[00:49] apachelogger: I love Austria! http://www.bbc.co.uk/news/world-europe-14135523 === emma is now known as em [08:16] morning [08:17] o/ [08:17] yofel: i have everything worked out except for the description :P [08:17] symbols are here http://pastebin.com/raw.php?i=yy1VWeee and http://pastebin.com/raw.php?i=NzvNgxV2 [08:41] morning [08:44] fun, kdelibs building for 9h now, 72% === Squt is now known as Sput [09:26] apachelogger: amarok (newest git) has problems building under current 11.10 [09:26] http://paste.kde.org/96205/ [09:26] I think it could have to do with libgcrypt.so moving to /lib from /usr/lib [09:26] cmake seems to find libgcrypt correctly but the -lgcrypt doesn't find it [09:36] yofel: Do we have kdelibs-mobile? [09:37] not that I know of [09:37] * yofel tries to get ccache and icecc working again in pbuilder [09:37] it's like the initial run of dpkg-buildpackage sets it's own PATH [09:37] once you drop to shell they work fine [10:03] yofel: it might just b executed from the wrong envrionment [10:04] could be, what irritates me is that an A99env hook prints PATH=/usr/lib/ccache:/usr/lib/icecc/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/X11R6/bin [10:04] while cmake sees PATH: /usr/share/pkg-kde-tools/bin:/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games [10:05] which doesn't even have the pbuilder base path [10:06] report a bug [10:06] ah [10:06] hold on [10:06] yofel: how do you export the path? [10:06] or where [10:06] or when [10:06] actually all of them :P [10:07] I export it in pbuilderrc [10:08] just reading pbuilder scripts [10:08] grep them for PATH ^^ [10:08] good idea [10:08] only gives pbuilder-checkparams: export PATH="/usr/lib/ccache:$PATH" :/ [10:09] which obviously works [10:14] * yofel adds some debugging output [10:27] I think it fails in the way pbuilder calls the build: [10:27] echo "cd tmp/buildd/*/ && dpkg-buildpackage -us -uc -j6 -rfakeroot" | chroot /var/cache/pbuilder/build//4560 env LOGNAME=pbuilder su -p pbuilder [10:32] or not o.O [10:41] helps to search the debian BTS - debian bug 606687 [10:41] Debian bug 606687 in pbuilder "ccache support fails in the face of su PATH mangling" [Normal,Open] http://bugs.debian.org/606687 [10:56] apachelogger: yofel need descriptions for libkateineterfaces4 and libktexteditorcodesnippetscore0. [10:57] lolwut? [10:58] apachelogger: public libs, need to split out of kate [11:01] like I know what those do [11:03] i have no idea myself [11:05] so read? [11:08] want me to put patched pbuilder and icecc into ninjas? This is pretty much unfunctional otherwise [11:09] sure whatever why not [11:12] The Kate application interfaces provide a better integration into the app like tool views. [11:12] doxygen++ [11:13] This library contains the code used by kate plugins that provide a better integration into the app like tool views [11:14] good? [11:14] apachelogger: did you see my libgcrypt problem? [11:15] shadeslayer: sounds like it does something, so good ^^ [11:17] O_O [11:17] libktexteditorcodesnippetscore0 is actually under kate/plugins [11:18] should i move it to kate-plugins? [11:18] what do you mean? /usr/lib/libktexteditor_codesnippets_core.so.0 [11:19] HorusHorrendus: rm -rf build [11:19] well yes, but in the source tarball its kate-4.6.90/kate/plugins/snippets_tng/lib/ [11:21] apachelogger: trust me ... I tried that [11:21] and many other things [11:21] checked if libgcrypt is installed [11:21] we also checked out a new FindLibgrcrpyt.cmake [11:21] shadeslayer: then ask them why the hell they install that in /usr/lib [11:22] yeah, emailing the author [11:23] HorusHorrendus: report a bug I guess [11:23] for libgcrypt? [11:24] whatever linker you are using [11:24] gold I think ... [12:32] bug 662433 [12:32] Launchpad bug 662433 in kcm-touchpad (Ubuntu) "Add "disable while typing" option" [Undecided,New] https://launchpad.net/bugs/662433 [12:32] didnt we like have that feature? [12:33] y u loose my features [12:33] yofel: y u no blog? [12:35] I'll spend most of today studying after wasting time on pbuilder, I'll think about it [12:42] yofel: ok, then I'll do it :P === ximion1 is now known as ximion [13:49] hey apachelogger, there are quite a lot of kde templates in the imports queue, and I'm wondering what to do with them. You said upstream had changed quite a lot of things wrt the l10n packages. Could you give me an overview of some of the changes so I know how I should approve or move around the templates in LP? [14:11] shadeslayer: what's the ETA for the kate package? [14:11] debfx: i've asked upstream if libktexteditorcodesnippetscore0 is indeed a public lib and whether or not to ship it in its own package [14:12] because libktexteditorcodesnippetscore0 is actually a part of kate plugins [14:14] apart from that i have everything ready ( well except for a description for libktexteditorcodesnippetscore0 ) [14:16] shadeslayer: maybe an external plugin also uses that library [14:17] if not we could just remove the header files from kate-dev and move the lib to kate-plugins [14:17] debfx: i leave it to your discretion, if you want, i have everything ready to go [14:25] shadeslayer: ok, I'll do that. could you push your changes? [14:26] debfx: sure, the only thing left is a description, lemme finish that up and then i'll push it :) [14:27] debfx: i have a question about shlibs [14:27] shadeslayer: don't worry about libktexteditorcodesnippetscore0, I'll move it to kate-plugins for now [14:27] alright [14:38] dpm: if only I knew [14:38] so [14:39] dpm: kdebase is now kde-baseapps, kdebase-runtime is now kde-runtime, kdebase-workspace is now kde-workspace [14:39] templates that formerly all came out of the kdegraphics source are now distributed across multiple source packages, though the template name should likely ahve changed the same [14:40] those include libksane, libkexiv2, libkdcraw, libkipi, ksaneplugin, kolourpaint, ksnapshot, gwenview, kruler, kcolorchoose, kgamma, kamera and okular [14:40] kate ealier was part of the kdesdk source, is now an independent source [14:41] konsole ealier was part of the kdebase source, is now an independent source too [14:41] kdeedu, like degraphics was split into multiple sources [14:41] blinken, cantor, kalgebra, kalzium, kanagram, kbruch, kgeography, khangman, kig, kiten, klettres, kmplot, kstars, ktouch, kturtle kwordquiz, libkdeedu, marble, parley, rocs and step [14:42] dpm: any new templates that are not covered by those changes need specific investigation I suppose [14:43] apachelogger, ok, that's really helpful already, so I've got something to start with. Thanks! :) [14:43] yw [14:48] [lp:~kubuntu-packagers/kubuntu-packaging/kate] Rohan Garg * 20 * debian/ (6 files) * Move arch indep files from kate to kate-data [14:49] debfx: ^^ [14:49] please review and kick me for my mistakes :) [14:55] who is in 4.7? [14:56] apachelogger: you mean the packaged 4.7? [14:56] or does Neon count? :P [14:56] anything that looks like 4.7 [14:56] I need a snazzy screenshot [14:57] anything in particular? [14:59] something snazzy [15:00] sheytan is probably your guy in that case [15:00] why? [15:00] i have a very boring desktop http://i.imgur.com/zO5Tc.jpg [15:00] he is not running 4.7 [15:00] apachelogger: iirc he has neon as well [15:00] oh good lord [15:00] * apachelogger reports a bug [15:00] ?? [15:06] shadeslayer: plz fix [15:06] hahaha [15:09] that is no fun business at all [15:16] Hi, do somebody know when the Kate will be updated to 3.7 RC1? At the moment it just crashes if I try to run it because KDE is 4.7 RC1 and Kate is still old version. [15:17] karmo: give shadeslayer a beating [15:17] he is holding it back [15:18] hey i pushed it to bzr, debfx is (hopefully) reviewing it [15:18] * yofel needs a break [15:22] ok:) i really miss my favorite editor [15:23] use emacs :P [15:24] nah, vim in yakuake [15:25] sooo... when are we wrapping up rc1? gwenview is stuck, I'm not sure how to split mobipocket, neither is it important, and smoke still has no sane copyright upstream [15:26] upload kate and kstars and move to rc2? [15:26] perlkde [15:26] thats left too iirc [15:26] what about networkmanager-kde? it doesn't work for wireless networks and i'm using gnome applet to connect wifi :) [15:26] shadeslayer: I can do perl in rc2 too... [15:26] karmo: broken due to new networkmanager backend iirc [15:26] if we postpone this any longer we might as well skip rc2 and wait for final to come out [15:27] karmo: the plasma widget? works fine for me in oneiric [15:27] (it's from the nm-0.9 branch) [15:28] i'm a bit busy for the next couple of days [15:28] won't be able to help until next week [15:28] apachelogger: can you bring your N900 to DS? [15:29] i want to test a PIM sync over a local network [15:33] hm, if mobipocket is too generic, how about okular-plugin-mobipocket and strigi-plugin-mobipocket then? [15:36] shadeslayer: any other phones I should bring? [15:37] yofel: kdegraphics-plugin-mobipocket [15:37] eh [15:37] kdegraphics-mobipocket [15:38] hm, if that's enought for scott then perfect [15:38] -t [15:47] [lp:~kubuntu-packagers/kubuntu-packaging/mobipocket] Philip Muškovac * 10 * debian/control rename into kdegraphics-mobipocket and run wrap-and-sort [15:47] I made it enhance okular too [15:49] apachelogger: can you look at it later? (but today if possible) [15:49] and someone please look at kstars [15:56] i'll look at kstars [15:57] I fixed some things, so should be ready for upload (except for indi being too old) === mischa is now known as mischasworld [15:59] JontheEchidna: bug 810529 [15:59] Launchpad bug 810529 in kubuntu-meta (Ubuntu) "[regression] Kubuntu contributors using broken editors" [High,Won't fix] https://launchpad.net/bugs/810529 [15:59] JontheEchidna: I think those random bug triagers get all smartass on me [15:59] LOL [16:00] :P [16:00] stop mis-using launchpad for personal differences :P [16:00] Riddell: you have libindi 0.7.2 in your PPA [16:00] yofel: tell that to apachelogger [16:00] QML session starting in #ubuntu-classroom now (btw) [16:00] ^^ i need that [16:00] * shadeslayer joins [16:01] shadeslayer: we need indi 0.8.0 [16:01] yofel: yeah, packaging that [16:01] k [16:01] indi is a mess to package [16:02] be careful with the plugins, they don't all have source or get released [16:04] fun [16:04] i just need the lib itself i think [16:05] weird how debian does not have the package [16:09] well you might just need the lib to compile against, but the lib won't do anything unless it has working plugins [16:10] akonadi [16:10] yofel [16:10] ppa/bzr [16:10] NEEDS UPLOAD [16:10] yofel: didnt I upload akonadi? [16:10] !info akonadi oneiric [16:10] Package akonadi does not exist in oneiric [16:10] fun :D [16:10] !info akonadi-server oneiric [16:10] akonadi-server (source: akonadi): Akonadi PIM storage service. In component main, is extra. Version 1.5.80-0ubuntu1 (oneiric), package size 524 kB, installed size 1996 kB [16:10] you uploaded 1.5.80, ppa has 1.6.0 [16:10] oh [16:10] bzr had .80 [16:10] uh... [16:11] * apachelogger does not release from ppa [16:11] Riddell: ah, are the plugins from your ppa any good? [16:11] apachelogger: refresh bzr? [16:11] well [16:11] dont have time [16:11] * apachelogger is writing blog post [16:11] well, not that important either [16:11] write the post [16:12] yofel: where did you get working networkmanager for kde in oneiric? [16:13] archive? [16:13] shadeslayer: hard to tell, I'll forward you the discussion I had with upstream last I looked [16:13] sure [16:13] !info plasma-widget-networkmanagement | karmo [16:13] karmo: plasma-widget-networkmanagement (source: plasma-widget-networkmanagement): Network Management widget for KDE4 Plasma. In component main, is optional. Version 0.9~svngit20110408-0ubuntu2 (natty), package size 62 kB, installed size 296 kB [16:13] bah [16:14] !info plasma-widget-networkmanagement oneiric | karmo [16:14] karmo: plasma-widget-networkmanagement (source: networkmanagement): Network Management widget for KDE Plasma workspaces. In component main, is optional. Version 0.9~svngit.nm09.20110610.0c83d8-0ubuntu1 (oneiric), package size 839 kB, installed size 3588 kB [16:14] as you see from the version, that's for nm09 [16:15] hmm [16:20] Riddell: http://sourceforge.net/projects/indi/files/indilib/0.8/ just has the lib [16:20] shadeslayer: right, upstream isn't very good at releasing plugins [16:20] [lp:~kubuntu-packagers/kubuntu-packaging/kdelibs] Philip Muškovac * 364 * debian/ (6 files) * New upstream release candidate [16:21] so we just use the older ones you packaged? [16:22] then you'll have to patch the cmake check, which I doubt is the right thing [16:29] does sound nasty [16:29] but i managed to build libindi [16:34] * yofel trys to get perlqt done [16:34] *tries [16:35] * debfx wonders if we should put katepart into an own package since it was part of kdelibs5-plugins [16:40] wasn't katepart in its own package? [16:40] iirc it was in libkatepartinterfaces4 [16:41] nope, it's in kate [16:42] shadeslayer: that's just the library katepart uses [16:43] oh derp [17:02] I need saxy screenshot of the 4.7 or I will start waving my fist again! [17:02] also I need someone to read blog post [17:02] Nightrose: honey :) :) :) [17:03] ah [17:03] I know [17:03] the lazy ol' yofel :P [17:04] yofel: you haz mail [17:04] * apachelogger leaves to play some golf [18:02] yofel: http://i.imgur.com/qp6kz.png << \o/ [18:02] i can finally haz memo sync [18:03] shadeslayer: Can you take a look at https://bugs.kde.org/show_bug.cgi?id=235020 ? [18:03] KDE bug 235020 in kdeui "crash while using character selector widget" [Crash,New] [18:04] * apachelogger beats shadeslayer to make a proper screenshot [18:04] apachelogger: whats a proper screenshot? [18:04] * Quintasan can't reproduce that while user in #ubuntu-pl gets this very often [18:04] one that does not contain unintersting plunder and uses a proper wallpaper [18:06] Quintasan: why dont you investigate? [18:06] ^^ yes [18:06] apachelogger: Since I don't get a shit what's going on in those backtraces nor I have coding skillz? [18:06] debug info is pretty good [18:07] Quintasan: but wouldnt it be because of that a perfect learning opportunity? [18:07] shadeslayer: no it is not [18:07] apachelogger: Learning what? [18:07] whoever wrote kcharselect did not know about the use of asserts clearly [18:07] apachelogger: whut [18:07] Quintasan: codingz, debugginz, shitz [18:08] shadeslayer: the debug info is crapz or the bug would be fixed already [18:08] apachelogger: how is that? you can clearly see where its crashing, you get the line number and everything [18:09] that does not make it good debug info as you dunno why shit goes wrong there [18:09] * Quintasan has no idea how to read that magic and even less ideas how to fix that [18:11] FWIW I cannot reproduce it either [18:12] ahh [18:12] Nor can I [18:12] apachelogger: lol @ comment 10 [18:13] apachelogger: the model index is out of range which is why it crashes? [18:15] *shrug* [18:15] shadeslayer: does that knowledge help? [18:16] i'm just speculating [18:17] well, usually it is that kind of thing with models [18:17] which is why one should be using plenty of asserts [18:18] hehehe [18:21] well [18:21] shadeslayer: I find the code there a bit weird [18:21] i haven't read the code, thats all i could infer from the backtrace [18:22] what type is htat anyway [18:22] ah [18:22] * apachelogger hates reading code in a webz browsa [18:23] it crashes at a ...brace ? 0.o [18:24] what [18:24] no [18:24] silly shadeslayer [18:24] QChar c = m_chars[pos]; [18:24] at least the most recent bracktrace [18:24] i was looking at the original report [18:25] ah yes, [18:25] chances are they are different issues [18:25] in my experience any bug that was thought to be fixed once but was reopened is representing at least 2 different issues [18:25] more likely >2 [18:26] phonon xine is a good example for that :P [18:26] then again it is shitty code and crashes all the time and everywhere, even without solar flares [18:27] now what is interesting about that most recent crash in kcharselect is that m_chars is a qlist [18:27] now as usual accessing a qlist with the [] operator is a bit of a dangerous thing because the index *must* be valid [18:28] to help comes line 680 [18:28] if (pos >= m_chars.size() || index.row() < 0 || index.column() < 0) { [18:28] inside only returns can happen [18:28] now the constraint for a valid index is 0 <= i < size IIRC [18:29] hahaha [18:29] sup? [18:30] anyhow, that pos >= m_chars.size() should ensure valid index [18:30] which leaves 2 options [18:31] a) pos exceeds int range and goes int -int which would of course then result in pos < size BUT pos < 0, which doesnt meet the requirement for valid index [18:31] looking at the calcuation for pos [18:31] int pos = m_columns * (index.row()) + index.column(); [18:32] I'd say that is a possibility as result from brekage elsewhere [18:32] Elsewhere being? [18:32] b) the qchar at m_chars[pos] is kaput, but since it is not a list of qchar* I find it unlikely [18:33] QChar c = m_chars[pos]; [18:33] that would at least result in a call to the assignment operator [18:33] and there it would crash [18:33] so, putting the evidance together.... [18:33] we have an if that checks half the valid index constraint [18:34] we have an index operator access on qlist [18:34] and we have a crash in ::data() even though the offending line should result in a deeper stack in terms of QChar copy/assignment [18:35] but oh, we are not there yet it seems [18:36] inline T &QList::operator[](int i) [18:36] { Q_ASSERT_X(i >= 0 && i < p.size(), "QList::operator[]", "index out of range"); [18:36] detach(); return reinterpret_cast(p.at(i))->t(); [18:37] that is the [] operator of qlist [18:37] and looking at it I wonder [18:37] if it were scenario a) as presented earlier [18:37] why does the stack not contain the assert call? [18:38] usually if you assert in a kapp the stack from assert position to what happens as part of an assert (qwarning and whatnot) should appear in the stack trace [18:38] but it is not in this case [18:38] however, perhaps their distros are built without debug in Qt [18:38] in which case the assert would not be fatal [18:39] and here comes Quintasan in [18:39] Quintasan: you said ubuntu users have seen that crash? [18:39] did they show you a backtrace? did they report a bug? [18:39] apachelogger: Indeed [18:39] https://bugs.kde.org/show_bug.cgi?id=235020 [18:39] Comment 44 [18:39] KDE bug 235020 in kdeui "crash while using character selector widget" [Crash,New] [18:39] I see [18:40] shadeslayer: what do we conclude? [18:40] users suck [18:40] and? [18:41] Quintasan: is that guy online by any chance? [18:41] apachelogger: Yes [18:41] asserts rule ? :P [18:41] Quintasan: ok, ask him to start kcharselect form a terminal and try to trigger the crash [18:41] shadeslayer: and? [18:42] dunno what else [18:42] ...and the code is crap and we dunno what the cause is because apachelogger eliminated a) and b) :P [18:42] unless!!!!!....!!!!#@12! [18:42] zomg [18:42] here is the thing [18:43] hmm? [18:43] iff I do not remember correctly and asserts do not show up as part of the stack trace in drkonqi, then it could still be a) [18:43] apachelogger: [18:43] http://wklej.to/Rrl9P [18:43] what we can say almost definitely is that it is not case b) [18:43] nothing interesting here [18:43] it would be crashing elsewhere if it were in the assignment [18:44] Quintasan: indeed [18:44] Quintasan: gdb kcharselect [18:44] r [18:44] ...trigger crash... [18:44] thread apply all bt [18:45] thread apply all bt <-- is that gdb command? [18:45] yuz [18:46] applies the bt (backtrace) command to all threads ^^ [18:49] apachelogger: http://wklej.to/hNrnH [18:52] apachelogger: Is that something more interesting? [18:53] hm [18:53] p pos [18:54] * apachelogger wonders if that works when the segfault already appeared ^^ [18:55] apachelogger: Fanfares [18:55] $1 = -469761976 [18:55] *shrug* [18:55] I win [18:55] a) right there [18:56] Quintasan: is he knowlegable enough to write a comment on the bug? [18:56] shadeslayer: ^ I get 50 bucks from you [18:56] apachelogger: Not at all, just your typical user [18:56] apachelogger: sure, 50 INR [18:56] :DDDDD [18:56] Quintasan: ok, I'll write the findings down, tell him the thanks for helping [18:56] thats like 1 USD [18:56] shadeslayer: more like 0.1 USD [18:56] and you call that moniez!!!! [18:57] your math skillz are screwed up [18:57] yes [18:57] I no [18:57] I am software engineer [18:57] google says its 1 USD :P [18:57] 1.1 rather [18:58] apachelogger: btw. do they sell Dr. Pepper somewhere near your dwelling? [18:59] sweet baby jesus [18:59] [40146.009081] Out of memory: Kill process 28238 (knotify4) score 889 or sacrifice child [19:00] hahah [19:00] swap: 700MB [19:00] o_O [19:00] WTF? [19:00] lol .. i'm using QMap::count incorrectly it seems [19:01] kernel oops inside "oom" killer itself... well... I will reboot with the 2.6.38 :D [19:01] bambee: Magic [19:01] hehe [19:02] Out Of Magic [19:02] heh [19:02] :P [19:02] Quintasan: ohoh [19:02] a good point for linux 3.0-rc7 => nouveau [19:02] it's very fast ! [19:02] Quintasan: if at all possible it would be cool to get the output of [19:02] Quintasan: LOL [19:02] p m_columns [19:02] p index.row() [19:02] and p index.column() [19:02] supposedly they are optimized out though at that point in the code [19:03] p m_columns - $2 = 18 [19:04] p index.row() - Cannot evaluate function -- may be inlined [19:04] p index.column() - See above [19:04] apachelogger: ^ [19:04] ah yeah, thx [19:05] apachelogger: I told him not to kill it yet [19:05] Quintasan: yeah, I have everything worth having :) [19:05] * Quintasan goes packing stuff for tomorrow's raid on the capital of Poland [19:12] Quintasan: https://bugs.kde.org/show_bug.cgi?id=235020#c46 [19:12]