[01:37] <Sarvatt> ahh crap, xserver master needs pixman >= 0.15.12 now too
[03:04] <Sarvatt> oookay, everything but wacom and joystick updated for xinput abi 7 here https://edge.launchpad.net/~sarvatt/+archive/xorg-testing
[03:04]  * Sarvatt expects to wake up to xinput abi 8 tomorrow after all that :D
[03:26] <Sarvatt> hmm, some fairly nasty problems with it vs the 0609 one. firefox 3.6 giving a BadWindow (invalid Window parameter) error, and gnome panel themes wont apply.
[03:35] <Sarvatt> maybe i didnt update pixman properly
[07:47] <Sarvatt> ah hah, gnome-settings-daemon is failing loading the keyboard plugin, if i strace it everything works right even though it crashes until i interrupt it
[07:58] <Sarvatt> i take that back, it was the gpointing-device-settings g-s-d plugin not working right with the latest XI2 driver. all fixed now.
[07:59] <Sarvatt> phew, gnome sure was ugly without themes working :D
[08:27] <hyperair> heh i can imagine =p
[10:07] <Unggnu> hi all
[10:07] <Unggnu> Is there already the X testing live cd with kms and so on?
[11:27] <hyperair> Sarvatt: i just noticed that with the latest mesa, killing compiz recovers a considerable amount of memory!
[11:27] <hyperair> it hit 1.4G, and when i killed compiz, it dropped to 0.4
[17:19] <Sarvatt> hyperair: let me know if you have any luck with the mesa thats building right now, cherry-picked the two compiz memory leak fixes from mesa_7_5_branch into 7.6
[17:20] <hyperair> Sarvatt: cool =p
[17:25] <Sarvatt> hopefully it wont take 11 retries to build right on amd64 thistime :D
[17:29] <hyperair> heheh did it take 11 retries the previous time? =p
[17:36] <Sarvatt> theres a race in the recursive make install only on amd64 (well sometimes lpia too but not 90% failures like amd64) and only on karmic for some reason. it started around when they changed the osmesa build process and coreutils was upgraded
[17:37] <Sarvatt> i tried changing it to $(MAKE) -j1 -f debian/rules $(INSTALL_TARGETS) but no luck
[17:42] <hyperair> hmm strange O_o
[17:43] <hyperair> this is why recursive make sucks =(
[17:43] <hyperair> did you bring this issue upstream?
[17:45] <Sarvatt> if i knew exactly what the problem was i would, it doesnt happen on the jaunty PPAs so i'm not sure exactly what the problem is
[17:45] <Sarvatt> oh wow built ok that time
[17:47] <Sarvatt> failed on lpia on the xorg-testing ppa https://edge.launchpad.net/~sarvatt/+archive/xorg-testing/+build/1086480/+files/buildlog_ubuntu-karmic-lpia.mesa_7.6.0~git20090621.df70d304-0ubuntu0sarvatt2_FAILEDTOBUILD.txt.gz
[17:50] <hyperair> oh it built already?
[17:50] <hyperair> i'll go upgrade then =p
[18:02] <Sarvatt> and amd64 failed on the radeon-kms ppa, sheesh
[18:21] <Sarvatt> mutter is sweet, cant wait till its packaged
[18:21] <Sarvatt> byebye compiz :D
[18:53] <cwillu> mutter=?
[18:56] <cwillu> Sarvatt, don't suppose you have any interest in whipping me up a 2.6.30 or 31 with 79f11c19a396e8cea7dad322dcfb46c0a8517fe6 reverted? :)
[18:57] <cwillu> a comment on http://bugs.freedesktop.org/show_bug.cgi?id=20520 suggests that it may help with suspend + kms
[18:57] <Sarvatt> its metacity with a clutter based opengl compositing backend instead of the crappy xrender one metacity compositing uses
[18:58] <cwillu> intriguing
[18:58] <Sarvatt> cwillu, have you tried a kernel before that was applied? i think that was in .30 rc2 or rc4
[18:59] <Sarvatt> could try the mainline daily kernel for rc2 out
[18:59] <cwillu> well, 2.6.29's been working for me.  Give me a sec, I'll check if I had a <=rc4 installed
[19:00] <Sarvatt> err not daily sorry
[19:00] <Sarvatt> just mainline 2.6.30-rc2
[19:00] <Sarvatt> finding out what rc it was in now, april 30th would put it around rc5 timeframe
[19:01] <Sarvatt> but i think it was right after rc5 released so its in rc6
[19:01] <cwillu> earliest I have installed is rc5, so I'll try an earlier one
[19:03]  * cwillu downloads the -rc2 mainline
[19:03] <Sarvatt> cant revert it from master right now at least
[19:08] <Sarvatt> i dont see that commit in rc5 if you have it
[19:08] <cwillu> okay
[19:08] <cwillu> hey, what was the url to that 2.6.31 you had again?
[19:09] <Sarvatt> http://sarvatt.com/downloads/2.6.31-pre-rc1/
[19:09] <cwillu> thanks
[19:18]  * cwillu reboots into rc5
[19:38] <Sarvatt> any luck cwillu?
[19:38] <cwillu> mild confusion on vt switching that I just got figured out (fbcon wasn't loaded)
[19:38] <cwillu> trying a suspend right now
[19:39] <cwillu> looks good so far
[19:39] <cwillu> usually it would have locked up by now
[19:39] <cwillu> (this is -rc2)
[19:39] <Sarvatt> yeah darn mainline kernels using jaunty's config causing that
[19:40] <cwillu> yep, rc2 is definitely looking good on the kms suspend front
[19:40] <cwillu> I'll check that 2.6.31-pre-rc1 as well
[19:40] <cwillu> yay for having two different suspend bugs :p
[19:45] <Sarvatt> it was added in rc6 btw so rc5 should be fine too
[19:46] <cwillu> 2.6.31-pre-rc1 hangs immediately on resume
[19:46] <cwillu> good ol' working mouse cursor that you can't do anything with :p
[19:49] <cwillu> give me a sec, I'll check rc5
[19:50] <Sarvatt> well hopefully something will come of it now that they know that was the problem :)
[19:50] <Sarvatt> is it 945?
[19:50] <cwillu> yes
[19:50] <cwillu> I'll add a 'works for me too' on the bug after I make sure rc5 is good