[10:37] <jtaylor> I'm looking at shiboken bug 1070772
[10:37] <jtaylor> would it be acceptable to put in a memleak in exchange for not crashing?
[10:37] <jtaylor> the code looks aweful
[10:38] <jtaylor> it deletes stuff in an object after it decrefs it to zero, but thats not even the cause of the segfault
[10:58]  * mitya57 doesn't like the idea, but if there is no alternative...
[11:00] <jtaylor> btw the issue occurs with all python versions not just 3
[11:00] <jtaylor> so its kind of imporant
[13:03] <jtaylor> great fixing another bug requires messing around with its code generator ._.
[13:03] <jtaylor> thats always fun
[13:26] <jtaylor> success!
[13:26] <jtaylor> but does adding a new optional parameter to a private function in c++ break abi?
[15:00] <cjohnston> jtaylor: there seems to be two different issues with shiboken.. I spent hours trying to sort out ftbfs, plus there is the test segfault.
[15:01] <jtaylor> cjohnston: I got a working build going
[15:01] <cjohnston> oh cool
[15:01] <jtaylor> but with an intentional memleak
[15:01] <cjohnston> :-/
[15:01] <jtaylor> debdiff is in the bug
[15:01] <jtaylor> just building pyside if it still works
[15:02] <jtaylor> seems to, but there is another problem with phonon packages ..
[15:02] <cjohnston> jtaylor: they really need to get cmake out of the upstream tree and use the installed version
[15:03] <jtaylor> they embed cmake? :O
[15:04] <jtaylor> didn't notice
[15:04] <jtaylor> thats aweful
[15:04] <cjohnston> https://qt.gitorious.org/pyside/shiboken/trees/4aa901db487cc1202680046091f074ea58544ea3/cmake/Modules
[15:04] <cjohnston> I don't know that they did the whole thing, but they did some stuff ^^ which are not completely fixed.
[15:04] <cjohnston> cmake itself is AFAIK
[15:08] <cjohnston> If some security problems in a package have been fixed in debian, what is the proper way to bring them to Ubuntu?
[15:09] <jtaylor> https://wiki.ubuntu.com/SecurityTeam/UpdatePreparation
[15:10] <jtaylor> :( pyside testsuite fails
[15:10] <cjohnston> :-/
[15:11] <jtaylor> but not too bad, one test detects the memleak, an expected phonon issue and one crash
[15:11] <jtaylor> unfortunately crash in shiboken ._.
[15:12] <cjohnston> So it looks like even if its fixed in debian you still have to do the patches in Ubuntu, you can't just bring the fixes over from debian
[15:12] <cjohnston> :-/
[15:13] <jtaylor> hm the tests are ignored, maybe I shouldn't have aborted :)
[15:13] <jtaylor> cjohnston: if the version in squeeze and ubuntu are the same you can just sync
[15:14] <cjohnston> It's django.. squeeze still has 1.2.3
[15:17] <jtaylor> django is main
[15:17] <jtaylor> the security team should deal with that
[15:18] <jtaylor> #ubuntu-hardenend if you want to help
[15:19] <cjohnston> I figured it's the weekend so I'd play a little..
[15:19] <cjohnston> they confirmed the bug, but haven't done anything yet
[15:19] <jtaylor> you should still contact them before you start working on it to avoid duplication
[15:20] <jtaylor> -hardened (or howevery you spell that word) is the right place for main and universe security issues
[15:20] <cjohnston> ok
[19:32] <Gaming4JC> Minor problem in this package: http://packages.ubuntu.com/quantal/slingshot -- Depends on FreeSansBold.ttf which is found in package ttf-freefont if I recollect.
[19:32] <Gaming4JC> someone should fix the dependency issue because the game crashes with "File "/usr/games/slingshot", ...Settings.font = pygame.font.Font(get_data_path("FreeSansBold.ttf"), 14) IOError: unable to read font filename"
[19:33] <Gaming4JC> I'd file a bug report excepting for the fact https://launchpad.net/slingshot is a different package ;)
[19:33] <Gaming4JC> let me keep digging around though, I'm sure it's on launchpad
[19:35] <lifeless> Gaming4JC: launchpad.net/ubuntu/+source/slingshot perhaps?
[19:35] <Gaming4JC> that looks right, thanks.
[20:22] <quidnunc> How do I confirm which version pdebuild is using?
[20:25] <jtaylor> version of what?
[20:26] <quidnunc> release
[20:26] <quidnunc> not version
[20:26] <jtaylor> look at the etc/os-release
[20:26] <quidnunc> I am getting "Dependency is not satisfiable: dpkg-dev (>= 1.16.1)" but that is not possible on quantal
[20:27] <quidnunc> jtaylor: Do I need to run a hook when running pdebuild to do that?
[20:27] <jtaylor> no just look at the file
[20:28] <jtaylor> pbuilder has its chroot in a tarball, cowbuilder its just in a folder
[20:28] <jtaylor> somewhere in /var/cache/pbuilder
[20:29] <quidnunc> jtaylor: I need to untar the chroot?
[20:29] <jtaylor> yes e.g. with file-roller
[20:29] <jtaylor> or you just login and look
[20:30] <TheLordOfTime> jtaylor, can't they use pbuilder to get into the chroot and run a shell?
[20:30] <quidnunc> jtaylor: I don't have /etc/os-release on login
[20:30] <jtaylor> hm
[20:30] <TheLordOfTime> or did pbuilder remove that since Lucid?
[20:30] <jtaylor> then look at apt/sources.list
[20:30] <jtaylor> TheLordOfTime: you can but its probably slower
[20:30] <quidnunc> jtaylor: Yeah, it's the wrong release. But why?
[20:31] <jtaylor> quidnunc: how did you create it?
[20:31] <quidnunc> jtaylor: pbuilder --create
[20:31] <jtaylor> pbuilder-dist raring amd64 create
[20:31] <jtaylor> quidnunc: that will create with some distribution specific default, probably the one your on
[20:31] <jtaylor> quidnunc: use DISTRIBUTION DIST or --distribution to change it, or more convinient pbuilder-dist
[20:31] <quidnunc> jtaylor: Which is what I want
[20:31] <jtaylor> or cowbuilder-dsit
[20:33] <quidnunc> jtaylor: Will pbuilder-dist overwrite the current chroot tarball?
[20:33] <jtaylor> no
[20:33] <jtaylor> it uses ~/pbuilder
[20:36] <quidnunc> jtaylor: How do I purge all my previous pbuilder environments?
[20:36] <jtaylor> delete the tarball from /var/cache/pbuilder
[20:56] <quidnunc> it turned out to be /etc/pbuilderrc. I had updated ~/.pbuilderrc
[21:24] <quidnunc> Can I not use backportpackage on packages from universe?
[21:27] <tumbleweed> sure you can
[21:28] <tumbleweed> your build environment will need to have universe enabled, of course
[21:31] <quidnunc> tumbleweed: On pbuilder do I have to recreate the base tarball after changing COMPONENTS?
[21:32] <tumbleweed> yes
[21:32] <tumbleweed> well, do an update
[21:32] <tumbleweed> with --override-config
[21:33] <quidnunc> tumbleweed: thanks
[21:37] <quidnunc> tumbleweed: Wait, backportpackage is complaining about raring. But my pbuilder target env is quantal. Why does my *build environment* need universe enabled?
[21:37] <quidnunc> 'llvm-3.2-dev' does not exist in the Ubuntu primary archive in raring...
[21:39] <tumbleweed> quidnunc: please paste the output you are seeing - I can't see what you can
[21:39] <quidnunc> I guess llvm-3.2-dev is not the source package name. In this case it is obvious but in general how do I map a package back to its source package name?
[21:39] <tumbleweed> paste.ubuntu.com
[21:48] <tumbleweed> quidnunc: apt-cache show package | grep Source
[21:49] <quidnunc> tumbleweed: thanks
[23:34] <quidnunc> How do I get the name of a package's source package when it is not in my /etc/apt/sources.list ?
[23:34] <quidnunc> e.g. a package in raring that is not in quantal when I am developing on quantal
[23:34] <jtaylor> I usually use the debian pts
[23:35] <jtaylor> d package-name with an appropriate keyword search
[23:35] <jtaylor> http://packages.qa.debian.org/common/index.html?src=%s
[23:46] <lifeless> quidnunc: I have sources.list entries for deb-src for all releases I'm interested in.