UbuntuIRC / 2008 /11 /30 /#ubuntu-mozillateam.txt
niansa
Initial commit
4aa5fce
[14:03] <fta> too bad, it's impossible to have ttf-mscorefonts-installer in build-deps :(
[14:05] <Jazzva> asac, ping
[14:07] <fta> asac, did you hear from Hobbsee yet?
[14:08] <fta> Jazzva, hi
[14:08] <Jazzva> fta, hello... how are you? :)
[14:08] <fta> i want holidays !!
[14:09] <Jazzva> here they come ;).. I want them too, but tough chances for me... exams and projects :(
[15:30] <asac> Jazzva: pong
[15:30] <asac> fta_: nope ;)
=== fta_ is now known as fta
[15:32] <Jazzva> 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] <Jazzva> 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] <fta> lol, bypassing revu using debian ;)
[15:34] <asac> thats the idea of revu ;)
[15:34] <asac> make it easier to get new packages into debian ;)
[15:34] <Jazzva> heh :)
[15:34] <asac> Jazzva: did you leave the comments in revu?
[15:34] <asac> where are the packages?
[15:34] <Jazzva> Yep...
[15:35] <Jazzva> well, it was few months ago, I'm not sure if they're still up
[15:35] <Jazzva> just to check
[15:35] <Jazzva> http://revu.ubuntuwire.com/details.py?package=pocketsphinx
[15:35] <Jazzva> http://revu.ubuntuwire.com/details.py?package=sphinxbase
[15:36] <Jazzva> RainCT also reviewed them...
[15:37] <Jazzva> 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] <asac> Jazzva: do we need them as separate packages?
[15:40] <asac> couldnt they just be included in gnome-voice-control?
[15:41] <Jazzva> asac: I guess they could, but isn't it considered bad practice to bundle libraries with the package?
[15:41] <asac> Jazzva: depends. if they are released in same tarball upstream its ok
[15:41] <asac> otherwise probably not
[15:42] <Jazzva> they're not
[15:42] <Jazzva> asac: gnome-voice-control is in separate tarball
[15:42] <asac> hmm
[15:42] <asac> Jazzva: ceck for ITP stuff in debian please
[15:43] <asac> 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] <Jazzva> i think that he would be the maintainer for pocketsphinx and sphinxbase
[15:45] <asac> Jazzva: ask him if he would want to maintain that in debian
[15:45] <Jazzva> asac, there RFP, which should be ITP, debian bug 439185
[15:45] <ubottu> Debian bug 439185 in wnpp "RFP: pocketsphinx -- Lightweight speech recognition engine" [Wishlist,Closed] http://bugs.debian.org/439185
[15:46] <Jazzva> asac, ok, I will.. I'll get back to you when I get the reply :).
[15:46] <Jazzva> and thanks :)
[15:46] <asac> welcome
[15:53] <fta> hmm, seems ctrl-tab is gone in b2..
[15:59] <asac> good ... finally something is done against tabs ;)
[16:00] <fta> ?
[16:00] <asac> isnt that used for switching tabs?
[16:00] <fta> yes
[16:00] <fta> it was
[16:00] <asac> so when it doesnt exist anymore, it will train people to use less tabs
[16:00] <fta> now, the tab list button is back
[16:01] <fta> but when i single click it (to see the list), it just jumps to the 1st tab :(
[16:01] <fta> i have to click and hold
[16:02] <asac> hehe
[16:02] <asac> yeah the glorious bug
[16:06] <asac> i assume that the issue should start in nsEventStateManager::CheckForAndDispatchClick
[16:06] <asac> //If mouse is still over same element, clickcount will be > 1.
[16:08] <asac> http://mxr.mozilla.org/mozilla-central/source/content/events/src/nsEventStateManager.cpp#2885
[16:10] <fta> is that new ?
[16:14] <asac> no its not new
[16:14] <asac> but the issue exists in 3.0 too
[16:15] <asac> its a huge chunk of shit code
[16:15] <asac> there probably is a bug ;)
[16:15] <asac> current idea is that ShiftFocus* stuff is called by click on menubar
[16:15] <asac> which moves the mCurrentTarget field to the menu
[16:15] <asac> and the code thinks its still the same
[16:16] <asac> then it dispatches the click ... and since the mouse position is negative you get some randomness ;)
[16:16] <asac> due to overflow
[16:31] <fta> 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] <asac> fta: http://pastebin.com/f1a682d3d
[16:32] <fta> can't you include that in the next upload ? it's zero risk
[16:32] <asac> fta: can you build with that
[16:32] <asac> and _slowly_ try to reproduce this click thing ;)
[16:32] <asac> and see if it prints this stuff at all then ;)
[16:33] <asac> currently tend to "yes"
[16:33] <asac> s/tend/lean/
[16:34] <fta> ?
[16:35] <asac> 17:32 < fta> can't you include that in the next upload ?
[16:35] <asac> 17:33 < asac> currently tend to "yes"
[16:35] <asac> 17:33 < asac> s/tend/lean/
[16:37] <fta> hm, still unclear. is it possible to add the build-system into the xul sdk in hardy ?
[16:40] <fta> asac, http://ubuntuforums.org/showthread.php?p=6274251
[16:41] <fta> i've been applying that for a while: https://wiki.ubuntu.com/UbuntuDevelopment/PatchTaggingGuidelines i kind of like it
[17:23] <tonyyarusso> 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] <tonyyarusso> 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] <asac> tonyyarusso: when was the last rebuild?
[17:41] <asac> tonyyarusso: what happens on the console when it crashes?
[17:43] <tonyyarusso> 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] <asac> tonyyarusso: try to do a rebuild
[17:43] <tonyyarusso> asac: normally on the console it just says somethign like "application crashed: segmentation fault"
[17:43] <asac> and see if things go away
[17:43] <tonyyarusso> Anything in particular I should be looking for while the output flies by?
[17:46] <asac> before crashing?
[17:46] <asac> no
[17:46] <asac> no idea ... you could try to get a backtrace
[17:46] <asac> but i would suggest to first do a rebuild locally and see if that elps
[17:47] <tonyyarusso> will do
[17:48] <tonyyarusso> 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] <asac> 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] <asac> 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] <tonyyarusso> 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] <asac> tonyyarusso: ensure that you pass the --disable-strip to configure in your package
[18:00] <asac> (and no --enable-strip of course)
[18:33] <fta> \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] <directhex> fta, can i press "enter" yet? :)
[18:34] <fta> i'm currently patching that stuff to detach the testsuite from the source tree
[18:39] <fta> but as i don't know the code, i'm slow
[18:40] <fta> especially with that scons build system, which i find far worse than the mozilla makefile based build system
[18:47] <fta> and without mxr, it's a pain in the *s to dig into such huge pile of code
[18:59] <tonyyarusso> asac: could you look at http://paste.ubuntu.com/78380/ and tell me which way it seems to be by default?
[19:00] <fta> disable is the default
[19:00] <asac> tonyyarusso: disable
[19:00] <asac> yeah
[19:00] <asac> if its default or not ... making it explicit doesnt hurt
[19:00] <asac> because thats what we want
[19:01] <asac> (and not whatever is default)
[19:04] <tonyyarusso> 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] <asac> tonyyarusso: in debian/rules?
[19:05] <asac> wherever you specify the other configuration options
[19:08] <tonyyarusso> my rules file just calls a makefile supplied by upstream
[19:10] <tonyyarusso> oh, this file looks useful - mozconfig.debian - has various variables defined that look similar.
[19:11] <tonyyarusso> has a line reading "ac_add_options --enable-strip-libs" currently
[19:13] <tonyyarusso> Looks like I need to add a line of "ac_add_options --disable-stip"
[19:14] <tonyyarusso> asac, fta: it currently has one of "ac_add_options --disable-debug" - what's that do?
[19:14] <fta> --disable-strip --disable-strip-libs --disable-install-strip is what we use in ff3
[19:15] <fta> see about:buildconfig in ff
[19:16] <fta> asac, btw, about:buildconfig in ff shows xul configure line, not ff
[19:19] <tonyyarusso> Man it's been WAY too long since I did this stuff last...forgotten a lot.
[19:22] <tonyyarusso> 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] <fta> 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] <asac> tonyyarusso: try the arguments fta suggests ... add them to ac_add_options
[19:43] <asac> take care that you dont do a typo ;)
[19:43] <asac> fta: not sure what you mean by "dquilt isnt good"
[19:43] <asac> fta: shouldnt dquilt detect the build-dir ?
[19:44] <tonyyarusso> asac: will do
[19:44] * tonyyarusso is creating a pbuilder atm
[19:46] <asac> pbuilder is for the weak ;)
[19:47] <tonyyarusso> what do you use?
[19:47] <fta> 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] <tonyyarusso> last I checked pbuilder was the recommended method...
[19:48] <asac> fta: yeah. imo thats a known cdbs bug
[19:48] <asac> not really sure whatfor the cdbs indexing really is
[19:48] <asac> maybe to figure out if the tarball changes
[19:50] <asac> yeah :-P
[20:05] <fta> asac, next meeting Sunday, 7th December, 19:00 UTC ???
[20:07] <asac> unlikely
[20:07] <asac> in MV ;)
[20:07] <fta> that will be 10am, and with the jet lag...
[20:08] <asac> fta: jet lag means that you will be awake at 3am .. if its in that direction ;)
[20:09] <fta> i know, i've been there sereval times
[20:10] <asac> so 10am isnt that early ;)
[20:17] <fta> we'll see
[21:00] <fta> TypeError: coercing to Unicode: need string or buffer, instance found:
[21:00] <fta> i hate python
[21:00] <asac> ok i get XXX: found popup overflowing event coordinates: relative x/y: 1500/660
[21:01] <asac> when this bug happens
[21:01] <asac> but sometimes i get that and it doesnt happen
[21:01] <asac> when i get that i also get a waring in debub build:
[21:01] <asac> XXX: found popup overflowing event coordinates: relative x/y: 1500/660
[21:01] <asac> 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] <tonyyarusso> asac: Okay, I rebuilt the package, and now it crashes on run. See http://paste.ubuntu.com/78437/
[21:03] <asac> nsCSSFrameConstructor::GetAbsoluteContainingBlock
[21:03] <asac> hmm
[21:04] <asac> tonyyarusso: install the -dbgsym package that you created (if you followed my instrutions)
[21:04] <asac> 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] <tonyyarusso> 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] <tonyyarusso> (and I don't appear to have one, so I'm guessing the latter)
[21:06] <asac> err, isnt that obvious ;)
[21:06] <asac> tonyyarusso: if you havent delete the biult tree you can just go into it
[21:06] <asac> and rebuild with that package installed using debuild -nc
[21:06] <asac> should be rather quick
[21:06] <asac> like 2 minutes ;)
[21:08] <tonyyarusso> 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] <asac> tonyyarusso: build depends? whatever. it has to be installed during build
[21:10] <asac> ;)
[21:10] <asac> why rerun pbuilder?
[21:10] <asac> debuild -nc is the rerun
[21:10] <asac> you have to do that in the old pbuilder chroot
[21:11] <asac> 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] <asac> whatever you do ... dont do it in pbuilder
[21:11] <asac> install build depends in main system or in a dedicated chroot
[21:11] <asac> and use that
[21:12] <asac> pbuilder is only to test if the biuld depends are right imo
[21:12] <tonyyarusso> Also handy just for the purpose of keeping my main system clean, but whatever.
[21:13] <asac> main system clean? ;) ... thats like keeping your school books in a library and not at home ;)
[21:14] <asac> i would suggest to setup a long living chroot if you want your main system to be virgin like
[21:14] <tonyyarusso> Yeah - would have to learn how to do that too :)
=== rzr is now known as RzR
[21:15] <asac> heh ;) https://wiki.ubuntu.com/DebootstrapChroot
[21:16] <asac> take care that once you mount bind stuff like your home you shouldn just rm -rf /path/to/chroot
[21:16] <asac> otherwise you will wipe your complete home ;)
[21:16] <asac> just because i ended up in that situation once ;)
[21:16] <tonyyarusso> icky
[21:16] <asac> but usually one doenst need to wipe a chroot ... so just remember to check that nothing is bind mounted in it ;)
[21:16] <asac> when you remove it at some point
[21:17] <tonyyarusso> how would you check that? run mount within it?
[21:25] <plun> hello... fta around ?
[21:26] <fta> plun, yes
[21:26] <plun> The test shell works !!! ;)
[21:26] <plun> I can also surf.... ;)
[21:27] <fta> good, but it's not supposed to, yet.
[21:27] <fta> i'm still working on it
[21:27] <plun> I got a update earlier this evening and checked it.... ;)
[21:28] <fta> yep, those are my test builds
[21:30] <fta> it's supposed to start like that: http://www.sofaraway.org/ubuntu/tmp/chromium.png
[21:30] <fta> not with an empty page like my last package
[21:30] <fta> plun, ^^
[21:31] <plun> Aha... I started with the test_shell script.... http://ubuntu-pics.de/bild/6589/screen_gD7pEY.png
[21:32] <plun> Your page also works.... ;)
[21:32] <fta> plun, don't expect a full browser anytime soon, upstream is still far behind
[21:33] <plun> Yup... this is just a happening... ;)
[21:34] <plun> Well it also crashes...
[21:34] <plun> Failed reading: /usr/lib/chromium-browser/../../webkit/tools/test_shell/resources/missingImage.gif
[21:34] <plun> Trace/breakpoint trap
[21:35] <fta> yep, that's what i'm working on
[21:36] <fta> (the evil "../../webkit")
[21:36] <plun> Okidok.... "playing" with this, just great anyway !! Good night
[21:36] <fta> thanks
[21:36] <fta> n8
[21:50] <tonyyarusso> asac: Okay, so I seem to have created a dbgsym package and installed it. Now what?
[21:53] <asac> tonyyarusso: reproduce crash
[21:53] <asac> get backtrace
[22:37] <fta> hm, ff3.1 crashed
[22:38] <fta> strangely, it didn't ask me which tabs I wanted to restore, it restored them all, like before
[22:38] <fta> another backout ?
[22:39] <fta> mconnor, ^^ ?
[22:41] <fta> asac, http://paste.ubuntu.com/78474/
[22:42] <mconnor> fta: mmm, iirc the plan was to only prompt if there were two crashes in swift succession
[22:43] <fta> hmm... ok. what about ctrl+tab ? is it gone in b2 ?
[22:45] <mconnor> yes
[22:45] <mconnor> in 3.1, really
[22:46] <fta> really? why?
[22:47] <fta> should i switch back to the addon then?
[22:47] <fta> mconnor, with ~150 tabs, the tab list is not usable
[22:48] <mconnor> fta: the UI wasn't good enough to ship
[22:49] <fta> mconnor, i see. is it completely gone or just prefed out ?
[22:49] <mconnor> it's selectively excised via #ifdef
[22:50] <mconnor> the plan is to get it in 3.2
[22:50] <mconnor> we just didn't like the way the interaction was working, and we weren't going to slip the release for it
[22:50] <fta> damn, ok then. i'll switch back to the addon then.
[22:50] <mconnor> you're still way way way out on the long tail :)
[22:51] <fta> when i 3.1 final planed?
[22:51] <fta> is