[14:03] too bad, it's impossible to have ttf-mscorefonts-installer in build-deps :( [14:05] asac, ping [14:07] asac, did you hear from Hobbsee yet? [14:08] Jazzva, hi [14:08] fta, hello... how are you? :) [14:08] i want holidays !! [14:09] here they come ;).. I want them too, but tough chances for me... exams and projects :( [15:30] Jazzva: pong [15:30] fta_: nope ;) === fta_ is now known as fta [15:32] asac, I've been planning to update gnome-voice-control to 0.3 for quite a long time now. But it needs pocketsphinx and sphinxbase. It was packaged by David (one of the sphinx devs, I think) and pushed to REVU. I tried to help him to get advocates few times, but nothing. [15:33] So I was wondering if you could help me to get the packages to Debian. I reviewed both of them, and they looked fine to me... [15:34] lol, bypassing revu using debian ;) [15:34] thats the idea of revu ;) [15:34] make it easier to get new packages into debian ;) [15:34] heh :) [15:34] Jazzva: did you leave the comments in revu? [15:34] where are the packages? [15:34] Yep... [15:35] well, it was few months ago, I'm not sure if they're still up [15:35] just to check [15:35] http://revu.ubuntuwire.com/details.py?package=pocketsphinx [15:35] http://revu.ubuntuwire.com/details.py?package=sphinxbase [15:36] RainCT also reviewed them... [15:37] maybe it would be good if we get them to debian. Then I would upgrade ubuntu package to 0.3, push it to debian, then sync it back to ubuntu... yay :) [15:40] Jazzva: do we need them as separate packages? [15:40] couldnt they just be included in gnome-voice-control? [15:41] asac: I guess they could, but isn't it considered bad practice to bundle libraries with the package? [15:41] Jazzva: depends. if they are released in same tarball upstream its ok [15:41] otherwise probably not [15:42] they're not [15:42] asac: gnome-voice-control is in separate tarball [15:42] hmm [15:42] Jazzva: ceck for ITP stuff in debian please [15:43] otherwise open an ITP ... ask upstream guy what role he wants to have ... maybe use you as maintainer and him as uploader or vv [15:43] i think that he would be the maintainer for pocketsphinx and sphinxbase [15:45] Jazzva: ask him if he would want to maintain that in debian [15:45] asac, there RFP, which should be ITP, debian bug 439185 [15:45] Debian bug 439185 in wnpp "RFP: pocketsphinx -- Lightweight speech recognition engine" [Wishlist,Closed] http://bugs.debian.org/439185 [15:46] asac, ok, I will.. I'll get back to you when I get the reply :). [15:46] and thanks :) [15:46] welcome [15:53] hmm, seems ctrl-tab is gone in b2.. [15:59] good ... finally something is done against tabs ;) [16:00] ? [16:00] isnt that used for switching tabs? [16:00] yes [16:00] it was [16:00] so when it doesnt exist anymore, it will train people to use less tabs [16:00] now, the tab list button is back [16:01] but when i single click it (to see the list), it just jumps to the 1st tab :( [16:01] i have to click and hold [16:02] hehe [16:02] yeah the glorious bug [16:06] i assume that the issue should start in nsEventStateManager::CheckForAndDispatchClick [16:06] //If mouse is still over same element, clickcount will be > 1. [16:08] http://mxr.mozilla.org/mozilla-central/source/content/events/src/nsEventStateManager.cpp#2885 [16:10] is that new ? [16:14] no its not new [16:14] but the issue exists in 3.0 too [16:15] its a huge chunk of shit code [16:15] there probably is a bug ;) [16:15] current idea is that ShiftFocus* stuff is called by click on menubar [16:15] which moves the mCurrentTarget field to the menu [16:15] and the code thinks its still the same [16:16] then it dispatches the click ... and since the mouse position is negative you get some randomness ;) [16:16] due to overflow [16:31] asac, what i'm supposed to do for http://launchpadlibrarian.net/20012919/buildlog_ubuntu-hardy-i386.prism_0.9.1%2Bsvn20081126r20278-0ubuntu1~fta2~hardy_FAILEDTOBUILD.txt.gz ? [16:32] fta: http://pastebin.com/f1a682d3d [16:32] can't you include that in the next upload ? it's zero risk [16:32] fta: can you build with that [16:32] and _slowly_ try to reproduce this click thing ;) [16:32] and see if it prints this stuff at all then ;) [16:33] currently tend to "yes" [16:33] s/tend/lean/ [16:34] ? [16:35] 17:32 < fta> can't you include that in the next upload ? [16:35] 17:33 < asac> currently tend to "yes" [16:35] 17:33 < asac> s/tend/lean/ [16:37] hm, still unclear. is it possible to add the build-system into the xul sdk in hardy ? [16:40] asac, http://ubuntuforums.org/showthread.php?p=6274251 [16:41] i've been applying that for a while: https://wiki.ubuntu.com/UbuntuDevelopment/PatchTaggingGuidelines i kind of like it [17:23] Hey, if there's anyone with a more skill than me and a bit of spare time, would you mind taking a look at the various bugs filed against kompozer in 8.10? For some reason in this release it's segfaulting under various circumstances, although the kompozer package itself hasn't changed since Feisty I think (just merged between repos). [17:23] I have no idea what the issue is unfortunately, but there are some useful-looking bug reports that might give one of you a clue. [17:40] tonyyarusso: when was the last rebuild? [17:41] tonyyarusso: what happens on the console when it crashes? [17:43] asac: rebuild meaning what exactly? The last thing I did manually, as in type in "debuild -S" myself was like three releases ago. As far as what the Launchpad automatic stuff or anything done by other release mgmt people, I have no idea. [17:43] tonyyarusso: try to do a rebuild [17:43] asac: normally on the console it just says somethign like "application crashed: segmentation fault" [17:43] and see if things go away [17:43] Anything in particular I should be looking for while the output flies by? [17:46] before crashing? [17:46] no [17:46] no idea ... you could try to get a backtrace [17:46] but i would suggest to first do a rebuild locally and see if that elps [17:47] will do [17:48] btw, what determines the icon that shows up next to a program in "Add/Remove Programs"? The icon is right in the Applications menu once installed, but there's just a generic one in that list. [17:48] tonyyarusso: at best create -dbgsym packages locally so you can get a backtrace later ... install pkg-create-dbgsym before building to achieve that [17:49] tonyyarusso: the icon is part of app-install-data ... maintainer of that is mvo. you can request a merge if you have a branch [17:49] asac: any switches I have to pass as well, or will just the presence of the package be sufficient? Should pkg-create-dbgsym be installed on the local system, or as a build-dep for pbuilder? [18:00] tonyyarusso: ensure that you pass the --disable-strip to configure in your package [18:00] (and no --enable-strip of course) [18:33] \o/ one baby step further.. http://launchpadlibrarian.net/20076538/buildlog_ubuntu-jaunty-i386.chromium-browser_0.3.155.0~svn20081120r5786-0ubuntu1~fta4_FULLYBUILT.txt.gz [18:34] fta, can i press "enter" yet? :) [18:34] i'm currently patching that stuff to detach the testsuite from the source tree [18:39] but as i don't know the code, i'm slow [18:40] especially with that scons build system, which i find far worse than the mozilla makefile based build system [18:47] and without mxr, it's a pain in the *s to dig into such huge pile of code [18:59] asac: could you look at http://paste.ubuntu.com/78380/ and tell me which way it seems to be by default? [19:00] disable is the default [19:00] tonyyarusso: disable [19:00] yeah [19:00] if its default or not ... making it explicit doesnt hurt [19:00] because thats what we want [19:01] (and not whatever is default) [19:04] how to I do that within the context of debuild? edit configure.in? [19:04] * tonyyarusso has never really messed with configure options much [19:04] tonyyarusso: in debian/rules? [19:05] wherever you specify the other configuration options [19:08] my rules file just calls a makefile supplied by upstream [19:10] oh, this file looks useful - mozconfig.debian - has various variables defined that look similar. [19:11] has a line reading "ac_add_options --enable-strip-libs" currently [19:13] Looks like I need to add a line of "ac_add_options --disable-stip" [19:14] asac, fta: it currently has one of "ac_add_options --disable-debug" - what's that do? [19:14] --disable-strip --disable-strip-libs --disable-install-strip is what we use in ff3 [19:15] see about:buildconfig in ff [19:16] asac, btw, about:buildconfig in ff shows xul configure line, not ff [19:19] Man it's been WAY too long since I did this stuff last...forgotten a lot. [19:22] aw crud, what file contains my secret key for gpg again? /me tried copying over secring.gpg, but apparently that wasn't sufficient [19:24] asac, dquilt is not that nice with embedded tarballs and cdbs. this is slow and cdbs keeps changing the config.sub/guess files :( [19:43] tonyyarusso: try the arguments fta suggests ... add them to ac_add_options [19:43] take care that you dont do a typo ;) [19:43] fta: not sure what you mean by "dquilt isnt good" [19:43] fta: shouldnt dquilt detect the build-dir ? [19:44] asac: will do [19:44] * tonyyarusso is creating a pbuilder atm [19:46] pbuilder is for the weak ;) [19:47] what do you use? [19:47] asac, it does, but i made it run pre-build so it unpacks the tarballs, the cons is that it also checks the tarball and performs some cdbs work each time: http://paste.ubuntu.com/78404/ [19:47] last I checked pbuilder was the recommended method... [19:48] fta: yeah. imo thats a known cdbs bug [19:48] not really sure whatfor the cdbs indexing really is [19:48] maybe to figure out if the tarball changes [19:50] yeah :-P [20:05] asac, next meeting Sunday, 7th December, 19:00 UTC ??? [20:07] unlikely [20:07] in MV ;) [20:07] that will be 10am, and with the jet lag... [20:08] fta: jet lag means that you will be awake at 3am .. if its in that direction ;) [20:09] i know, i've been there sereval times [20:10] so 10am isnt that early ;) [20:17] we'll see [21:00] TypeError: coercing to Unicode: need string or buffer, instance found: [21:00] i hate python [21:00] ok i get XXX: found popup overflowing event coordinates: relative x/y: 1500/660 [21:01] when this bug happens [21:01] but sometimes i get that and it doesnt happen [21:01] when i get that i also get a waring in debub build: [21:01] XXX: found popup overflowing event coordinates: relative x/y: 1500/660 [21:01] WARNING: Positioned frame that does not handle positioned kids; looking further up the parent chain: file /home/asac/Development/upstream/mozilla/mozilla-central/layout/base/nsCSSFrameConstructor.cpp, line 7773 [21:03] asac: Okay, I rebuilt the package, and now it crashes on run. See http://paste.ubuntu.com/78437/ [21:03] nsCSSFrameConstructor::GetAbsoluteContainingBlock [21:03] hmm [21:04] tonyyarusso: install the -dbgsym package that you created (if you followed my instrutions) [21:04] 18:48 < asac> tonyyarusso: at best create -dbgsym packages locally so you can get a backtrace later ... install pkg-create-dbgsym before building to achieve that [21:05] asac: well, I installed that package, but in my local environment. I never got an answer as to whether it was supposed to be there or in the pbuilder environment. [21:05] (and I don't appear to have one, so I'm guessing the latter) [21:06] err, isnt that obvious ;) [21:06] tonyyarusso: if you havent delete the biult tree you can just go into it [21:06] and rebuild with that package installed using debuild -nc [21:06] should be rather quick [21:06] like 2 minutes ;) [21:08] Okay, so just to be clear: 1) edit debian/control to include it in the build-depends, 2) cd .. && debuild -nc, 3) re-run pbuilder build [21:10] tonyyarusso: build depends? whatever. it has to be installed during build [21:10] ;) [21:10] why rerun pbuilder? [21:10] debuild -nc is the rerun [21:10] you have to do that in the old pbuilder chroot [21:11] if you have removed it you probably notice why pbuilder is suboptimal for packages that take longer than 4 minutes ;) [21:11] * tonyyarusso is lost [21:11] whatever you do ... dont do it in pbuilder [21:11] install build depends in main system or in a dedicated chroot [21:11] and use that [21:12] pbuilder is only to test if the biuld depends are right imo [21:12] Also handy just for the purpose of keeping my main system clean, but whatever. [21:13] main system clean? ;) ... thats like keeping your school books in a library and not at home ;) [21:14] i would suggest to setup a long living chroot if you want your main system to be virgin like [21:14] Yeah - would have to learn how to do that too :) === rzr is now known as RzR [21:15] heh ;) https://wiki.ubuntu.com/DebootstrapChroot [21:16] take care that once you mount bind stuff like your home you shouldn just rm -rf /path/to/chroot [21:16] otherwise you will wipe your complete home ;) [21:16] just because i ended up in that situation once ;) [21:16] icky [21:16] but usually one doenst need to wipe a chroot ... so just remember to check that nothing is bind mounted in it ;) [21:16] when you remove it at some point [21:17] how would you check that? run mount within it? [21:25] hello... fta around ? [21:26] plun, yes [21:26] The test shell works !!! ;) [21:26] I can also surf.... ;) [21:27] good, but it's not supposed to, yet. [21:27] i'm still working on it [21:27] I got a update earlier this evening and checked it.... ;) [21:28] yep, those are my test builds [21:30] it's supposed to start like that: http://www.sofaraway.org/ubuntu/tmp/chromium.png [21:30] not with an empty page like my last package [21:30] plun, ^^ [21:31] Aha... I started with the test_shell script.... http://ubuntu-pics.de/bild/6589/screen_gD7pEY.png [21:32] Your page also works.... ;) [21:32] plun, don't expect a full browser anytime soon, upstream is still far behind [21:33] Yup... this is just a happening... ;) [21:34] Well it also crashes... [21:34] Failed reading: /usr/lib/chromium-browser/../../webkit/tools/test_shell/resources/missingImage.gif [21:34] Trace/breakpoint trap [21:35] yep, that's what i'm working on [21:36] (the evil "../../webkit") [21:36] Okidok.... "playing" with this, just great anyway !! Good night [21:36] thanks [21:36] n8 [21:50] asac: Okay, so I seem to have created a dbgsym package and installed it. Now what? [21:53] tonyyarusso: reproduce crash [21:53] get backtrace [22:37] hm, ff3.1 crashed [22:38] strangely, it didn't ask me which tabs I wanted to restore, it restored them all, like before [22:38] another backout ? [22:39] mconnor, ^^ ? [22:41] asac, http://paste.ubuntu.com/78474/ [22:42] fta: mmm, iirc the plan was to only prompt if there were two crashes in swift succession [22:43] hmm... ok. what about ctrl+tab ? is it gone in b2 ? [22:45] yes [22:45] in 3.1, really [22:46] really? why? [22:47] should i switch back to the addon then? [22:47] mconnor, with ~150 tabs, the tab list is not usable [22:48] fta: the UI wasn't good enough to ship [22:49] mconnor, i see. is it completely gone or just prefed out ? [22:49] it's selectively excised via #ifdef [22:50] the plan is to get it in 3.2 [22:50] we just didn't like the way the interaction was working, and we weren't going to slip the release for it [22:50] damn, ok then. i'll switch back to the addon then. [22:50] you're still way way way out on the long tail :) [22:51] when i 3.1 final planed? [22:51] is