=== jono is now known as Guest3821 [07:06] good morning [09:58] yofel: pingly [09:59] yofel: http://paste.kde.org/776720/ [09:59] what I ran : bzr-buildpackage-ppa -d saucy -y 13.10 [09:59] oh right [09:59] nvm [10:18] Good evening. === vinay is now known as Guest5274 [10:19] How's kde-workspace? [10:21] Erm, nobody fixed it for me? [10:21] :O [10:21] I thought I left it with someone able to takeover... [10:22] debian/tmp/usr/share/kde4/services/searchproviders/freshmeat.desktop [10:22] why that looks like a fun search provider [10:22] Nice:) [10:22] Some of my exam results came out today. Most crapped is Chinese:( [10:23] Best subjects are chemistry , biology and computer [10:24] congrats :) [10:24] atleast you passed :P [10:24] shadeslayer, eh hum it's still Form 3, I pass everything (I do have a good result at school) [10:25] smartboyhw: IMHO passing is all that matters :P [10:25] because even that was super hard at my uni [10:25] shadeslayer, in Hong Kong, passing is just s***t [10:25] maybe you should move to Austria [10:25] where they complete their Bachelors in 12 years [10:25] shadeslayer, :O [10:25] 12 years for a single bachelor degree!? [10:26] ask apachelogger [10:26] smartboyhw: I had no access to the machine and it's an expensive machine that so I had to turn it off [10:26] Riddell, oh dear.... [10:26] Can someone who has a fast machine build that? [10:26] smartboyhw: ok I'll take a look [10:26] It takes SO long for me to build kde-workspace.. [10:26] Riddell, thx:) [10:26] it is a beast with its double compilation [10:32] dpkg-source: warning: unknown information field 'Kde-Window-Manager-Creplaces' in input data in package's section of control info file [10:32] smartboyhw: I don't suppose you worked out what that random warning is? [10:32] I couldn't work it out at all last time i looked [10:32] Riddell, I dunno what the hell is that... probably the extra C makes the issue... [10:32] I never heard of that thing before though. [10:34] having 4 builds of kwin is a bit nuts, I'm really not convinced that makes sense [10:34] Hmm, we got a problem in kmajohgg [10:34] and I remember martin saying we could get rid of one at least [10:34] smartboyhw: what's up? [10:34] Riddell: why do you have 4 builds? [10:34] oh hello :) [10:34] Riddell, python-twisted depends on 7 packages but they can't be installed... [10:34] https://i142731151.restricted.launchpadlibrarian.net/142731151/buildlog_ubuntu-saucy-i386.kajongg_4%3A4.10.80-0ubuntu1~ubuntu13.10~ppa2_FAILEDTOBUILD.txt.gz?token=c89fa49fefe5edabc1bf8d42de71e186 [10:34] mgraesslin: debian/control:Package: kde-window-manager [10:34] debian/control:Package: kde-window-manager-gles [10:34] debian/control:Package: kde-window-manager-active [10:34] debian/control:Package: kde-window-manager-active-gles [10:35] we have 4 packages, but it only builds twice [10:35] good [10:35] once for desktop, once for active [10:35] just wanted to point out that it only needs two build [10:35] right, once with -DKWIN_PLASMA_ACTIVE=true and once without [10:36] for active the gles package does not make sense [10:36] active should always be gles [10:36] because it's always gles? [10:36] right [10:36] makes sense, lets drop that then [10:36] for desktop you could merge kde-window-manager and kde-window-manager-gles [10:37] I think that's from the time when the build system was not able to build both binaries at the same time [10:37] mgraesslin: merge as in put in the same binary package? [10:37] yes [10:37] it would then have kwin and kwin_gles in one package [10:37] that reminds me: I need to send a mail to packagers about the Wayland dependencies [10:38] we've added libwayland-dev to the kde-workspace build-depends [10:38] anything else needed? [10:38] runtime dependencies [10:39] without Weston it doesn't make much sense [10:39] mm [10:39] so I would make that a recommends [10:39] but do we want weston installed for our default X only setup [10:40] mgraesslin: if kwin and kwin_gles are in the same binary how does it choose between them (in startkde presumably)? [10:40] like before: you have to set the env variable in case you want kwin_gles [10:40] right [10:40] http://userbase.kde.org/Desktop_Effects_Performance#kwin_gles [10:41] smartboyhw: instead of building the whole of kde-workspace twice I think it'll be no problem to just build kwin twice [10:42] Riddell, OK. [10:42] yes that works [10:42] KWin can be built without kde-workspace [10:42] For now, let me fix some analitza red lights. [10:42] * mgraesslin uses that in Jenkins [10:42] Missing symbols (again) [10:42] smartboyhw: analitza needs SO version bumping [10:42] shadeslayer, alright. [10:42] apol hasn't done that :/ [10:43] meh:P [10:43] So? [10:43] hm? [10:43] 2 public symbols were changed [10:43] breaking ABI [10:43] so they need to bump the SO version [10:43] Alright, kdegraphics-mobipocket then. [10:44] Coding error it seems.... [10:44] smartboyhw: there was something in https://notes.kde.org/p/kubuntu-ninjas about analitza [10:44] Riddell, just read it:) [10:44] yeah the so version [10:44] Looks the kdegraphics-mobipocket got a completely broke generator_mobi.cpp [10:45] huh [10:45] GENERIC_LIB_SOVERSION [10:45] set_target_properties(analitzagui PROPERTIES VERSION ${GENERIC_LIB_VERSION} SOVERSION ${GENERIC_LIB_SOVERSION} ) [10:47] set by kdelibs I think [10:47] oh [10:47] means he needs to start setting it manually [10:47] this is the first time I've encountered GENERIC_LIB_SOVERSION [10:47] I guess I'll email Aleix so that he remembers to do this [10:49] Riddell: if you want to pass something on about this Mir/KWin thread: 11 files changed, 1588 insertions(+), 1020 deletions(-) [10:50] that's the difference I created since yesterday in the Wayland backend - the file Thomas looked at [10:51] plus uncommitted changes: 5 files changed, 108 insertions(+), 5 deletions(-) [10:51] it's kind of insane to try to do any adoption of KWin right now [10:52] but we would love to see the Wayland help we got promised years ago ;-) [10:53] Riddell, shadeslayer confirmed: kdegraphics-mobipocket failed due to coding error. [10:53] oh [10:53] pastebin error? [10:53] tsdgeos: ^^ [10:53] you just move so fast mgraesslin all anyone else can see is a blur! [10:53] http://paste.kde.org/776810/ [10:53] smartboyhw: could be a gcc 4.8 issue [10:54] shadeslayer, look at the paste and determine:) [10:54] smartboyhw: places to check for a fix would be upstream git and if other distros have managed to compile it [10:55] also check if 4.10.4 compiles or not in saucy [10:56] Riddell, second sentence confirmed it does. [10:56] Riddell: that's the reason why the Wayland backend is marked as experimental in 4.11 - it's just a random development snapshot which works [10:56] in saucy [10:56] (works with limitations, e.g. gles doesn't work) [10:57] Riddell, last commit is @ 6/6/2013, aligning to okular patch. [10:57] I'm starting to think that since okular didn't build then it can't build also. [10:58] 4.10.80 I mean. [10:58] Riddell: did I get correct that the discussion is mainly about how the stack will look like in 14.04? [10:58] okular can't build due to plasma-active-dev:O [10:58] kde-runtime is the prob I think... [11:01] mgraesslin: only timescale I've seen (I've not looked hard) is robert ancell saying "We're aiming to be able to preview XMir in 13.10. We're doing the work right now to integrate Unity 7 with XMir. [11:01] smartboyhw: oh wouldn't okular be a dep of that? [11:02] Riddell, yep. Okular (>=4.10.80) should be the dep. But it failed to build also... [11:02] mgraesslin: so should we build against libwayland or just ignore it? [11:03] Riddell, speaking of it, you have a nice launchpad karma: 88888 [11:04] 8 is good luck in china isn't it? :) [11:04] smartboyhw: did you get this error?..-- DebianABIManager: unable to find CMake target 'solidcontrol' for package 'libsolidcontrol4abi2'. Please set X-CMake-Target [11:04] -- DebianABIManager: unable to find CMake target 'solidcontrolifaces' for package 'libsolidcontrolifaces4abi2'. Please set X-CMake-Target [11:04] Riddell, well, you have to remove these two packages. [11:04] It's KOed. [11:04] Riddell, yeah, 8 is gd luck in China [11:04] * smartboyhw has painfully discovered that:P [11:05] I mean the build thing. [11:05] what's there to remove? [11:06] yofel, the solid::control. [11:06] ah right [11:06] Check yesterday's backlog. [11:06] Riddell: I highly recommend to build against libwayland [11:07] that's why I need to send a mail to packagers [11:07] Riddell: yeah, that got remove in 4.11, it was used by networkmanagment it seem and was killed with it [11:07] *removed [11:07] Riddell: ok, let's consider 14.04 - you will be shipping KWin 4.11 [11:07] Riddell: anything else won't make sense [11:07] Riddell: we had already the dependency freeze for 4.11 and that does not include Mir - neither with an XMir nor directly [11:08] we will not support any setup which is not X11 based in the long term support version of KWin 4.11 [11:08] so if 14.04 would have a setup with XMir replacing the normal X Server we would not be able to support it [11:08] yofel: btw where have you traditionally put the qmldir ? in a install file or a .dirs file [11:09] this means we would just have to close all bug reports of Kubuntu users [11:09] have -./usr/lib/kde4/imports/org/kde/dirmodel/qmldir [11:09] mgraesslin: mm in kubuntu 13.10 we will be shipping kwin/kde sc 4.11 [11:09] shadeslayer: uh, install file [11:09] mgraesslin: 14.04 would be kde 4.12 no? [11:09] any reason not to put it in a .dirs file? [11:09] shadeslayer: does debian do it differently? [11:09] Riddell: there is no 4.12 of KWin [11:09] no idea, I haven't checked [11:09] there will be a KDE SC 4.12 [11:09] but not with kde-workspaces [11:10] yofel: any recommendations on where would be a good place to look? [11:10] mgraesslin: ah right [11:11] getting back to my point: with an XMir in between we have no chance to figure out whether a problem is a general problem or a Ubuntu specific problem [11:11] and we would not have the manpower to carry such an overhead [11:11] all bugs would be RESOLVED DOWNSTREAM [11:11] mgraesslin: right, I hate the very idea of running kwin on x on mir, I'd be very reluctant to do that for the reasons you say [11:11] shadeslayer: hm, any of the merged packages that could have a qmldir [11:11] I don't think that this would look good for the distribution [11:12] right, like? [11:12] when I asked if they would run unity on mir on wayland I got a "nope" which is pretty telling [11:12] we have like a bazillion packages [11:12] lets see, a game that was ported to QML maybe [11:13] debian/plasma-widgets-workspace.install:usr/share/kde4/apps/plasma/plasmoids/org.kde.notifications/contents/ui/NotificationDelegate/qmldir [11:13] debian/plasma-widgets-workspace.install:usr/share/kde4/apps/plasma/plasmoids/org.kde.notifications/platformcontents/touch/ui/NotificationDelegate/qmldir [11:13] install file it is [11:15] mgraesslin: so my current thinking is we expect kubuntu 14.04LTS to have KWin on X and for 14.10 we can see if switching to wayland is ready using packages synced from debian [11:15] yes [11:15] which will kill any ability to run ubuntu desktop and kubuntu switching at login but oh well [11:15] but you should be prepared that also in the Wayland world you might need to provide X [11:15] if NVIDIA and Catalyst do not support Wayland at that time you are forced to provide X [11:16] one worry is for the live CD stuff, that ubiquity-dm is faffy [11:16] so from looking at Robert's mail - none of the possibilities a to g suit us [11:17] that is upstream will support: KWin on X [11:17] and KWin on Wayland [11:17] anything else can be done downstream [11:17] but at least without any support from me [11:17] maybe someone else will help [11:17] but to my knowledge no KWin developer is using Ubuntu [11:18] The question is: Who will help? [11:18] smartboyhw: we have offers from e.g. thomas voss but I really don't want to take them up [11:18] there's no way he can support kwin on mir part time [11:18] and do a good enough job [11:18] unlikely - quite true [11:18] I just don't believe it's possible [11:18] ^^ [11:18] especially without any help from upstream [11:19] and we don't make it easy to fork kwin - we have a huge churn [11:19] plus with the amount of code churn that mgraesslin says happens in KWin, it'll be impossible to do it part time [11:19] and that would be the mother of all distro only patches which we have a policy against [11:19] will require full time work [11:19] (for adding the Wayland backend I changed our internal interfaces) [11:19] also some assumptions in the thread are wrong [11:20] I do not plan to have KWin with a pluggable windowing system interface [11:20] KWin is a Wayland compositor [11:20] just currently only supporting X11 [11:20] that has been my planning for a few years now [11:20] I don't change that just for the fun of it :-) [11:22] a wayland compositor only supporting x11? I'm not sure I understand that [11:22] I consider Wayland as our primary windowing system [11:22] we just don't support it yet [11:24] I can't wait to try out wayland stuff in 4.11 :P [11:25] smartboyhw: you can install weston now and run it on X, works fine [11:25] shadeslayer: rather [11:25] I was able to use it two days in full production, then my Intel card froze [11:26] http://wstaw.org/m/2013/06/18/plasma-desktopt14572.png - Weston with KWin as client and XRender compositing backend [11:27] mgraesslin: btw any ideas why this would happen on a multi monitor setup? http://i.imgur.com/yALluEB.png [11:27] when using cover switch [11:27] yes [11:28] oh?\ [11:28] * mgraesslin cannot do math [11:28] I see :D [11:28] it works fine on the bigger monitor [11:29] mgraesslin: bug already reported? [11:29] or should I report one for 4.11 if I can reproduce it there [11:29] of course it is reported [11:29] years ago [11:29] hah [11:29] https://bugs.kde.org/show_bug.cgi?id=184874 [11:30] KDE bug 184874 in effects-tabbox "'Cover switch' view is cut off at the left/right edge on a low-res screen" [Normal,Confirmed] [11:30] 2009 - nice [11:30] KDE 4.2.0 @_@ [11:31] my Math skills got worse since then [11:31] my bug is slightly different btw, windows are fine, but the black gradient is misplaced [11:31] no the windows are wrong [11:31] oh [11:32] they should be at the edge of the screen [11:32] ah, and I thought the windows were fine but the gradient was misplaced [11:32] ( which looks more likely, since the windows IMHO appear to be placed correctly ) [11:34] you know better obviously :) [11:34] yeah I know that I lost all my math skills :-( [11:35] hehe [11:36] one of the reasons why I don't want to write effects anymore [11:53] so hot again -.- === greyback is now known as greyback|lunch [11:55] apachelogger baby you're always hot [11:56] Riddell, :O..... [11:57] lol [11:57] here it's not as hot as yesterday, + the fridge isn't empty ^^ [11:57] It's very hot here. [11:57] 29C [11:57] Tomorrow 33C [11:58] well, I am totally going to spend the day at the pool, see you in the evening :P [11:59] yesterday it was 36°C here, today it's 32 [12:08] well .. good thing I didn't go for my vacation [12:08] would have been swept away by all the water due to cloud bursts and rainfall [12:09] shadeslayer, :) [12:09] Riddell: Netflix doesn't have Dr Who S5 [12:09] OTOH it's 31 degrees here, but 80% humidity [12:10] do you at least have a fan? I don't -.- [12:11] oh you wimps, heat is a feature of living in india and continental europe, if you want cloud and cold come to scotland [12:12] shadeslayer: depending on your numbering system that might be 40 years old [12:13] Riddell: the one that features Matt Smith playing the doctor for the first time [12:14] fishfingers and custard! [12:16] heh [12:19] JontheEchidna: I get Muon crashing when trying to use Firefox installed [12:19] Installer* [12:19] logs coming soon [12:21] JontheEchidna: http://paste.ubuntu.com/5776939 [12:22] Quintasan: looks like bug https://bugs.kde.org/show_bug.cgi?id=320230 [12:22] KDE bug 320230 in installer "Muon Software Center Crash" [Crash,Unconfirmed] === greyback|lunch is now known as greyback [12:23] JontheEchidna: Quite possible, want me to add the backtrace there? [12:23] Quintasan: nah, that's fine [12:27] yofel: kcm_pimactivity seems like it could go in kontact? [12:41] shadeslayer: possibly. I pondered libkdepim4 as that has plugin files too and should always be there. (you can use kmail without kontact) [12:47] possibly [12:49] yofel: similar thing with usr/bin/asapcat [12:49] which is a developer tool [12:50] but I just put it in akonadi-server for now [12:50] Hiyas all [12:52] bbl [12:59] shadeslayer: may be that you're building against an old okular, but hey let's blame people coding! [13:18] tsdgeos: well, technincally it is a upstream issue because someone didn't bump the minimum requirement on Okular in CMakeLists.txt [13:19] shadeslayer: you know where the repo is [13:19] *nod* === francisco is now known as Guest7324 [13:30] tsdgeos: review plz http://paste.kde.org/777050 [13:32] shadeslayer: don't we have a versioned build-depends on okular in that package? [13:33] it wasn't part of kdesc-dev-latest-saucy.txt though, I just added it [13:34] yup okular-dev (>= 4:4.10.80), [13:34] ah okay [13:36] then why the compile error? [13:36] tsdgeos: that version was recently changed [13:36] ah [13:36] Riddell: but kdegraphics-mobipocket should complain if it finds a old okular imho [13:37] shadeslayer: if the aptch works, go on [13:37] shadeslayer, that's the prob I'm having here:P [13:38] tsdgeos: it works in the sense that cmake can't find okular 4.10.80 on a 4.10.4 system :) [13:38] The first thing though: Someone repair okular:P [13:38] good then [13:38] * shadeslayer nominates smartboyhw for that [13:38] shadeslayer, it involves kde-runtime. [13:38] I uploaded that [13:39] shadeslayer, hmm but it also involes plasma-active... [13:39] huh? [13:39] -- DebianABIManager: CMake target 'okularcore' SOVERSION does not match package name 'libokularcore2abi1' [13:39] ^^ [13:39] that looks pretty straight forward [13:39] how does that involve plasma active [13:40] yofel, shadeslayer https://i142414905.restricted.launchpadlibrarian.net/142414905/buildlog_ubuntu-saucy-amd64.okular_4%3A4.10.80-0ubuntu1~ubuntu13.10~ppa2_FAILEDTOBUILD.txt.gz?token=822c3ac9ce5ef35e6ee788890e39fda1 [13:40] your token is showing [13:40] Somebody added it as a strict build-dep. [13:40] oh ok, archive skew [13:41] all this OptiPNG does add to the build time a lot [13:41] yofel: btw thanks for mentioning hp systray the other day [13:41] yofel: fixed the issue where printing PDF's would cut off the top on my printer because page scaling wasn't happening [13:41] shadeslayer: that turned out to be actually useful? ^^ [13:41] aah [13:41] need to check if I can do the same here [13:42] Riddell, you didn't fix calligra for me:( [13:42] smartboyhw: it's on he todo! [13:42] yeah, I couldn't for the life of me figure out how to make it print PDF's properly [13:42] Riddell, he!? [13:42] okular cuts off parts when I have the duplexer in the printer but I set it to single-sided [13:42] *the you mean? [13:42] smartboyhw: it's on the todo! [13:42] chrome / okular both cut them off [13:42] Riddell, :) [13:42] great:) [13:42] yofel: dunno, cuts it off on single sided as well [13:43] yeah, that's why I used evince lately when I needed properly printed PDF's :( [13:43] so I had to copy the PDF's over to a Windows machine and printed using Adobe thingamajing [13:43] *thingamajig [13:43] lol, evince didn't work? [13:43] wait, we have the adobe thing for linux too [13:44] we do? [13:44] also, evnice, I prefer not to install gtk apps, and the desktop was already on the network, was a simple copy and print [13:44] acroread is in partner [13:45] but yeah, the systray thing has a option to scale documents [13:45] 9.5.5 at least [13:45] worked for the 5 pages I printed today [13:45] * smartboyhw is expecting ninjas PPA build failure messages:P [13:47] Now it makes sense [13:47] Unpacking plasma-active (from .../plasma-active_3.0-0ubuntu1_amd64.deb) ... [13:47] dpkg: error processing /var/cache/apt/archives/plasma-active_3.0-0ubuntu1_amd64.deb (--unpack): [13:47] trying to overwrite '/usr/lib/kde4/imports/org/kde/dirmodel/qmldir', which is also in package kde-runtime 4:4.10.80-0ubuntu1~ubuntu13.10~ppa1 [13:47] dpkg-deb: error: subprocess paste was killed by signal (Broken pipe) [13:48] shadeslayer, fix fix please:) [13:48] :< [13:48] I think that's the qml move that was done [13:48] Yes it is:) [13:48] :P [14:00] shadeslayer, how to determine okular's abi version? [14:00] yofel, ^ [14:00] * smartboyhw can't determine... [14:00] you'll find it in some CMakeList.txt [14:01] * yofel checks [14:01] hmmm [14:01] how do we handle this [14:01] set_target_properties(okularcore PROPERTIES VERSION 3.0.0 SOVERSION 3 ) [14:01] smartboyhw: ^ [14:01] because dirmodel is shipped in 3.0 [14:01] and will be removed in 4.0 [14:02] so do I just break it against 4.0? [14:02] Argh so it is abi3 now is it? [14:02] it's libokularcore3 [14:02] X-Debian-Abi: 0 [14:02] i.e Breaks: plasma-active (<< 4.0-0ubuntu1) [14:03] how about removing the files from active and depending on kde-runtime >= 4.10.80 ? [14:05] but then is that compatible with what PA is using? [14:05] good question [14:05] this has to be backportable too... [14:05] well, breaking active is probably ok ^^ [14:06] hah [14:06] I'll ask in #active [14:07] what is the installgen file for in debian/ ? [14:07] !find usr/lib/kde4/imports/org/kde/dirmodel/libdirmodelplugin.so [14:07] !find usr/lib/kde4/imports/org/kde/dirmodel/libdirmodelplugin.so saucy [14:07] File usr/lib/kde4/imports/org/kde/dirmodel/libdirmodelplugin.so found in plasma-active [14:08] last I heard about installgen was that it was something from before my time [14:08] YAY IT BUILDS!!!!!!!!!! [14:08] ^^ [14:08] * smartboyhw then waits for a lot of symbols issues;P [14:09] yofel: it's compatible with PA3, exact same code [14:10] so lets drop it from PA3 [14:10] heh, typical [14:10] second try. [14:12] hmm, installgen.. [14:12] "It is a system used to automatically update the .install files. The [14:12] installgen files in kde-sc packaging however have not been updated for [14:12] a long time and we do not actively use them anymore. You can ignore [14:12] them." [14:12] yofel: Breaks/Replaces : plasma-active ( << 3.0-0ubuntu1) [14:12] * Riddell ignores them [14:12] sound fine? [14:12] not 4? [14:12] why 4? [14:13] I'll drop them from 3.0-0ubuntu2 [14:13] << 3.0-0ubuntu2 then [14:13] ah right [14:13] beause 1 has the files as well [14:16] wonder why the Windows 7 L2TP/IPsec vpn connection is so much faster than the Linux equivalent ...must be missing an important setting , but haven't been able to track it down [14:17] shadeslayer, yofel: What's the final decision for the control file of okular? [14:17] * smartboyhw will upload it soon.... [14:18] Probably in an hour [14:18] * smartboyhw still has to fix the HUGE number of symbols. [14:18] uh, as I said [14:18] it's libokularcore3 [14:18] X-Debian-Abi: 0 [14:18] yofel, I mean the build-deps [14:18] active, etc. etc.... [14:18] wait for shadeslayer [14:19] yofel, OK:) [14:19] well [14:19] I'll have to upload plasma-active to ninjas [14:19] and the archive [14:19] so that you can build without issues [14:19] shadeslayer, OK. [14:19] or maybe not [14:19] !? [14:19] if you have -proposed enabled, then it'll work [14:20] ninjas depends on proposed so you shouldn't need to [14:20] yeah, fine [14:20] britney won't auto migrate because p-a will depend on kde-runtime > 4.10.80 [14:20] and there is no 4.10.80 in the archive :) [14:20] so you'll have to run -proposed [14:20] shadeslayer, I don't since it messes up calligra:( [14:21] fun, the PA install file has a bunch of free space at the top [14:21] for no reason [14:21] hmm, should I package kdenetwork-strigi-analyzers? [14:21] it's in the 4.10.80 release but it didn't build in old kdenetworks [14:22] in neon it builds: [14:22] /opt/project-neon/lib/strigi/strigiea_tiff.so [14:22] /opt/project-neon/lib/strigi/strigiea_dvi.so [14:22] no idea if something uses that [14:23] For symbols files, should I change it to 4:4.10.80 only? [14:24] smartboyhw: where, what, why? [14:24] yofel, symbols [14:24] yeah, but what exactly is the context? [14:24] What: All the symbols like _ZNK60kular10Annotation5flagsEv@ABI_3_0 [14:24] After that it is originally 4.9.80 [14:24] oh okular, yeah, you could simply re-create the file from scratch [14:25] LOL [14:25] smartboyhw: see http://pkg-kde.alioth.debian.org/symbolfiles.html [14:25] or you change the lib name and version in the header, build and use batchpatch [14:25] smartboyhw: quick review : http://paste.kde.org/777146/ [14:25] that's what I usually do [14:26] erm [14:26] yofel: ^^ [14:26] LOL [14:26] It looks fine 'ere though:P [14:27] yofel: that's kdegraphics-strigi-analyzer [14:27] yofel: kdenetwork is /usr/lib/strigi/strigita_torrent_analyzer.so [14:27] uh... right [14:27] anyway, packaged :) [14:27] * yofel should read what he types... [14:28] Continue:) [14:28] :P [14:28] Third build, 73% [14:28] Oops [14:28] shadeslayer: ship it [14:30] ssh! that's the secret debian upload password! [14:30] fourth build with fixed .install files:P [14:30] Riddell, really? [14:30] :P [14:36] - _ZN6Okular10Annotation6Window7setTextERK7QString@ABI_3_0 4:4.10.80 [14:36] +#MISSING: 4:4.10.80-0ubuntu1~ubuntu13.10~ppa3# _ZN6Okular10Annotation6Window7setTextERK7QString@ABI_3_0 4:4.10.80 [14:36] Can somebody explain!? [14:37] that function was removed? [14:37] symbol went missing, which is fine as the SOVERSION changed [14:37] ^^ [14:37] smartboyhw: just remove all MISSING lines from the file [14:38] kde-runtime fixed and uploaded as well [14:38] but as shadeslayer said, it happened because that function was removed: [14:38] $ c++filt _ZN6Okular10Annotation6Window7setTextERK7QString [14:38] Okular::Annotation::Window::setText(QString const&) [14:39] alternatively, the signature changed [14:40] so maybe its called setMagicText now, who knows :P [14:42] * shadeslayer takes up kstars [14:42] Riddell: any progress on the split packages? [14:43] shadeslayer: I'm doing kdenetwork now [14:43] any ideas if I should worry about this? [14:43] E: kde-zeroconf: library-not-linked-against-libc usr/lib/kde4/kio_zeroconf.so [14:43] \o/ [14:43] uhm [14:43] that looks odd [14:44] no sure, check whether it acutally uses libc stuff [14:44] ^^ [14:44] surely everything uses libc? [14:44] I checked one lib from nepomuk, and that had no libc method call from what I could see [14:44] kdeworkspace has the same thing on various bits so it might be some facet of saucy gcc [14:44] what Riddell said [14:45] would everything really use libc? [14:45] I mean: if I write a lib with only one class with one function, that callls one function from Qt, would it link against libc? [14:45] * smartboyhw really can't understand how humans invented symbols, so difficult... [14:45] dunno I assumed it included some fundamentals [14:46] like I think libstdc++.so.6 includes the definition of "new" which is pretty important to c++ [14:46] considering our linking is really conservative lately, I can imagine it not linking against libc if it uses none of its symbols [14:46] hm [14:47] well, doesn't it work like this : Qt app uses some symbols of Qt and Qt itself uses some symbols of libc, hence app is linked to libc [14:47] because the same thing happend with amarok and GStreamer 1.0 [14:48] Amarok was compiled against QtWebkit which used GStreamer 0.10 [14:48] and the pgst port I was trying to test used GStreamer 1.0 [14:48] yeah, it depends where it takes the symbol from. If Qt has all of the symbols it needs, then there's no need to link against libc [14:50] sound reasonable, I'll stop worrying then [14:51] I'm not 100% convinced myself yet.. apachelogger ^ [14:51] yofel: but wouldn't something as basic as say new require libstdc++ ? [14:52] shadeslayer: it does link against libstdc++ [14:52] so that class would definitely need to link against libstdc++ [14:52] just not libc [14:52] oh [14:52] hmm [14:52] the I /suppose/ it's possible [14:53] eh [14:53] well, looking at the source that kioslave is rather small and from what i see pure c++, so this sounds possible [14:53] yofel: what's the question? [14:54] apachelogger: whether something can work without linking aginst libc6 [14:54] E: kde-zeroconf: library-not-linked-against-libc usr/lib/kde4/kio_zeroconf.so [14:54] no [14:54] that's not a library though [14:54] that's a plugin [14:54] :P [14:54] how does that make a difference? [14:56] plugins are loaded after runtime linker resolution such that in theory it needs no libs at all iff all the libs it would need are loaded prior to the plugin loading :P [14:57] anyway [14:57] cmake allows you to link non-recursively [14:57] different example then: [14:57] E: libnepomukcleaner4: library-not-linked-against-libc usr/lib/libnepomukcleaner.so.4.11.0 [14:57] so unless libfoo actually uses libc crap it won't be linked against libc [14:58] so in a qt context you can get around libc, you'll probably not get around libc++ as Riddell already mentioned that operator new is defined in there [14:59] ok, so lets conclude that we can ignore it [14:59] well [14:59] it works, doesn't it? :P [14:59] It is theoretically possible to have a library which doesn't use any symbols from libc, but it is far more likely that this is a violation of the requirement that "shared libraries must be linked against all libraries that they use symbols from in the same way that binaries are". [15:00] from lintian [15:01] I HATE SYMBOLS. [15:01] * smartboyhw faints. [15:02] good, you're starting to understand them :P [15:02] ^^ [15:03] interesting : * Add kubuntu_link_pthread.diff to link to pthread, fixes compile on i386 [15:03] from yofel [15:03] uh.... that's DSO stuff IIRC [15:04] please drop if possible [15:04] lets see if it we still need it? [15:04] shadeslayer: that kstars? [15:04] yes [15:04] though then we might need it for raring still? [15:04] shadeslayer: I've not worked out why we need that, and neither does upstream [15:04] okay [15:05] well, drop it and we'll see. Worst case we can add that back for raring-only or so [15:05] yeah [15:07] shadeslayer: I'm pretty sure we still needed it when I merged kstars this cycle [15:10] hurray [15:10] we can drop the patch [15:10] oh? [15:10] yeah, compiled fine [15:11] on i386 [15:11] ? [15:11] oh ... good point :/ [15:12] * shadeslayer forgot that it had to compiled against i386 [15:17] ready to upload now:) [15:17] shadeslayer, fixed already? [15:17] fixed? [15:17] whats fixed? [15:18] bag [15:18] bah [15:18] /usr/bin/ld: note: 'pthread_create@@GLIBC_2.1' is defined in DSO /lib/i386-linux-gnu/libpthread.so.0 so try adding it to the linker command line [15:18] k will fix post dinner [15:23] shadeslayer, plasma-active? [15:28] Now if okular doesn't build, it will be super-weirdo. [15:34] phew, kde-workspace uploaded [15:38] Good afternoon. [16:01] anyone else unable to suspend from a session in saucy? [16:01] I wonder if that's a consolekit thing [16:01] being replaces by logind [16:03] Riddell: me [16:04] pm-suspend itself works, just not from UI [16:05] well [16:05] Riddell: I couldn't until I installed pm-utils [16:05] doesn't pm-suspend come from pm-utils? [16:05] http://paste.kde.org/777194/ [16:06] yeah [16:06] I can suspend from UI and pm-suspend [16:06] but I didn't have pm-utils installed [16:09] wait what [16:09] libpython2.7 : Depends: libpython2.7-stdlib (= 2.7.5-1ubuntu2) but 2.7.5-5ubuntu1 is to be installed. [16:09] how [16:09] saucy has 2.7.5-5ubuntu1 [16:11] can't be archive skew, Uploaded by: Matthias Klose on 2013-06-05 [16:18] yofel: any ideas why this happens? [16:20] not really, unless you're using an out-of-date chroot [16:25] hm well [16:25] I ran --update [16:27] yofel: http://paste.kde.org/777212/ [16:27] ahh [16:27] marble [16:58] hello people [16:58] new symbols galore in marble [17:00] hm [17:00] there's a usr/bin/marble-qt [17:00] yofel: ^ thoughts? [17:10] shadeslayer: make a marble-qt package (As I guess that's the same thing as quassel-qt4?) [17:10] that's all fine and dandy ... but ... [17:10] there is no other library or anything to go with it [17:10] just marble-qt [17:10] which is the perplexing part to me [17:11] well, it doesn't depend on kde I guess? [17:11] and it has no shared libs? [17:11] libmarblewidget itself doesn't either [17:11] hm [17:11] So I can see a purpose for it [17:12] you'll just have to make sure it works without 'marble' installed [17:12] in a non kde env [17:13] yeah. Using marble in unity would be cool :P [17:13] hehe [17:13] you know ... I've never tried out Unity [17:13] like, never ever [17:13] I've only seen screenshots [17:13] used it a few times. [17:14] it *is* usable, it just tends to annoy me because it gets in my way at a few places and I can't tell it not to [17:14] the look and feel though is nice [17:15] interesting : http://www.ubuntu.com/phone/carrier-advisory-group [17:15] indeed [17:15] can only mean one thing tbh :P [17:17] well, if they ever do make a really usable ubuntu phone I might even buy it ^^ [17:17] why this is odd [17:17] dh_install: cp -a debian/tmp/usr/lib/kde4/plugins/marble/15/QtMobilityPositionProviderPlugin.so debian/marble-plugins//usr/lib/kde4/plugins/marble/15/ returned exit code 1 [17:17] except there is no QtMobilityPositionProviderPlugin.so in the install files [17:17] anywhere [17:17] oh [17:17] debian/marble-plugins.install.linux:usr/lib/kde4/plugins/marble/15/QtMobilityPositionProviderPlugin.so [17:18] what's install.linux ? 0.o [17:18] debian has !linux stuff [17:21] * yofel on the road, back in an hour [17:23] * shadeslayer will finish marble and be done for the day [17:40] shadeslayer: Unity is ok and usable, it is just Compiz that is holding it back [18:23] * yofel wonders where vHanda went missing [18:23] what's up? [18:25] vHanda: where's the current shared-desktop-ontologies repository? We have things failing to build in project neon as it needs s-d-o 0.11, but git://oscaf.git.sourceforge.net/gitroot/oscaf/shared-desktop-ontologies has 0.10.51 [18:25] git://git.code.sf.net/p/oscaf/shared-desktop-ontologies [18:25] sourceforge :/ [18:26] now that looks a lot better :) [18:26] thanks! [18:36] he sent an email about this btw :P [18:37] who wants to take care of marble symbols [18:44] okay done [18:46] 8 files changed, 5644 insertions(+), 4609 deletions(-) [18:49] night [18:56] Damn. That's a lot of changes. ha [18:56] My imx died :( [19:06] shadeslayer: who, where, what mail? === francisco is now known as Guest75516 [19:25] * yofel puts kcm_pimactivity in kmail together with the identity manager [19:30] meh, I'll add a new package that kontact and kmail pull in. That's probably the most sensible thing (at least more sensible than putting it in libkdepim4) [19:31] no other common package? [19:34] well, kdepim doesn't really have a common package. There's kdepim-kresources if anything [19:37] nah that's not much use [19:38] my thought was to add kde-config-pimactivity as that's the only thing in there. I could add a kdepim-common, but for only 2 files... [19:39] I'd think kde-config-pimactivity unless we expect more to be added [20:49] Riddell: hello [20:49] hi ahoneybun [20:49] that was fast [20:50] lol [21:29] Did anyone look if the libdcraw vulnerability that got a USN for 12.04 affected newer releases? [23:23] What, if any, frontend to ufw is preferred now for Kubuntu? guarddog is gone ( but still referred to by ubottu ) [23:26] gufw I think, there's no usable kde UI for firewall management that I know of. [23:27] yofel: OK, thanks