[00:33] <vbrummond> hi I have a question about the stable updates X.org ppa, it seems that the xorg version is higher in Ubuntu Lucid than in the PPA I wondered why
[00:34] <RAOF> I havn't checked, but I'd guess it's because the stable updates PPA grabs the drivers.
[00:34] <vbrummond> I see
[00:34] <vbrummond> no wait, its higher.. ;/ odd I cant seem to update to it, ill figure it out
[00:35] <RAOF> apw: I haven't looked at that bug since I saw it, saw that Colin and you appeared to be on it, and that it looked decidedly kernel-ish.
[00:40] <vbrummond> i think I see I am a bit confused because of metapackages, xserver-xorg-core says a newer version than the ppa is installed, but its essentially the same version
[01:23] <Sarvatt> whoa, 2.6.36-rc1 is an insane amount more responsive than .35
[01:23] <Sarvatt> apw: thanks for patching out comedi so it built!
[01:24] <Sarvatt> thats crazy, updating 31 packages extracting 600mb of lzma archives and the gnome-do dock is still instant
[02:53] <Sarvatt> well my mind is blown, nothing I do with 2.6.36-rc1 slows down the desktop. it hasn't responded anywhere remotely this good since 6.06
[03:04] <Sarvatt> hmm, jockey could probably use an update to offer libgl1-mesa-dri-experimental, trying to activate effects says its not possible unless the nvidia drivers are installed still but works when you cancel it
[03:20] <warewolf> Sarvatt: poke
[03:36] <RAOF> Sarvatt: I don't think we particularly want to offer libgl1-mesa-experimental; does that occur when libgl1-mesa-experimental is installed already?
[03:55] <bjsnider> Sarvatt, why is it more responsive, is it because of those disk activity patches?
[03:55] <bjsnider> is it the scheduler?
[07:38] <knittl> hi guys. my display stays blank after standby. if rhythmbox was playing it plays, but stops after the current song
[13:30] <Sarvatt> darn, 12 hours later i'm getting some heavy swapping going on and it killed the performance, knew it was too good to be true :)
[16:14] <barry> hi folks.  what's the current best recommendation for ati support on maverick?  i have a radeon hd 4670 and cannot install fglrx (bug 573748) from main archive, or the swat or edgers ppa.  the free driver gives me a usable dual-headed desktop, but no gl.
[16:16] <jcristau> "no gl"?
[16:17] <barry> glxgears says: "extension GLX missing on display :0.0"  also no "visual effects" i.e. no compiz
[16:18] <jcristau> i suspect something in your install is broken then
[16:19] <barry> jcristau: i'm sure it is, but i can't suss it out ;).  i was working with #ubuntu+1 yesterday to try many different combinations, without luck.  otoh, it seems like bug 573748 is affecting many other people
[16:20] <barry> but i guess my first question is: should i be using edgers and/or swat ppa or just the main archive?  i think i ran lucid entirely off of edgers, but i'd like to avoid that for maverick if possible
[16:45] <Sarvatt> barry: fglrx doesn't work with maverick and probably wont until the very last minute before release, you need to purge the fglrx package you probably have lying around
[16:58] <barry> Sarvatt: cool, will do.  does it help at all to use the ppas?  i don't mind helping to test things as long as i can get some help when the desktop breaks horribly :)
[17:00] <minimec> barry: the x-edgers ppa has a package called purge-ppa. Be sure to install that one ;)
[17:01] <Sarvatt> nope, but the r600 mesa driver is much better in xorg-edgers than whats in maverick if you find it too slow. ati is notoriously bad about supporting new stuff in fglrx, from what I've read 10.10 will be the first release with any chance of supporting xserver 1.9 and 10.8-10.9 are just supposed to be minor bugfix releases
[17:01] <barry> minimec: oh yes, purge-ppa is my favorite :)
[17:01] <Sarvatt> ppa-purge is in maverick and lucid-backports as well now :)
[17:01] <Sarvatt> (thanks BlackZ!)
[17:02] <barry> Sarvatt: cool thanks.  i think i'll stick with the main archives then and just wait.  thanks for the info!
[17:10] <Sarvatt> heads up to anyone using edgers on intel, if you have a ~/.drirc you need to delete it or manually enable ARB_fragment_shader support in it
[17:11] <Sarvatt> well the people using i915_dri.so that is
[21:13] <Sarvatt> uhoh, nouveau getting ready for a new api? hopefully just declaring 0.0.16 as stable 1.0.0 :) http://cgit.freedesktop.org/mesa/drm/commit/?id=b61e81a191d3a5c269c5f7c40199aebc9ebc034c
[23:12] <penguin42> Have people run the X test suite (xts) recently - and is it normal for it give a load of fails?
[23:12] <penguin42> (I've never run it before so I don't know if it's actually trustworthy)
[23:17] <jcristau> yes
[23:17] <jcristau> and yes
[23:18] <penguin42> hmm thanks for the reply; is there any trustworthy test then that actually checks rendering?
[23:18] <jcristau> what are you trying to do?
[23:20] <penguin42> jcristau: I've got a reliable bug where any 3d rendering on the cirrus card emulated inside kvm is gently screwed up in an odd way - it looks like it's wrongly addressing or something; given that it won't actually be trying to do hardware 3d (because the card doesn't have any) I wondered if there was a way to narrow it down to maybe which operation was broken
[23:20] <jcristau> see, that's a more useful question.
[23:21] <penguin42> jcristau: Well true; but it would be nice if xts worked and passed!
[23:21] <jcristau> for gl, i'd try piglit
[23:22] <penguin42> any 2d ones? I could swear there used to be a 2d x render test many years ago
[23:22] <jcristau> you said the problem occurred with 3d rendering..
[23:22] <penguin42> jcristau: Yes, but since the card doesn't have any 3d support I'm assuming it's one of the 2d opps that the 3d emulation is using
[23:23] <penguin42> 2d seems fine - the odd pixel dropping here or there, but it seemed sensible to try a 2d test if there was one
[23:24] <jcristau> without a 3d driver stuff gets rendered in software by swrast
[23:25] <penguin42> and swrast would just use all the normal 2d back end routines I assume?
[23:25] <jcristau> no
[23:26] <jcristau> well.  not really.
[23:26] <jcristau> swx11 does that.
[23:26] <jcristau> swrast does the rendering itself, and then uses PutImage to display it.
[23:27] <penguin42> ok, so in principal if PutImage is working swrast should be pretty much independent
[23:27] <penguin42> hmm a PutImage screw up would potentially fit the symptoms - it does look like a misrendered image
[23:28] <jcristau> putimage just copies bits around
[23:28] <jcristau> there's no magic there
[23:29] <penguin42> ok, it just looks almost like an image with the wrong depth/width 
[23:30] <penguin42> jcristau: http://launchpadlibrarian.net/50279619/Screenshot.png
[23:38] <penguin42> hmm cmake, not used that before, still it's pretty
[23:42] <penguin42> jcristau: Ah that's interesting, piglit's sanity check passes even though from the display it's obviously wrong
[23:55] <penguin42> Piglit's sanity test fails on his Hosts Radeon (open source driver/maverick)