2012/07/23/#kubuntu-devel.txt

=== ronnoc_ is now known as ronnoc
[08:04] <Riddell> morning
[11:53] <BluesKaj> Hey all
=== shadeslayer is now known as shadeslayer_
=== shadeslayer_ is now known as shadeslayer
[12:31] <shadeslayer> Ok, I'm going to get cracking on the RC
[12:32]  * shadeslayer looks at debfx's automation stuff
[12:34] <jussi> shadeslayer: do you know which ppa has a reasonably stable, but reasonably recent version of Calligra?
[12:35] <Riddell> shadeslayer: he threw bits into kubuntu-ninjas but had to wait for it to be in the rightorder
[12:35] <Riddell> jussi: see kubuntu.org
[12:35] <jussi> Riddell: thanks
[12:36] <shadeslayer> Riddell: I don't see anything related to 4.8.97 in the ninjas ppa
[12:36] <Riddell> shadeslayer: I'm thinking we should just diverge from debian and drop the kde-sc-dev-latest (>= 4:4.8.90) stuff in favour of versioned depends on kdelibs5-dev (>= 4:4.8)
[12:36] <jussi> Riddell: ahh thanks...
[12:36] <shadeslayer> Riddell: why?
[12:36] <Riddell> shadeslayer: that would mean it can just be thrown into the PPA without worrying about the build order, launchpad should work out the build order
[12:37] <shadeslayer> really? :O
[12:37] <shadeslayer> why can't it do it right now?
[12:37] <Riddell> shadeslayer: because launchpad's way of working out when to rebuild isn't as complete as debian's
[12:37] <Riddell> there's a bug about it
[12:38] <Riddell> but if you upload kde4libs and a kdegames using kde-sc-dev-latest you'll need to retry kdegames after kde4libs is done
[12:38] <Riddell> which you wouldn't if we just did simple kdelibs5-dev (>= 4:4.8.90)
[12:39] <Riddell> shadeslayer: oh and let me know if you need an ec2
[12:39] <shadeslayer> nah, have a VPS now :P
[12:39] <shadeslayer> will poke if I happen to run out of memory
[12:40] <Riddell> VPS?
[12:40] <shadeslayer> Riddell: Virtual Private server :)
[12:41] <Riddell> shadeslayer: you hired one from somewhere?
[12:41] <shadeslayer> yeah, 15 USD for a year
[12:42] <shadeslayer> just about enough for building stuff like KDE Telepathy and smaller packages
[12:42] <shadeslayer> need to script kde telepathy packages though
[12:42]  * shadeslayer adds to TODO
[12:43] <shadeslayer> apachelogger: ping me when you're about
[12:45] <shadeslayer> Riddell: I was thinking of writing a python script that queries launchpad for build state of KDE SC packages, then if they're in a build wait state, and if the build dep was alredy built, it retries the build
[12:45] <shadeslayer> the latter part is already done
[12:45] <shadeslayer> the 'monitor' bit needs work
[12:45] <Riddell> shadeslayer: that seems r
[12:46] <shadeslayer> r? :P
[12:46] <Riddell> shadeslayer: that seems pretty faffy compared to just working with launchpad and doing versioned build-depends like it expects
[12:46] <debfx> shadeslayer: that's already implemented in https://bazaar.launchpad.net/~debfx/+junk/kubuntu-automation/view/head:/kubuntu-retry-builds
[12:46] <shadeslayer> oh ... awesome
[12:47] <shadeslayer> I'm not fond of the idea of diverging from debian .....
[12:47] <shadeslayer> even though we change a couple of lines, merging stuff from debian later on becomes iffy because of a larger diff
[12:48] <debfx> ideally someone would just fix that launchpad bug :P
[12:48] <shadeslayer> :P
[12:48] <shadeslayer> can you point me to the bug?
[12:48] <yofel_> we're talking about build-deps here, we have to merge that anyway - and you wouldn't have to worry about merging kde-sc-dev-latest anymore
=== yofel_ is now known as yofel
[12:48] <shadeslayer> I'm don't exactly understand what the problem is ...
[12:49] <yofel> shadeslayer: builds failing instead of being put into dep-wait
[12:49] <shadeslayer> because it finds that kde-sc-dev-latest was built but kdelibs5-dev wasn't
[12:49] <yofel> as launchpad ignores the Breaks from kde-sc-dev-latest
[12:49] <shadeslayer> oh
[12:50] <yofel> I once tried to do that, but back then it wasn't scripted and too much work, if you just put that in a script I would prefer using versioned depends over kde-sc-dev-latest
[13:18] <shadeslayer> hmm
=== shadeslayer is now known as shadeslayer_
[13:18] <shadeslayer_> something is horribly broken between my router and my modem it seems
=== shadeslayer_ is now known as shadeslayer
[13:18] <shadeslayer> anyway, drop kde-sc-dev-latest then?
[13:19] <shadeslayer> in favor of versioned build depends on kdelibs5-dev
[13:19] <shadeslayer> ( will need modifications in debfx's scripts as well then )
[13:20] <debfx> there's not only kde4libs5-dev
[13:21] <Riddell> dropping kde-sc-dev-latest gets my vote
[13:22] <Riddell> but yes it'll need everything else in kde-sc-dev-latest used by the source package as a versioned build-depened for all the packages
[13:22] <shadeslayer> right
[13:27] <shadeslayer> seems like alot of work .... ah well ..
[13:29] <debfx> the code to bump the kde-sc-dev-latest build-dep version is already there
[13:29] <debfx> so it's mostly a matter of keeping the package list synchronized with kde-sc-dev-latest
[13:31] <shadeslayer> *nod* but I'll have to version unversioned build depends, add kdelibs5-dev etc to most packages right?
[13:32] <Riddell> yes, for the kde sc packages
[13:33] <shadeslayer> yeah, so the initial work that needs to be done is quite alot :P
[13:33] <ScottK> If you're trying to get to build-dep instead of failed, dropping kde-sc-dev-latest won't help much.
[13:34] <Riddell> ScottK: why not?
[13:34] <ScottK> Because the first time another build-dep is uninstallable on a slow arch, if fails.
[13:35] <ScottK> Think of ~everything that depends on pimlibs for example.
[13:36] <ScottK> You'd have to drop kde-sc-dev-lastest and do the trick I used to do of uploading the packages backwards to do it.
[13:36] <ScottK> Once a package is in depwait, it'll go back to it.  The trick is getting it there first.
[13:37] <yofel> shouldn't it be put into dep-wait with the versioned build-deps?
[13:40] <ScottK> If the necessary version is not available, it'll depwait.  If the dependency is not installable, it'll FTBFS.
[13:40] <ScottK> That's why the reverse upload trick works.
[13:40] <debfx> ScottK: why would kdepimlibs be uninstallable after it built?
[13:40] <ScottK> debfx: It'll be built on some archs and not others.
[13:41] <ScottK> Maybe we changed enough -dev packages to arch any it's not such a big deal anymore.
[13:41] <debfx> afaik all -dev packages are arch:any
[13:42] <Riddell> -dev packages should always be arch any, header files can differ between arches
[13:42] <debfx> it could only be a problem if a library has a strict dependencies on a data package that is arch:all
[13:42]  * Riddell uploads some test stuff to ~jr ppa to see
[13:44] <debfx> e.g. libkdeedu-dev
[13:55] <Riddell> successful dep waot 
[13:55] <Riddell> dep wait https://launchpad.net/~jr/+archive/ppa/+build/3676000
[14:15] <shadeslayer> hmm ...
[14:19] <highvoltage> Darkwing: hey there, are you around?
[15:26] <Riddell> hmm well it doesn't magically retry the package https://launchpad.net/~jr/+archive/ppa/+build/3675999
[15:26] <Riddell> I find that most strange
[15:26] <micahg> hrm, there's no ${calligra:Enhances} in calligra-l10n...I'm wondering if I should just demote to suggests for alpha3 and fix the enhances later
[15:27] <Riddell> micahg: yeah that's fine too
[15:34] <Riddell> ScottK: csync-owncloud and owncloud-client in New should you be in the mood
[15:35] <ScottK> OK.
[15:38] <kubotu> ::qt-bugs:: [1022479] bzr crashed with SIGSEGV in QGtkStyle::drawComplexControl() @ https://bugs.launchpad.net/bugs/1022479 (by Logan Rosen)
[15:38] <ubottu> Launchpad bug 1022479 in bzr (Ubuntu) "bzr crashed with SIGSEGV in QGtkStyle::drawComplexControl()" [Medium,Confirmed]
=== nowardev2 is now known as Peace-
=== rdieter_laptop is now known as rdieter
=== cmagina-lunch is now known as cmagina
[15:54] <shadeslayer> Riddell: did you retry kdepimlibs manually?
[15:55] <Riddell> on i386 yes
[15:55] <Riddell> not on amd64
[15:55] <shadeslayer> ok, amd64 seems to have started 
[15:55] <Riddell> ooh it's started
[15:55] <shadeslayer> yeah
[15:55] <Riddell> ok so just launchpad being slow
[15:56] <shadeslayer> yeah
[15:56] <shadeslayer> Riddell: so, shall we start switching deps?
[15:58] <Riddell> shadeslayer: it still has my preference
[15:59] <ScottK> I'd try one with the arch all -common package first.
[15:59] <ScottK> txwikinger is in Canada.  He should just go make infinity implement the BD-Uninstallable state in soyuz.
[15:59] <ScottK> That's be easiest.
[16:00] <ScottK> Well, for everyone except txwikinger.
[16:04] <shadeslayer> hahaha
[16:07] <ScottK> I just gave infinity a ping.
[17:43] <kubotu> ::qt-bugs:: [1028039] package libqtgui4 4:4.8.1-0ubuntu4.1 failed to install/upgrade: short read on buffer copy ... @ https://bugs.launchpad.net/bugs/1028039 (by dragan juric)
[17:43] <ubottu> Launchpad bug 1028039 in qt4-x11 (Ubuntu) "package libqtgui4 4:4.8.1-0ubuntu4.1 failed to install/upgrade: short read on buffer copy for backend dpkg-deb during `./usr/lib/x86_64-linux-gnu/libQtGui.so.4.8.1'" [Undecided,New]
[18:14] <micahg> \o/, no longer getting calligra on upgrade without kubuntu-desktop/calligra previously installed
[18:40] <Riddell> thanks micahg 
[19:51] <kubotu> ::qt-bugs:: [1013868] Crashes at any attempt to open a file @ https://bugs.launchpad.net/bugs/1013868 (by Thomas Malenfant)
[19:51] <ubottu> Launchpad bug 1013868 in qt4-x11 (Ubuntu) "Crashes at any attempt to open a file" [Undecided,New]
[20:03] <shadeslayer> Riddell: ping
[20:05] <shadeslayer> Riddell: I could use a AWS instance about now
[20:14] <Riddell> shadeslayer: hi
[20:14] <Riddell> shadeslayer: what sort?
[20:14] <Riddell> 32 or 64?
[20:14] <Riddell> precise or quantal?
[20:14] <shadeslayer> quantal, preferably on SSD
[20:14] <Riddell> mm I've no idea if they do SSD
[20:15] <shadeslayer> iirc they introduced new options
=== rdieter_laptop is now known as rdieter
[20:15] <Riddell> "High I/O Quadruple Extra Large Instance"
[20:15] <Riddell> sounds expensive
[20:16] <shadeslayer> gah
[20:16] <shadeslayer> meh, leave it then
[20:16] <shadeslayer> The I/O time is not worth it
[20:19]  * shadeslayer will get more resources on his VPS when stock is available
[20:21] <Riddell> $3.10 an hour, will get very expensive very quickly
[20:21] <shadeslayer> ouch
[20:21] <shadeslayer> not worth it then
[20:22] <Riddell> lots at $0.08 an hour would be preferable :)
[20:22] <shadeslayer> :D
=== rdieter_laptop is now known as rdieter
[20:57] <kubotu> ::qt-bugs:: [1028039] package libqtgui4 4:4.8.1-0ubuntu4.1 failed to install/upgrade: short read on buffer copy ... @ https://bugs.launchpad.net/bugs/1028039 (by dragan juric)
[20:57] <ubottu> Launchpad bug 1028039 in qt4-x11 (Ubuntu) "package libqtgui4 4:4.8.1-0ubuntu4.1 failed to install/upgrade: short read on buffer copy for backend dpkg-deb during `./usr/lib/x86_64-linux-gnu/libQtGui.so.4.8.1'" [Undecided,Invalid]
=== skreech__ is now known as Daskreech
[22:35] <kubotu> ::workspace-bugs:: [1028163] plasma-desktop crashed with SIGSEGV when changing battery monitor settings @ https://bugs.launchpad.net/bugs/1028163 (by johannes)
[22:35] <ubottu> Launchpad bug 1028163 in kde-workspace (Ubuntu) "plasma-desktop crashed with SIGSEGV when changing battery monitor settings" [Undecided,Invalid]