[06:12] <mlankhorst> morning
[06:25] <RAOF> Good morning!
[06:54] <mlankhorst> I'm looking into nouveau vblank thing, seems the problem is on newer cards we don't get the vblank interrupts yet :o
[06:58] <mlankhorst> But it's fun, basically invalid card mathods used to generate interrupts, and those interrupts are used to implement software calls.
[07:27] <mlankhorst> RAOF: oh yeah, the fun of conflicting version names :-)
[07:27] <mlankhorst> RAOF: shouldn't fakesyncs be named -fakesync1 ?
[07:28] <RAOF> As far as I'm aware there's no automation on that.
[07:29] <mlankhorst> shrug, ok :)
[07:55] <RAOF> mlankhorst: Incidentally, I'm going to be amending the version numbers of all the X stuff to -0ubuntu1 rather than -1; we're not syncing from released Debian packages, and things would go badly if Debian ended up pushing a slightly different -1 ;)
[07:56] <mlankhorst> sure
[07:57] <RAOF> Man the DDXen have *so many* compiler warnings :(
[07:57] <mlankhorst> indeed, it's horrid
[07:58] <mlankhorst> RAOF: want the patch to make geode compile on new X? I don't think it's in debian version control anywhere so just grab current git and apply the debian tree on top
[07:59] <RAOF> mlankhorst: I'll build all the drivers I care about first, and check that it minimally works. Then I'll go to drivers noone cares about :)
[08:00] <mlankhorst> RAOF: well it's part of xorg-video-all so some care
[08:00] <RAOF> Yeah, we won't be pushing out of -proposed until video-all is installable.
[08:01] <jcristau> we should trim down video-all.
[08:01] <RAOF> But I can at least check that ati, intel, and nouveau provide a working server first.
[08:01] <mlankhorst> it will be installable with the changes, but geode should be gone from video-all and maybe pushed to universe
[08:01] <jcristau> to like intel+ati+nouveau+modesetting+vesa+fbdev
[08:01] <mlankhorst> and video-all might need to be renamed to video-base
[08:01] <mlankhorst> jcristau: udl too
[08:02] <jcristau> leio was talking about fixing up geode i think
[08:02] <mlankhorst> ok :)
[08:02] <mlankhorst> well my patches are on the mailing list
[08:05] <RAOF> The next -ati release will really, really be 7.0.0, right?
[08:06] <jcristau> i should certainly think so, given 'drop ums' landed
[08:06] <mlankhorst> RAOF: yes
[08:06] <mlankhorst> that's the reason I needed git, to pick up the ums patches
[08:07]  * RAOF is just hoping to ensure the *next* upload of -ati isn't 6.99.99+is.really.goddamnit.something.else
[08:07] <mlankhorst> might still happen :P
[08:07] <mlankhorst> couldn't you just upload it as 6.999~somegit then?
[08:07] <mlankhorst> :D
[08:08] <mlankhorst> we shall just append more 9s to make
[08:08] <mlankhorst> up*
[08:09] <mlankhorst> RAOF: http://people.canonical.com/~mlankhorst/xf86-video-geode.tar.gz git tree + debian directory inside to build
[08:12] <mlankhorst> I have only verified it builds, I can pretty much guarantee it's broken until someone fixes exa acceleration
[08:13] <jcristau> hopefully leio can fix that :)
[18:02] <Prf_Jakob> RAOF: you are the one working on Wayland in 12.10 right?
[18:02] <Prf_Jakob> Whats the story there with hardware that doesn't support 3D?
[23:16] <RAOF> Prf_Jakob: Hardware that doesn't support 3D won't get the system compositor; it'll look exactly as it does today.
[23:16] <mlankhorst> RAOF: got vblank working on nouveau today, but might have to start over after nouveau changes again ;s
[23:16] <Prf_Jakob> RAOF: ok
[23:17] <RAOF> mlankhorst: At least you know what to do now :)
[23:17] <RAOF> Prf_Jakob: Over time the system compositor path will gain features that the non-composited path doesn't, but as long as the nvidia and fglrx drivers are important the non-composited path isn't going away.
[23:18] <Prf_Jakob> RAOF: ok
[23:18] <mlankhorst> RAOF: yeah have to do the prime stuff over as well due to the kernel rework, but i could have expected that
[23:18] <RAOF> We'll also possibly want a cut-down, software only “compositor” at some point; that's entirely possible.
[23:22] <RAOF> Prf_Jakob: We should be able to do egl+kms on vmwgfx at some point though, right?
[23:26] <Prf_Jakob> yes
[23:26] <Prf_Jakob> I think right now, I just haven't tested it.
[23:28] <RAOF> I should probably test that, shouldn't I :)
[23:28] <Prf_Jakob> Also 3D might not be enabled.
[23:29] <RAOF> That would make it a little bit more difficult to test :)
[23:30] <Prf_Jakob> Well its a VM option.