[00:02] <Sarvatt> 7.9 might be releasing now
[00:02] <Sarvatt> oh yeah RAOF, no 3D on evergreen even though we have 7.9, it needs the updated ddx for sure
[00:04] <RAOF> Does that inconvenience you a lot?
[00:04] <RAOF> I wonder if that performance problem in 6.14 I saw reported on IRC got (a) filed or (b) resolved?
[00:04] <Sarvatt> not at all, I wasn't sure and I think you weren't sure if we got it so was letting you know :)
[00:06] <RAOF> Ok.  Maybe I should get an evergreen; my poor geforce 7600go struggles mightily with Civ V :)
[00:32] <Sarvatt> someone linked this in #radeon earlier and its a really nice fast openarena benchmark run - http://hifi.iki.fi/oa_benchmark.tar.gz
[00:33] <Sarvatt> was getting sick of using phoronix test suite and its taking 45 minutes per test :)
[00:46] <Sarvatt> ok here goes, note I only have 1366x768 available on my tv :)
[00:47] <Sarvatt> default options are mostly low-medium, wonder if it was autodetected
[00:49] <Sarvatt> oops, wrong window :)
[01:34] <Sarvatt> dang RAOF, now you got me hooked on civ 5..
[01:35] <RAOF> Don't worry - the archive is almost locked down :)
[02:32] <Sarvatt> demo over, boo!
[02:32]  * Sarvatt buys
[03:20] <Sarvatt> RAOF: uhoh, I hope what that test case was made for did not regress *again*
[03:20] <RAOF> I haven't changed my mesa.  I think?
[03:20] <RAOF> Um...
[03:20] <Sarvatt> ... white screen
[03:20]  * RAOF checks his notes.
[03:21]  * Sarvatt hopes he has a non archive package somewhere
[03:21] <RAOF> No, white screen isn't what that testcase tests.
[03:21] <RAOF> It tests the “No clutter for you!  Bad GLX drawable!” problem.
[03:22] <Sarvatt> CLUTTER_VBLANK=none required..
[03:22] <RAOF> Where?
[03:22] <RAOF> What mesa?
[03:22] <Sarvatt> 945
[03:22] <Sarvatt> hopefully its just something local screwing it up
[03:22] <Sarvatt>  7.9~git20100924-0ubuntu2
[03:22] <RAOF> GM45 is very happily running unity here.
[03:22] <RAOF> On, indeed, that mesa.
[03:23] <RAOF> libcluter-1.0-0 1.2.12-0ubuntu13?
[03:23] <Sarvatt> yep
[03:23] <RAOF> Arse.
[03:23] <Sarvatt> now it's starting up black instead of white
[03:24] <RAOF> Yeah.  That's the work around for the big white flash on startup being increadibly annoying.
[03:24] <Sarvatt> now it worked
[03:24] <Sarvatt> what is this crap
[03:24] <RAOF> What changed?
[03:24] <Sarvatt> nothing
[03:24] <Sarvatt> I ran it 6 times in a row
[03:27] <Sarvatt> CLUTTER_VBLANK=none unity works 100%, it didn't use to pass CLUTTER_VBLANK when launching it with unity
[03:27]  * Sarvatt wonders what the heck changed post RC..
[03:27] <RAOF> Well, we uploaded a new mesa.
[03:27] <RAOF> 0ubuntu2
[03:28] <RAOF> Are you running unity from gdm, or from within a session?
[03:28] <Sarvatt> I was using 0ubuntu2 before it was uploaded to confirm it fixed things
[03:28] <Sarvatt> inside a session
[03:28] <Sarvatt> desktop one, trying unity and unity -p
[03:28]  * RAOF tries to reproduce there.
[03:29] <bjsnider> is unity built on webkit, like gnome-shell?
[03:30] <RAOF> No, and gnome-shell isn't built on webkit.
[03:31] <RAOF> (It's built on gjs, the gobject-javascript bindings which (ab)use firefox's JS engine)
[03:31] <bjsnider> gnome-shell's appearance properties is a css file
[03:31] <bjsnider> there must be a layout engine at work
[03:31] <bjsnider> i assumed it was webkit
[03:31]  * Sarvatt squints at http://launchpadlibrarian.net/56915042/unity_0.2.46-0ubuntu3_0.2.46-0ubuntu4.diff.gz
[03:32] <RAOF> Sarvatt: Yeah, I can reproduce this in a desktop session with clutter -p
[03:32] <RAOF> I mean, um, unity -p
[03:34] <Sarvatt> hopefully launching it with mutter directly wont be broken like this, busy looking at what might have broken it because we absolutely made sure this was fixed for RC
[03:34] <Sarvatt> yeah mutter is fine
[03:35] <Sarvatt> maybe its just busted with compiz
[03:35] <RAOF> I'm using metacity
[03:39] <Sarvatt> mutter --replace --mutter-plugins=libunity-mutter and the session works fine here so i'll just chalk it up to more busted unity fun and not freak out again. it still doesn't work without GL_ARB_non_power_of_two (aka ATI) so this is minor :)
[03:43] <RAOF> I suspect that unity when run as a window might be selecting the wrong drawable to listen to swap_complete on?  Maybe?
[03:45] <RAOF> bjsnider: gnome-shell just uses a css-parsing library, IIRC, it doesn't embed an html renderer.  Same as the css-based gtk theme engine.
[03:58] <bjsnider> i see
[05:24] <cwillu_at_work> has anything been pushed out to 8.04 intel recently?
[05:25] <cwillu_at_work> a friend of mine lost his x after a normal apt-get upgrade with no special repositories enabled, and is kinda frustrated with the whole concept right now
[05:25] <cwillu_at_work> (intel, netbook interface)
[05:25] <cwillu_at_work> er, 10.04 I mean
[05:33] <Sarvatt> nope
[05:34] <RAOF> Most recent thing which could possibly have triggered it is the gdm update on 2010/09/27
[05:34] <RAOF> Next back would be linux-2.6.32-25.44 on 2010/09/18
[05:35]  * cwillu_at_work checks the changelog
[05:36] <Sarvatt> http://launchpadlibrarian.net/56057053/gdm_2.30.2.is.2.30.0-0ubuntu3_2.30.2.is.2.30.0-0ubuntu4.diff.gz
[05:36] <Sarvatt> try reverting that in /etc/init/gdm.conf
[05:37] <RAOF> That's a pretty long shot.
[05:37] <RAOF> But, hey!  Why not try.
[05:39] <Sarvatt> thats a hell of a kernel update changelog
[05:39] <Sarvatt> https://lists.ubuntu.com/archives/lucid-changes/2010-September/011691.html
[05:39] <RAOF> Yeah.
[05:39] <RAOF> A bunch of i915 changes there.
[05:39] <RAOF> It's the more likely culprit.
[05:41] <cwillu_at_work> got some more logs from him; I'm no longer sure that it's an x crash
[05:41] <cwillu_at_work> could a malformed .profile cause the session to die?
[05:42] <RAOF> I don't know.
[05:43] <cwillu_at_work> okay, I don't think this is an x problem;  he commented out the sketchy thing in .profile, and although it's still not logging in, it's also not immediately returning to the login screen
[05:45] <cwillu_at_work> seems more likely that he got hit by an update to dash that tightened up validation or some such
[05:46] <cwillu_at_work> he's got a bunch of things being complained about now
[05:48] <cwillu_at_work> okay, .profile errors _will_ cause an xsession to break
[05:49] <cwillu_at_work> which probably shouldn't be the case, but whatever
[05:49] <cwillu_at_work> sorry to make you guys dig through changesets :(
[05:52] <cwillu_at_work> AngryParsley> so I guess the problem was that I copied my .profile over, but never restarted my machine for weeks afterwards
[05:52] <cwillu_at_work> yay for a user who admits fault :p
[05:55] <RAOF> :)
[05:55] <JanC> even wrong permissions on some files could cause troubles with Xorg 
[05:56] <JanC> or with the session at least
[11:43] <tseliot> mvo: did you receive the message that I wrote yesterday on IRC?
[11:51] <mvo> tseliot: hello - I did. update-manager will not explicitely remove nvidia, only if it e.g. conflicts with xserver-xorg-core. so if 173 works u-m should be happy
[11:55] <tseliot> mvo: perfect, thanks
[13:17] <penguin42> isn't sure what to do with bugs 651994 and 654515
[13:17] <ubot4> Launchpad bug 651994 in xserver-xorg-video-intel (Ubuntu) "Dual-monitor possible with top-bottom screen configuration only (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/651994
[13:17] <ubot4> Launchpad bug 654515 in xserver-xorg-video-intel (Ubuntu) "Dual-head setup only possible with stacked screen combination (affects: 2) (heat: 12)" [Undecided,Invalid] https://launchpad.net/bugs/654515
[13:17] <penguin42> they're over 2k horizontal, but both claiming they had working configs on Lucid
[17:45] <Duke`> latest xserver-xorg-video-intel (20101004) is broken
[17:46] <Duke`> on lucid
[17:46] <Duke`> X segfaults
[17:47] <Sarvatt> got a log? someone else said it was broken on maverick too, i haven't tested it yet
[17:48] <Sarvatt> http://paste.ubuntu.com/506247/
[17:49] <Duke`> same log for me
[17:49] <Duke`> after "[    16.553] (II) intel(0): Setting screen physical size to 338 x 211", it breaks
[17:50] <Sarvatt> what gpu?
[17:50] <Duke`> btw, the given screen physical size is strange (338 x 211)
[17:50] <Duke`> i945
[17:51] <Duke`> http://paste.ubuntu.com/506627/
[17:51] <Sarvatt> any chance you could try earlier x-x-v-intel's to narrow down where it broke?
[17:51] <Duke`> the 20101002 is ok
[17:51] <Duke`> the 20101004 is ko
[18:04] <Duke`> I fail at driver building from git
[18:04] <Duke`> at autogen.sh stage :-(
[18:04] <Duke`> ./configure: line 11185: syntax error near unexpected token `RANDR,'
[18:04] <Duke`> ./configure: line 11185: `XORG_DRIVER_CHECK_EXT(RANDR, randrproto)'
[18:05] <Sarvatt> Duke`: install xserver-xorg-dev and xutils-dev?
[18:06] <Duke`> and this f****** laptop has serious hardware bugs (it just lost power link, now on battery only until I reboot -___-) 
[18:07] <Duke`> better with xserver-xorg-dev
[18:07] <Duke`> but now :
[18:07] <Duke`> Requested 'libdrm >= 2.4.22' but version of libdrm is 2.4.20
[18:07] <Duke`> ;_;
[18:08] <Sarvatt> not building against xorg-edgers?
[18:08] <Duke`> I use xorg-edgers packages
[18:08] <Duke`> but for lucid
[18:08] <Duke`> maybe there isn't the latest libdrm
[18:08] <Sarvatt> libdrm 2.4.22 is in edgers/lucid
[18:09] <Sarvatt> intel wouldn't build otherwise
[18:09] <Sarvatt> weird
[18:09] <Duke`> yeah
[18:10] <Sarvatt> whats it say when you sudo apt-get install libdrm-dev?
[18:10] <Sarvatt> maybe held back for some reason
[18:11] <Duke`> it says it's up to date
[18:11] <Duke`> I see the deb packages in pool but not in list
[18:11] <Sarvatt> oh Duke`
[18:11] <Sarvatt> you got caught in the accidental libdrm epoch bump
[18:11] <Sarvatt> i bet
[18:11] <Duke`> ;_;
[18:12] <Sarvatt> there's a note on the edgers page about it, the epoch was bumped for a few days by accident  :(
[18:12] <Duke`> hu ho
[18:12] <Duke`> how can I fix that?
[18:12] <Sarvatt> looks like i deleted the note because it was like 3 months ago
[18:13] <Sarvatt> sudo apt-get install libdrm2/maverick libdrm-dev/maverick libdrm-radeon1/maverick libdrm-nouveau1/maverick libdrm-intel1/maverick libkms1/maverick etc
[18:13] <Sarvatt> err lucid instead of maverick
[18:13] <Sarvatt> sorry about that man
[18:14] <Duke`> np
[18:14] <Sarvatt> leaving the bumped epoch would have been more of a PITA in the long run
[18:14] <Sarvatt> it screwed up all the symbols and stuff
[18:14] <Duke`> how can I list all packages installed from xorg-edgers?
[18:14] <Sarvatt> hmm maybe I should add that feature to ppa-purge
[18:15] <Sarvatt> oh you can look in software center
[18:17] <Sarvatt> key software-center doesn't help much, just shows stuff in there not available in the archive :)
[18:18] <Sarvatt> oh nevermind I dont even have xorg-edgers enabled, its just showing me stuff I have installed from here still
[18:18] <Sarvatt> renderbench and intel-gpu-tools
[18:20] <Sarvatt> Duke`: you can just ppa-purge xorg-edgers and reactivate it too, easier :)
[18:20] <penguin42> Sarvatt: So I now have two bugs both about >2048 wide displays that worked on Lucid
[18:20] <penguin42> possibly 3
[18:21] <Sarvatt> i dont have any more insight into them
[18:21] <penguin42> any idea what to do with them?
[18:21] <Duke`> ppa-purge? how does it works?
[18:22] <Sarvatt> Duke`: it goes over the list of all packages in the PPA, compares it to what you have installed, deactivates the PPA and for each one adds it to a list to sudo apt-get install package1/dist package2/dist so it downgrades them to the archive ones
[18:23] <Duke`> Warning:  Could not find package list for PPA: xorg-edgers ppa
[18:23] <Duke`> oh noes :(
[18:25] <Sarvatt> Duke`: <danvet> Sarvatt, latest xf86-video-intel looks indeed broken
 I'll take a stab at it, otherwise just prod ickel ;)
[18:26] <Duke`> hopefully maverick will be released in a week, so I'll can upgrade and clean a bit my packages ;p
[18:29] <Duke`> I upgraded libdrm & libkms manually (download and install)
[18:34] <Duke`> ok, I could compile it
[18:34] <Duke`> let's restart
[18:37] <albert23> penguin42: sounds like it might be bug 619663 (fixed with libdrm commit from freedesktop bug 28515)
[18:37] <ubot4> Launchpad bug 619663 in xserver-xorg-video-intel (Ubuntu) (and 2 other projects) "[maverick] Non-mirrored dual-screen gives narrow display on secondary monitor (affects: 15) (dups: 2) (heat: 88)" [Undecided,Confirmed] https://launchpad.net/bugs/619663
[18:37] <ubot4> Freedesktop bug 28515 in DRM/Intel "[i915] Failed to allocate framebuffer when exceed 2048 width" [Normal,Resolved: fixed] http://bugzilla.freedesktop.org/show_bug.cgi?id=28515
[18:51] <Duke`> 7c7294ec00d6c3a454a17a1b9983d14d0655162c is the first bad commit
[18:51] <Duke`> "shadow+dri2: Allow dri2 to be independently enabled with shadow"
[18:51] <Duke`> http://paste.ubuntu.com/506664/
[18:58] <penguin42> albert23: Do you think that is good enough for me to merge the others into it?
[19:01] <Sarvatt> Duke`: yeah none of the other commits touch 945/gm45 thats broken
[19:05] <Sarvatt> Option "Shadow" "True" probably works if you need it working now
[19:05] <albert23> penguin42: I'm not sure. The others don't seem to have logs from running with dual monitors
[19:06] <penguin42> albert23: If they did what would you be looking for in it?
[19:08] <albert23> penguin42: (EE) intel(0): Failed to allocate framebuffer. 
[19:09] <penguin42> albert23: OK, I'll ask the victims to give the logs
[19:09] <Duke`> Sarvatt: it seems to be a big commit that may break pretty bad ^^
[19:12] <penguin42> albert23: And if it has that then it's OK to dupe ?
[19:12] <albert23> penguin42: in my opinion, yes
[19:12] <penguin42> ok, will do - I spotted another one
[19:13] <Sarvatt> oh how I wish we released at the normal time and could have pulled in all the Q3 releases, quite a few x-x-v-intel and libdrm fixes
[19:14] <jcristau> Sarvatt: but 101010 was *so* important!
[19:16] <Sarvatt> I think the schedule is permanently moving back a few weeks now so it'll be tight every time from now on too
[19:53] <Sarvatt> RAOF: https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/651294
[19:53] <ubot4> Launchpad bug 651294 in xserver-xorg-video-intel (Ubuntu) (and 1 other project) "X crash on KDM logout (still - yes, really) (affects: 3) (heat: 14)" [High,Confirmed]
[20:24] <jcristau> Sarvatt: it seems hard to make the kdm people understand that relying on regen is not a winning plan
[22:13] <penguin42> So what do you do for someone whose hardware is giving bogus resolutions about the panel in their laptop?
[22:56] <RAOF> Sarvatt: Oh, *BAH*. That's no fun.