[10:27] <tseliot> jcristau, tjaalton: what package does it launch X with parameters such as "-bg" or "-fg" in debian and ubuntu? I can't remember
[10:41] <tseliot> ah, right it's /etc/X11/xinit/xserverrc in xinit. Never mind
[21:42] <RAOF> Woop!  Woop!
[21:43] <RAOF> Nouveau is now faster, cleaner, smoother than nv, *and* *supports* *suspend*.
[21:51] <tormod> RAOF, cool. I copied your new stuff to xorg-edgers also. you might want to get a newer libdrm from there.
[21:51] <RAOF> tormod: Oh, should I be uploading there?
[21:52] <tormod> would be nice if you can copy the ddx and the modules package there, so people can get everything in one place
[21:52] <RAOF> In future I'll just upload straight there, then.
[21:53] <tormod> yeah
[21:53] <tormod> I was just thinking you wanted to let people test nouveau without updating the whole xorg stack
[21:54] <RAOF> Well, maybe.  So, I guess I'll upload it to both places.
[21:54] <tormod> but they would profit from newer mesa as well
[21:54] <RAOF> Not for nouveau they wouldn't.
[21:54] <RAOF> Surely we don't build the nouveau DRI module, do we?
[21:55] <tormod> you can do a no-build copy for the module package, and a rebuild-copy of the ddx
[21:55] <tormod> oh well I guess not
[21:55] <RAOF> Upstream would be most displeased if we were shipping their DRI :)
[21:58] <tormod> ok :) I just have seen various nv50 and nouveau commits but I guess that's forbidden fruit then
[21:59] <RAOF> Oh, apparently it works quite well - to the point of developers running compiz & such.
[21:59] <RAOF> It's just aggressively unsupported still.
[22:00] <tormod> ok, but users can profit from newer sw rendering then :)
[22:00] <RAOF> Fair 'nuff :)
[22:03] <RAOF> So, nouveau TODO: (1) enable kms by default in the ppa, so users get working suspend/resume. (2) play with the DRI. (3) Get people discussing nouveau as default open-source nvidia driver for lucid.
[22:03] <bryce> hi guys
[22:04] <RAOF> Howdie.
[22:13] <Amaranth> RAOF: I've been told nouveau with metacity or kwin compositing feels as fast as nvidia with compiz
[22:13] <Amaranth> so that's pretty cool, that means the XRender acceleration is _awesome_
[22:16] <RAOF> Amaranth: I think they're underplaying the speed of nouveau.  I find it _faster_ than nvidia with compiz.
[22:17] <Amaranth> hehe
[22:17] <RAOF> And, yes.  The XRender acceleration is _awesome_.
[22:17] <Amaranth> It's as good as intel but the hardware is like 6x faster
[22:17] <RAOF> Right.
[22:17] <Amaranth> hopefully people don't start thinking using cairo for video games is a good idea :)
[22:18] <RAOF> This ~3 year old 7600go is tremendously faster than my new intel laptop.
[22:42] <tormod> bryce, seems like a couple of intel issues out there in 9.10? what do you think?
[22:43] <bryce> tormod, which ones?
[22:43] <bryce> unfortunately all the checkbox bugs filed against X have been consuming my attention
[22:44] <tormod> hangs with i8xx (but that's maybe nothing new) and resolution problems on i915
[22:45] <tormod> but I am pretty new to intel bugs, so maybe it is just business as usual what I am seeing
[22:46] <bryce> the resolution problems could be due to dropped monitor quirks due to KMS
[22:46] <bryce> jbarnes might have an inkling
[22:46] <tormod> and some have upgrade issues because they have cruft in xorg.conf I think
[22:46] <bryce> yeah
[22:48] <tormod> what are checkbox bugs?
[22:48] <jbarnes> tormod: not sure if the 8xx fixes are in the ubuntu kernel (I think most should be) but there have definitely been a lot of fixes in that area in the past few months
[22:48] <jbarnes> 9xx resolution stuff is new to me, what are you seeing?
[22:50] <tormod> for instance bug 433326
[22:53] <jbarnes> sounds like it's failing to get a decent edid?
[22:54] <jbarnes> hm no xrandr looks good
[22:54] <jbarnes> indicates things are running at 1360x768 even
[23:03] <bryce> tormod, checkbox bugs -> there is now a 'System Test' tool under the System menu, which after you go through it lets you file bug reports against xorg
[23:03] <bryce> bug 448913 is an example
[23:04] <bryce> bunch more are in the xorg package that I haven't had time to move/close/deal-with
[23:05] <bryce> jbarnes, on that bug I gather they hardcoded the h/v sync rates in xorg.conf to get their resolution
[23:05] <tormod> bryce, great report there ;)
[23:07] <bryce> bug 471246 is more typical
[23:08] <pwnguin> RAOF: the only place where it's slow is with the UNR clutter UI
[23:08] <bryce> jbarnes, ^ what do you think?  I'm not even certain what info is needed to debug such an issue
[23:08] <jbarnes> xrandr --prop from a startup w/o an xorg.conf
[23:08] <RAOF> pwnguin: Well, have you actually installed the DRI component?  Otherwise you're not really testing nouveau, you're testing how good mesa's software rasteriser is!
[23:09] <jbarnes> that would tell us if it finds an edid at least
[23:09] <bryce> jbarnes, wouldn't xrandr --verbose show the same?
[23:09] <bryce> (that's what is attached by default)
[23:09] <jbarnes> yeah that should have it, maybe I missed that
[23:10] <bryce> oh wait
[23:10] <bryce> checkbox seems to only use -q not --verbose
[23:10] <jbarnes> oops
[23:10] <bryce> wow, it's even more useless than I had thought