[10:01] <unggnu> hi all
[10:01] <unggnu> Wasn't the plan to push everything needed for Nouveau before Alpha 1 to got enough testing data? So what is the plan now? I have reade something about the kernel team but without a ppa it doesn't look like I can test nouveau right now in Lucid
[10:04] <tjaalton> you should ask the kernel team..
[10:05] <unggnu> tjaalton: How can I check nouveau then?
[10:06] <unggnu> under Lucid circumstances of course
[10:06] <unggnu> :)
[10:06] <tjaalton> no idea
[10:06] <unggnu> I just think that testing is important because of LTS and the state of the driver but I guess you all know this :/
[10:07] <tjaalton> it would still need the bits in the kernel
[10:07] <tjaalton> and that hasn't happened
[10:08] <unggnu> According to a newsletter the changes aren't so severe if I recall it right
[10:08] <unggnu> Radeon looks great though at least with an older card and also with 4870 with some minor issues. :)
[10:17] <unggnu> tjaalton: Couldn't you push a Kernel with the needed patches with a ppa?
[10:19] <tjaalton> unggnu: me? no
[10:24] <unggnu> tjaalton: no, in the X ppa
[10:24] <unggnu> tjaalton: it was a general you :D
[10:24] <tjaalton> too much work anyway, we have the kernel team for that
[10:24] <unggnu> ok
[10:24] <unggnu> Just kick their ... ;)
[10:25] <unggnu> If it doesn't work it have to be removed which also needs time so the earlier the better
[10:28] <unggnu> At least many radeon standby problems should be fixed I guess. Intel KMS should work with overlay again - but I am not sure if this changes were already applied in the 2.6.32 kernel.
[10:29] <jcristau> they weren't
[10:30] <unggnu> that's bad and I guess they are too severe for backporting?
[10:30] <jcristau> no
[10:30] <unggnu> Ok, lets hope they make it since pre i915 couldn't use kms then
[10:30] <unggnu> afaik
[10:30] <jcristau> it could.  just without video.
[10:30] <unggnu> or no video either way :)
[10:30] <unggnu> +accelerated
[10:34] <unggnu> Accelerated video is a basic and more important than 3D and of course KMS imho. Btw. does KMS work with i815? Hasn't Intel dropped support for this chipset?
[10:36] <jcristau> kms is in the i915 driver.  so i830+
[10:36] <jcristau> i81[05] is still using ums, xaa, dri1, ...
[10:36] <tjaalton> hmm so 2.10 will drop support for them
[10:37] <jcristau> well the 3 of them that still exist.
[10:37] <unggnu> jcristau: :D When was their release date?
[10:37] <unggnu> So X will be adaptive in future? Starting without root rights if possible and vice versa?
[10:37] <jcristau> http://en.wikipedia.org/wiki/Intel_810
[10:37] <jcristau> 10 years ago
[10:38] <unggnu> Ok, if vesa still works and if it has kms in the future ... :D
[10:39] <unggnu> Without enough RAM Ubuntu should be run just fine on P3s. Maybe the Linux Intel graphic guys have more time in the future since the big changes should be done now. :)
[10:40] <tjaalton> you wish
[10:41] <unggnu> tjaalton: that's for sure :D
[10:41] <unggnu> They probably just create zxa or something like that :D
[10:45] <tjaalton> jcristau: I'll pull in the xorg.conf.d/inputclass stuff for testing, tseliot promised to fix the fallbacks ;)
[10:46] <unggnu> No response from the Kernel team but I guess it is the time :)
[10:46] <unggnu> anyway, bye guys :)
[10:54] <tseliot> tjaalton: do you mean in a PPA, right?
[10:54] <tjaalton> tseliot: just for git
[10:55] <tjaalton> for now
[10:55] <tseliot> tjaalton: in a separate branch, not in our master ubuntu branch, right?
[10:55] <tjaalton> no, in the master branch
[10:55] <tjaalton> as separate patches
[10:56] <tseliot> tjaalton: I was asking as I will have to upload a packaging change in X to get the alternatives stuff to work
[10:56] <tjaalton> just push it
[10:56] <tseliot> and I would like to upload that without uploading the other patches
[10:57] <tseliot> how would it work?
[10:57] <tjaalton> I haven't pushed anything yet
[10:57] <tseliot> ok
[10:58] <tseliot> tjaalton: in this case maybe I don't need to put UNRELEASED in the distro
[10:59] <tjaalton> if you have no other changes..
[10:59] <tseliot> no, nothing else
[10:59] <tjaalton> otherwise use several commits with just one functional change in each
[11:05] <tjaalton> well, on a second thought I won't be pushing this to the branch
[11:10] <tseliot> good
[11:13] <tjaalton> apart from the inputclass commit bumping the ABI, they apply ok
[11:17] <tseliot> yes, I noticed the input ABI bump
[11:24] <tseliot> tjaalton: ok, pushed
[11:25] <tseliot> I can upload it later
[11:28] <jcristau> tjaalton: the main problem with the fallbacks (besides the fact that the code is a bit painful) is i never figured out what behaviour i wanted.  also i need a screen and a device section for each driver.  and add each screen to the serverlayout.  so do i do that only if there's no serverlayout, or if there's a serverlayout with no screen, or a screen with no device, or a device with no driver, or ...
[11:28] <tjaalton> jcristau: :/
[11:32] <jcristau> and when you get to the point in the code where you want to add a few devices/screens, the list of screens in the layout is already allocated so you have to go back to add more
[11:32] <jcristau> so it needs some sitting down which i wasn't able to do yet :)
[11:33] <tjaalton> yeah it sounds troublesome
[11:35]  * tseliot nods
[14:30] <jneves> hi there - I'm having a strange problem with 9.10 and gdm-2.20 - every time I start, gdm fails (the screen is screwed up so I can't read it) - if I git thru recovery more and do "resume" it works fine
[14:30] <jneves> anyone has a hint on what I should look at?
[14:33] <jneves> the closest thing to a clue I have is this on gdm logs: Jan  4 13:20:33 net gdm[1366]: DEBUG: gdm_server_start: Server :0 died during startup!
[14:50] <jneves> brb
[15:54]  * jneves found the problem - failsafe is starting before gdm - disabling failsafe "solves" this
[15:55] <jneves> https://bugs.launchpad.net/ubuntu/+source/gdm-2.20/+bug/491483
[16:04] <superm1> why are you using gdm-2.20 in the first place?
[16:04] <superm1> i personally don't think it should have ever been introduced to the archive
[16:05] <superm1> it was introduced so xubuntu didn't need to pull in so much of gnome, but i dont think it got enough testing
[16:07] <jneves> superm1: because in my case gdm doesn't work
[16:07] <superm1> jneves, what's your problem with gdm?
[16:07] <jneves> superm1: groups of users go to different networks depending on their authentication (the change is done by pam)
[16:08] <jneves> superm1: when I logout from one user, gdm fails in strange ways (the user input box does not appear)
[16:09] <superm1> jneves, so wouldn't the proper thing to do be file bugs on gdm then for these cases?
[16:10] <jneves> superm1: yes, but I also need to work around the problem in the meantime
[16:10] <superm1> gdm-2.20 is gonna need work to  be reworked as an upstart job
[16:10] <superm1> otherwise failsafe-x will kick in as you found
[16:10] <jneves> superm1: it's working at the moment
[16:10] <jneves> superm1: it's not that simple - right know, if you have gdm installed, but choose another dm, failsafe-x does the wrong thing
[16:11] <jneves> superm1: right now, failsafe-x only works with gdm :(
[16:12] <superm1> jneves, ah
[16:13] <jneves> superm1: so, yes this is working around a bug - no, the current situation isn't right either ;)
[18:26] <Duke`_> ^£$@+*!!
[18:26] <Duke`_> execbuf error WTF?!!1
[19:45] <schilli> I just installed Lucid today.  Ran into an issue on the Eee 1005hab setting up an external monitor...
[19:47] <schilli> Works fine mirroring, or with one or the other monitor disabled.  When start taking the external monitor to higher resolutions, things get hairy...
[19:48] <tjaalton> schilli: poulsbo gfx, not supported properly
[19:49] <schilli> it is known about and being dealt with?
[19:49] <tjaalton> known, nothing we can do about
[19:49] <tjaalton> binary blobs ftw
[19:50] <schilli> but doing the same thing works fine in Karmic.
[19:50] <jcristau> then use karmic
[19:52] <schilli> OK
[19:53] <schilli> Where, if anywhere, should I be of any use to anyone while messing around and trying things out with Lucid?  
[19:57] <tjaalton> triaging bugs on lp
[20:07] <bryyce> heya tjaalton
[20:08] <tjaalton> bryyce: hi, how was the vacation?
[20:08] <bryyce> very good, got lots of house/shop projects done (or mostly done)
[20:09] <bryyce> tjaalton, how things going for you?
[20:09] <tjaalton> I played a lot of Rock Band (& The Beatles), and read two books :P
[20:11] <bryyce> nice
[20:11] <bryyce> I played a lot with my son :-)
[20:12] <bryyce> hey, I saw some email about libdrm on ubuntu-x
[20:12] <tjaalton> yep
[20:12] <bryyce> I'm about to email Intel to get some extra priority on that issue, but are there any recent updates about it?
[20:12] <maxb> I am finding that _intermittently_ Lucid goes into low graphics mode on first boot, but will then work if you restart gdm. I hypothesize some boot-time race condition that leaves X not knowing about its display devices properly. Can anyone suggest how I would pursue this?
[20:12] <tjaalton> not on the bug report at least, since dec 31
[20:13] <bryyce> tjaalton, ok but the issue is still occurring?  Anyone beyond sarvatt seeing it?
[20:13] <tjaalton> maxb: yes, happens here too, haven't figured out why
[20:13] <tjaalton> bryyce: yes, apparently it affects 945 chips
[20:13] <bryyce> tjaalton, also you mentioned setting up a ppa with the new libdrm and mesa; did this get set up?
[20:13] <tjaalton> though I haven't tried it on my 965
[20:14] <tjaalton> bryyce: no, not yet
[20:14] <tjaalton> but I did update libdrm git
[20:14] <bryyce> ok
[20:15] <bryyce> I guess this could go into the x-testing ppa:  https://edge.launchpad.net/~ubuntu-x-swat/+archive/ppa
[20:16] <tjaalton> yes
[20:28] <tjaalton> we probably need to (again) use the drm headers from libdrm instead of the kernel, for the new intel at least
[20:28] <tjaalton> probably radeon too
[20:28] <bryyce> yeah
[20:29] <bryyce> that should be fine for x-testing
[20:29] <tjaalton> bryyce: btw, since intel 2.10 will drop support for !kms, it also means that it won't "support" i81{0,5} anymore
[20:29] <bryyce> tjaalton, good grief
[20:29] <tjaalton> but, who cares
[20:30] <bryyce> well, I care ;-)
[20:30] <tjaalton> those are ancient
[20:30] <tjaalton> bah :)
[20:30] <bryyce> um, not *so* ancient...
[20:30] <tjaalton> 10y
[20:31] <bryyce> is 2.10 == 2009Q4?
[20:31] <tjaalton> yes, slipped though
[20:31] <bryyce> hmm
[20:32] <bryyce> oh apparently I haven't had enough coffee
[20:32] <bryyce> I thought you wrote i8*{0,5}
[20:32] <tjaalton> hehe
[20:32] <bryyce> I'm much less concerned with dropping i810/815 support... those have been more or less dead for a while now
[20:33] <tjaalton> right
[20:33] <superm1> well hopefully those gracefully fallback to vesa though rather than just intel driver not working and kicking in failsafe-x
[20:34] <bryyce> superm1, good idea
[20:34]  * bryyce adds this to his todo list
[20:34] <tjaalton> the server will be fixed yes
[20:34] <tjaalton> not to try intel on those
[20:38] <bryyce> tjaalton, is someone else already planning on doing that?  If so I'll leave it off my todo
[20:38] <bryyce> btw, iirc we talked some weeks ago about ideas for improving the xserver autodetection stuff...  did those ideas ever make it into a wiki page?  I've largely forgotten what we talked about specifically
[20:39] <tjaalton> bryyce: it'll probably happen upstream for 1.8, when 2.10 is released. other than that, I don't know
[20:39] <bryyce> ok, I'll leave it on my todo list to verify it when we merge 2.10
[20:40] <tjaalton> bryyce: well, no wiki page exists for that. jcristau mentioned some problems earlier today
[20:41] <Duke`_> Sarvatt, I got an execbuf error with your commit-reverted packages :-(
[20:44] <Sarvatt> Duke`: yeah me too, both reverts is still ok though but the other doesn't revert cleanly from 2.4.17 :(
[20:46] <tjaalton> huh, autosyncs are clearly not happening
[20:46] <tjaalton> libxcb is still at 1.4-1
[20:53] <tjaalton> building mesa 7.7 locally
[20:53] <tjaalton> I was wondering to add ubuntu-test branches to git for stuff like this
[20:54] <jcristau> tjaalton: iirc jd_ switched libxcb source to 3.0 format, maybe that stopped it?
[20:55] <tjaalton> jcristau: xcb-proto is synced though, lp supports it now
[20:55] <jcristau> ok
[21:12] <bryyce> tjaalton, are you noticing any other broken autosyncs besides libxcb?
[21:13] <tjaalton> bryyce: no, it's the only one, the others are not in testing yet
[21:14] <bryyce> ok
[21:32] <bdmurray> bug 434032 is likely a duplicate of something but I can't remember which bug
[21:32] <Duke`> damin it
[21:32] <Duke`> google maps raised execbuf error
[21:36] <tjaalton> bdmurray: replied to the bug
[21:37] <bdmurray> tjaalton: okay thanks!
[22:34] <tjaalton> bryyce: libdrm and mesa uploaded to uploaded to the ubuntu-x ppa
[22:34] <bryyce> tjaalton, excellent
[22:34] <tjaalton> sigh, mesa tarball size has tripled
[22:35] <jcristau> argh indeed.
[22:35] <jcristau> -rw-r--r-- 1 julien julien  8773112 Dec 29 10:28 ../mesa_7.6.1.orig.tar.gz
[22:35] <jcristau> -rw-r--r-- 1 julien julien 26176639 Dec 25 16:12 ../mesa_7.7.orig.tar.gz
[22:35] <jcristau> and 7.7 has the glut sources removed.
[22:35] <tjaalton> it did?
[22:36] <tjaalton> I diffstat'ed against rc2
[22:36] <bryyce> whoa, wtf?
[22:36] <jcristau> i didn't include MesaGLUT-7.7.tar.gz in my 7.7.orig
[22:36] <tjaalton> ah ok
[22:37] <tjaalton> progs/objviewer is  now huge
[22:37] <jcristau> wtf is this stuff?
[22:37] <tjaalton> well, 7.6.1 didn't have it
[22:40] <bryyce> tjaalton, I see libdrm in x-testing but not mesa?
[22:41] <tjaalton> bryyce: still uploading
[22:41] <bryyce> aha great
[22:44] <tjaalton> bryyce: so I was wrong about the i810/815, they still use ums on 2.10 (which was tagged a moment ago)
[22:45] <bryyce> ahh, ok good
[22:45] <tjaalton> mesa upload finished
[22:45] <bryyce> tjaalton, I've emailed Intel about this bug.  I sent them the link to the ppa and asked that they up their priority on it
[22:45] <tjaalton> bryyce: good
[22:46] <bryyce> also asked yingying's team to try to reproduce (sometimes that helps get things moving)
[22:46] <bryyce> would be nice to get comment from developers on it
[22:46] <tjaalton> yep
[22:46] <bryyce> maybe now that holidays are over it'll get attention
[22:47] <tjaalton> hope so
[23:57] <Sarvatt> thats odd, gdm isn't starting randomly on boot, just getting a login prompt sometimes now. I get a could not create session error every startup for a second or two before it draws the panel when it does boot also
[23:57] <Sarvatt> running that libdrm 2.4.17/mesa from x-testing now to see how things fare re: execbuff while wedged errors
[23:57] <bryyce> thanks
[23:58] <bryyce> Sarvatt, is "could not create session" gdm speak for 'X crashed' perhaps?