nmwidgets' even though it's not used in this function [15:48] # as importing it as the side effect of registering [15:48] # NetworkManagerWidget which we DO use in the Wireless step UI. [15:49] that is in the gtk code, right? [15:49] yeah [15:50] Riddell: this import was there before but was importing something else [15:50] it makes the ./tests/run-pyflakes test fail [15:51] oh before it was [15:51] - # NOTE: Import 'nm' even though it's not used in this function as [15:51] yes [15:51] I wonder how this got through pyflakes [15:52] maybe that's because it imports "misc" on the same line? or was it whitelisted in the test? [15:52] nm is used elsewhere in the file [15:52] I don't see a whitelist [15:53] ah that's why pyflakes does not complain [15:53] it's not smart enough [15:54] a simple and clean fix would be to *not* register things at import time (which is bad practice anyway) and add an explicit register() function, which would be called from the code importing nmwidgets [15:54] Riddell: we had a guard / whitelist for run-pyflakes, might need adjustment. [15:55] Riddell: agateau: adjust ./tests/pyflakes.exclude [15:55] xnox: ah that's the one [15:56] xnox: right, thanks [15:56] pushing the change in 10 secs [15:57] Riddell: pushed [15:59] agateau: this however is just unexcusable... [15:59] ubiquity/frontend/gtk_components/nmwidgets.py:12:1: E302 expected 2 blank lines, found 1 [16:00] damn [16:00] Riddell: mmm, I have no blank lines there [16:01] agateau: no me neither and now I run it again the issue disappears [16:01] and run-pep8 is happy there [16:01] Riddell: reliable tooling :) [16:01] maybe it was me who made the unexcusable error in my testing and it get tidied up when I merged with yours [16:02] * Riddell hangs head in shame [16:08] thou shall not mess with pep8! [16:12] agateau: this looks more problematic http://paste.kde.org/pf5322790/ [16:13] oww [16:13] Riddell: did not notice there was unittests for these classes [16:15] agateau: able to tidy those up? [16:16] Riddell: looking into it [16:16] Riddell: if I don't get it done today, I'll be working on it tomorrow [16:16] groovy [16:26] Riddell: agateau: why do i have email "2 revisions removed from brunch lp:ubiquity" ? [16:26] is it ok, or did you push override my upload =) [16:26] xnox: I have no idea [16:27] ok. [16:27] xnox: I uncommitted [16:27] because of the test failures [16:27] Riddell: ack. [16:43] yofel: since you know about the issue, can you have a look at what's causing the weirdness wrt nepomuk and db? [16:46] uhhh [16:46] this build log looks weird [16:57] shadeslayer_: where? [16:58] trello [16:58] that thing where nepomuk doesn't start [16:58] because it can't start the db, or connect to it or sth [17:02] shadeslayer_: there's 2 issues I can think of: [17:02] 1) people using 3rd party virtuoso (or something coming from debian experimental) leading to a lib lookup issue like we had in neon [17:02] 2) system crashes or something like that corrupting soprano-virtuoso.trx which requires a manual recovery [17:03] yofel: regarding 1, who even uses a 3rd part virtuoso o_o [17:03] vHanda wanted to look at 2) at some point because it happened to me during akademy, not sure what came out of it [17:03] ack [17:03] shadeslayer_: no idea, that's why I'm not too worried about that (but maybe extending the RPATH with the multiarch dir might be a good idea) === yofel changed the topic of #kubuntu-devel to: Kubuntu - Friendly Computing | https://trello.com/kubuntu | https://notes.kde.org/p/kubuntu-ninjas 4.10.97 saucy/archive raring/beta quantal/todo precise/ninjas | 13.10 Alpha 2 testing | 13.10 milestoned bugs tagged Kubuntu http://goo.gl/vHRjj [17:29] kde status cleanup [17:49] Riddell: :) http://cli-apps.org/CONTENT/content-pre1/159751-1.jpeg [17:49] apachelogger: qtwebkit still OOM'ing? [17:49] stupid thing still hangs on fontconfig [17:54] yes [17:54] well, it doesn't oom [17:54] it forkbombs [17:55] * yofel tries a build [17:57] apachelogger: why does it need that syncqt.pl call? [18:01] header creation [18:06] the question is, is that the recommended way of building it [18:06] since there seem to be atleast 3 build systems [18:06] -.- [18:06] the archive package uses qmkae [18:06] and the README is empty [18:06] and the other 2 systems don't work for me [18:06] !find synqt.pl [18:06] Package/file synqt.pl does not exist in raring [18:06] !find synqt.pl saucy [18:07] Package/file synqt.pl does not exist in saucy [18:07] where be that file [18:07] project-neon-qt5 [18:07] I see [18:07] project-neon5-qt5 actually [18:09] wtf, 'qmake' forkbombs. 'mkdir build; cd build; qmake ..' doesn't [18:10] 0.o [18:27] yofel: stops at Checking for fontconfig for me [18:28] it did finish after a while here [18:28] ah yes [18:29] yofel: so, IIRC there's an option to specify the build dir right? [18:29] -B I think [18:30] let's try using that :P [18:33] why does this not happen to the archive package... [18:33] because, magic === Blizzzek is now known as Blizzz [18:41] uh... and why does this not happen in git?!? [18:45] :O [18:45] indeed [19:13] yofel: maybe something the mk files do? [19:14] well... [19:14] this doesn't happen if I create the package from git [19:14] this doesn't happen if I create the package from bzr [19:14] now I'm trying to run dailydeb by hand to see what happens [19:14] oh [19:41] lolwhat [19:42] when building qtwebkit built by bzr dailydeb it forkbombs @_@ === claydoh_ is now known as claydoh [20:50] yofel: o_o [20:50] what about using apachelogger's custom build scripts [20:51] well, I'm wondering if the foldername with the complex package version is the issue [20:51] need to try that [20:53] why would that be a factor 0.o [20:53] unless qmake trips on the path and goes bezer [20:53] *bezerk [20:53] ( like it did with the space in my folder name in the Sailfish SDK ) === shadeslayer_ is now known as shadeslayer [21:01] Are we able to go without a debian import for terminus? I've made a .deb for the latest version, which allows it to be run on kubuntu again [21:02] Noskcaj: what's terminus? [21:02] Riddell, a font, that breaks bits of kubuntu. bug 812134 [21:02] bug 812134 in xfonts-terminus (Ubuntu) "Konsole + Terminus font, wrong line alignment (Update to terminus 4.38)" [Low,In progress] https://launchpad.net/bugs/812134 [21:02] mm, why would we want it then? [21:04] shadeslayer: yeah, seesm to be the folder name O.O [21:04] Riddell, because i just fixed it, and it's a very popular font. [21:04] try to rename the git clone to project-neon5-qtwebkit-0.0+git20130726+r164~84d199a+neon6~test1, then run qmake in there [21:05] O_O [21:05] yofel: okay yeah, that's just freaking broken [21:05] are you saying it trips on saucy1? [21:06] it trips on *something*, I'm renaming now [21:06] yofel: btw http://bcache.evilpiepirate.org/ [21:06] Noskcaj: aah [21:06] might be beneficial for you :) [21:06] Noskcaj: such we can get that in, is it on the bug? [21:06] seen the initial post when it was merged, not sure what to do with it [21:07] could be useful or virtualbox I guess [21:07] Riddell, I'll make the .dsc now [21:07] *for [21:10] shadeslayer: seems like '+' is the fuse [21:10] project-neon5-qtwebkit-0.0+ BOOM, project-neon5-qtwebkit-0.0 WORKS [21:10] (WTF) [21:11] apachelogger: ^ [21:11] what the flying fuck [21:15] I changed the version string to 0.0.git{date}.r{revno}~{git-commit}.neon{revno:packaging} (unless someone has a better idea) [21:16] fine with me [21:16] actually [21:16] yofel: you could change the build for Qt [21:16] * shadeslayer does that himself [21:16] what do you mean? [21:16] well [21:16] currently the qt build disables mm and webkit [21:17] we could change the version on qt and try building with mm and webkit] [21:17] I would rather have qtwebkit seperate... [21:17] why? :D [21:17] takes too long to build would be one reason [21:17] okay [21:18] lemme try mm then [21:26] Riddell, I've added the files from the new debian version. The debian maintainer hasn't gotten to it yet, and it would be nice to fix this by 13.10 [22:14] yofel: yep, mm compiles [22:14] when you set the version without the + signs [22:56]