[00:36] <bryce> jbarnes: is the gpu dumping stuff particular to 9xx?  is it at all worth trying for 8xx freeze bugs?
[00:40] <jbarnes> so far it just decodes 9xx stuff mainly
[11:44] <seb128> bryce_: did you try to trigger the intel crash with a virtual config?
[11:45] <seb128> there has been 2 new comments from users who don't have the issue and have virtual settings today
[18:58] <hyperair> is there a reason i965 has been blacklisted in jaunty?
[18:59] <hyperair> i commented out the line in the compiz loading script and it seems to run fine
[19:00] <mnemo> hyperair: if you do "aptitude changlog compiz" maybe there is a comment about the reason on that change?
[19:01] <hyperair> mnemo: i'll poke it later when i'm back in jaunty.
[19:01] <hyperair> the installer hung on me.
[19:01] <hyperair> bah
[19:01] <mnemo> ;(
[19:01] <hyperair> well then
[19:01] <hyperair> off i go
[19:01] <hyperair> with another attempt with the usb startup thing
[19:21] <jbarnes> apw: get a chance to put together those mchbar patches yet?
[19:47] <bryce> http://people.ubuntu.com/~bryce/drivers.svg
[19:51] <jcastro> bryce: that is ouchy
[19:56] <bryce> jcastro: yeah
[20:04] <paran> but that is a nice graph.. I didn't know gnuplot could output SVG :)
[20:45] <jbarnes_> bryce: those trends are pretty scary... all of the used drivers have positive slopes
[20:47] <bryce> jbarnes: yep
[20:48] <bryce> well, except for -nv and -openchrome
[20:48] <jbarnes> yeah that's why I qualified it with "used" :)
[20:48] <bryce> heh
[20:48] <jbarnes> we're clearly in the worst shape though... and only partly because there are so many intel chips out there
[20:49] <bryce> -ati at least has fewer bugs now than it did at the start of  the period
[20:49] <jbarnes> yeah that's nice
[20:50] <jbarnes> I wonder how many ati bug reports are supressed when people just move to the proprietary driver though
[20:50] <bryce> actually for jaunty, we've seen the reverse situation
[20:50] <jbarnes> if they find a bug in fglrx I wonder if they report it directly to ati or just live with it too
[20:50] <jbarnes> oh really?
[20:51] <bryce> -fglrx was not available for the longest time, so people were using -ati, and then when we finally did get it, they dropped support for all pre-R6xx chips
[20:51] <jbarnes> oh right
[20:51] <bryce> so this cycle a LOT of people have been forced to shift to -ati.
[20:52] <bryce> also the functionality gap between -ati and -fglrx is a lot narrower than between -nv and -nvidia
[20:52] <jbarnes> yeah
[20:52] <bryce> I've given up on -nv pretty much.  I think in karmic we're going to switch to -nouveau like fedora has
[20:53] <jbarnes> yeah might be in ok shape by then... the fedora guys have been putting a ton of work into it
[20:54] <bryce> with -intel, I'm a bit concerned in that with the EXA -> UXA transition, is upstream going to care any longer about bugs when using EXA?  If not, what do we do with all these existing bug reports?  wontfix them?
[20:55] <bryce> I guess, once we have new -intel bits in place, we'll ask people to retest against that, and close the bugs if they can't reproduced there
[20:57] <jbarnes> yeah that's one option
[20:57] <jbarnes> we'll continue to do 2.7.x releases though
[20:58] <jbarnes> which means asking users to try uxa might become more viable over time with juanty, since the driver has it and it should get more usable over the 2.7.x period
[21:01] <bryce> that's good to hear
[21:01] <bryce> I'd like in karmic to retain the option of switching back to exa for users who have trouble with uxa