[01:36] <ubotu> New bug: #193111 in xserver-xorg-video-unichrome (universe) "Unsupported VT3371" [Undecided,New] https://launchpad.net/bugs/193111
[08:18] <Q-FUNK> guys, are we including the CPUID patch from bug #180742 for Hardy?
[08:18] <ubotu> Launchpad bug 180742 in xorg-server "xf86-video-amd: switching to a vcons fails" [Unknown,Confirmed] https://launchpad.net/bugs/180742
[08:19] <Q-FUNK> it's the last of 3 x86emu patches needed to bring x86emu back into line.
[08:40] <ubotu> New bug: #193192 in xorg (main) "DPMS is broken" [Undecided,New] https://launchpad.net/bugs/193192
[08:43] <tjaalton> Q-FUNK: I'll add it today
[08:44]  * tjaalton drops eyes to the floor... ati 6.8.0 released
[08:44] <tjaalton> sheesh
[08:46] <Q-FUNK> thanks!
[08:46] <Q-FUNK> hehe
[08:51] <bryce> heya
[08:52] <Q-FUNK> hey :)
[08:53] <tjaalton> bryce: I guess we want the new ati in hardy?-) Unstable already has it, I'm looking at the merge or sync
[08:55] <bryce> tjaalton: yeah probably, but we should review the changelog to get a sense of the proportion of new features to bug fixes
[08:55] <bryce> and also testing results of course...  if there's a lot of confirmed bug fixes, we definitely want it
[08:56] <tjaalton> it's a stable release, we have a six weeks old snapshot.. :)
[08:56] <tjaalton> sorry, eight weeks
[08:56] <bryce> yes quite true
[09:06] <ubotu> New bug: #193198 in xorg (main) "Screen errors in ATI Mobility Radeon X700 in Hardy" [Undecided,New] https://launchpad.net/bugs/193198
[09:08] <tjaalton> we should be able to sync it
[09:09] <tjaalton> the only patch is from upstream
[09:09] <bryce> ok
[09:10] <tjaalton> Ng: upstream want's you to try git master to fix the graphics bugs with 855
[09:10] <tjaalton> -'
[09:12] <tjaalton> well, the last commit on the stable branch
[09:12] <tjaalton> http://gitweb.freedesktop.org/?p=xorg/driver/xf86-video-intel.git;a=commitdiff_plain;h=aa1813e4807e500fb122f36af988cf60f91e05a5;hp=1cea254b70158600d9dfff8ba66fb2ec0a6e0f67
[09:17] <Ng> tjaalton: can I build that independently of xorg*
[09:17] <Ng> +?
[09:18] <bryce> Ng, yup
[09:18]  * Ng installs git
[09:18] <Ng> is it just git co http://thaturl     ?
[09:18] <tjaalton> just wget that patch and apply it on the packages
[09:18] <tjaalton> -s
[09:19] <tjaalton> or build the git version
[09:19] <tjaalton> and copy the drv.so
[09:19] <Ng> aha
[09:20] <ubotu> New bug: #149479 in linux-restricted-modules-2.6.24 (restricted) "Graphic set-up is problematic" [Undecided,New] https://launchpad.net/bugs/149479
[09:20] <mvo> tjaalton: I would like to talk to you about the fglrx issue when they drop support for cards (you filed a update-manager bug about it). do we have a list of pci ids available of the cards that are supported or where support got dropped?
[09:22] <Ng> hmm, don't spose anyone still has a copy of xserver-xorg-video-intel 2.2.0+git20080107-1ubuntu2 lying around, do they? if this doesn't work I'd like to be able to revert back to the working driver ;)
[09:24] <bryce> Ng, the (probably crude) approach I've taken when working with git drivers, when done is to do apt-get remove <driver>; apt-get install <driver>.
[09:24] <bryce> there's probably a cleaner solution tho
[09:24] <tjaalton> mvo: hmm, what bug #?
[09:25] <bryce> I believe you can do apt-get install <driver>=<version> or something to force it to a particular version
[09:26] <Ng> bryce: sure, but aiui, older packages get purged from the archive, so apt won't be able to find it
[09:27] <saispo> hi
[09:27] <bryce> Ng, ahh, yes that's true
[09:27] <bryce> saispo: heya!
[09:27] <mvo> tjaalton: #110388 you reassigned this one to update-manager on saturday
[09:27] <saispo> :)
[09:28] <saispo> little question: anyone know how can i do not invoke xorg postinst in gutsy installer ? because it build a xorg.conf automatically and i don't want it :)
[09:30] <bryce> saispo: there isn't a way to suppress the initial generation of it, however you should be able to delete it at any point during install by switching to a terminal console and deleting it.  (Perhaps I misunderstand what you need tho.)
[09:30] <tjaalton> mvo: oh that one
[09:30] <saispo> bryce: yes, i see no other solution :|
[09:31] <saispo> bryce: you see right ;)
[09:31] <saispo> bryce: i will use a d-i command for doing this
[09:31] <bryce> saispo: if you have X trouble during installation, another workaround is to use the alternate install CD, which does a text-mode install
[09:31] <saispo> thanks
[09:31] <bryce> sure
[09:31] <saispo> bryce: i use this, but it's for a custom distro
[09:31] <saispo> bryce: with feisty i have no trouble
[09:31] <tjaalton> mvo: lrm includes the modules.alias.override -files which could be used
[09:31] <bryce> ah
[09:33] <Q-FUNK> seb128: could you tell me what was the name of the gnome componet that draws the bacground image?
[09:33] <tjaalton> nautilus?
[09:33] <seb128> Q-FUNK: depends, gnome-control-center or nautilus
[09:33] <seb128> what is the issue?
[09:34] <tjaalton> bryce: do I file the sync request for ati?
[09:34] <bryce> tjaalton: sure go for it
[09:34] <tjaalton> ok cool
[09:34] <Q-FUNK> seb128:  zoom image behavior changed. it used to center and zoom, now it left-aligns and zooms
[09:35] <Q-FUNK> seb128: in hardy
[09:36] <seb128> not sure, you can open it on gnome-control-center if you want, attach an example to the bug maybe, I don't notice the issue on my installation
[09:37] <Q-FUNK> seb128: I tried both with the left-click option on the dektop and with the appearnce dialog. the result is the same in both cases.
[09:38]  * Ng goes with a ghetto solution of tarring up the contents of the current working package ;)
[09:40] <seb128> Q-FUNK: the left click option starts the appearance capplet
[09:43] <Q-FUNK> seb128: ok
[09:46] <tjaalton> Ng: just copying intel_drv.so is enough
[10:10]  * Ng lost in a sea of version numbers. I want to be building 2.2.0.90, right?
[10:11] <ubotu> New bug: #193211 in xserver-xorg-video-ati (main) "Please sync a new version from Debian unstable" [Wishlist,Confirmed] https://launchpad.net/bugs/193211
[10:12] <tjaalton> Ng: yes, and the patch on top of that
[10:14] <Ng> hmm, then I reckon it's working. I thought I'd built the previous snapshot version that was working, but I seem to have build 2.2.0.90. I'd like to try again with absolutely-definitely-guaranteed pristine source though ;)
[10:14] <tjaalton> ah :)
[10:19] <Ng> I have some hardy updates to apply first too
[10:59] <bryce> tjaalton: wow, long changelog for -ati.  I'll keep my fingers crossed that it's stable
[11:07] <tjaalton> sure it is
[11:13] <bryce> night
[11:20] <ubotu> New bug: #193233 in linux-restricted-modules-2.6.24 (restricted) "nvidia api mismatch when installing nvidia-glx-new" [Undecided,New] https://launchpad.net/bugs/193233
[11:25]  * Ng up to date. building driver again...
[15:24] <tjaalton> bryce: btw, the changelog was 6.6.3 -> 6.8.0, so it's considerably shorter for 6.7.197 -> 6.8.0
[16:01] <ubotu> New bug: #193323 in xserver-xorg-video-cirrus (universe) "x fails to startup in qemu" [Undecided,New] https://launchpad.net/bugs/193323
[16:36] <ubotu> New bug: #187645 in linux-restricted-modules-2.6.24 (restricted) "Evolution calendar consumes 100% CPU with mouse-over info" [Low,New] https://launchpad.net/bugs/187645
[16:51] <seb128> I reassigned this evolution calendar issue
[16:51] <seb128> that's basically cairo rendering being slow on fglrx when not using composite
[16:51] <seb128> not sure if that's cairo, xorg or the driver to blame though
[17:04] <Ng> tjaalton: it looks to me like that patch works
[17:04] <Ng> either that or I'm utterly confused about versions ;)
[17:05] <bryce> tjaalton: ahh, that explains it
[17:23] <tjaalton> Ng: cool, I'll let upstream know
[17:23] <Ng> :)
[17:23] <Ng> can has a new upload with the patch? :D
[17:24] <tjaalton> just apt-get reinstall the driver to see it really was the patch that fixed it?
[17:24] <tjaalton> we'll see :)
[17:25] <Ng> I'll do the reinstall when I get home (couple of hours)
[17:25] <Ng> I'd prefer not to break things while I'm working ;)
[17:28] <tjaalton> heh, ok
[18:53] <ubotu> New bug: #192994 in xorg (main) "X doesnt shut down. Requires CTRL ALT BACKSPACE to kill it so it can shut down" [Undecided,New] https://launchpad.net/bugs/192994
[20:01] <ubotu> New bug: #193419 in xserver-xorg-video-intel (main) "External monitor loss - intel 945" [Undecided,New] https://launchpad.net/bugs/193419
[20:22] <mvo> bryce, tjaalton: what is the latest status with regard to i965 and exa/xaa ? I got some mails asking about this for compiz and I was wondering if there is a decision about this yet
[20:23] <bryce> mvo, in fact I'm just now working on my patch for it
[20:24] <bryce> mvo, so I should have more news by the end of the day
[20:24] <mvo> bryce: the patch that makes it happy with exa?
[20:24] <bryce> yes
[20:24] <mvo> thanks
[20:24] <tjaalton> well, 965
[20:25] <tjaalton> but the rest should use XAA
[20:25] <bryce> tjaalton: aside from my patch, is there anything else we need to do to ensure <965 are using XAA?
[20:25] <mvo> I have still distorted window shadows on the 965, but I guess you know about this, right?
[20:26] <mvo> <i965 should be fine, as long as the patch I did for gutsy to use the hardware overlay for Xv is still applied
[20:26] <mvo> (compiz-wise all should be fine for those)
[20:26] <tjaalton> bryce: yes, there's a commit which switched EXA on, so revert that and then add to your patch the rule to use EXA for 965
[20:26] <bryce> yup, in fact that's what I'm using to check if this is working
[20:27] <bryce> mvo, I have a simple patch to fix those borders which has been tested to work, but if this greedy patch solves it, we don't need that patch
[20:27]  * mvo nods
[20:27] <bryce> tjaalton: hmm, sounds like I'll need to also do a patch to force EXA for 965... currently I just check if both those are enabled, and then turn on greedy too
[20:28] <mvo> great, thanks for the update. I'm looking forward to it
[20:28] <tjaalton> mvo: ah so it was your patch :)
[20:28] <tjaalton> bryce: right
[20:28] <mvo> tjaalton: eh, did it broke anything :) ?
[20:28] <tjaalton> mvo: no, I was just wondering where it came from :)
[20:29] <mvo> aha, ok :)
[20:29] <mvo> I think we talked about it when I uploaded the packages, but it was a while ago. it bugged me to no end that i965 wouldn't work, until I discovered that the hardware is missing
[20:29] <mvo> yeah for no docs from intel (fixed now fortunately)
[20:32] <bryce> there's an article on lwn complaining that the community invests more resources in things like reverse engineering nvidia drivers where there aren't docs, and none in -intel where there are (now) plenty of docs.
[20:32] <bryce> I don't know if I would agree there - I suspect it's just that Intel ends up hiring everyone that works on -intel ;-) - but an interesting point of view
[20:36] <pwnguin> id argue as well that nvidia drivers are more valuable
[20:36] <pwnguin> given that the chips are more capable
[20:36] <bryce> pwnguin: do you mean because of nvidia's hardware performance potential?
[20:37] <pwnguin> yea
[20:37] <pwnguin> unless intel has some magic i havent read about since i last went laptop shopping
[20:38] <bryce> whoohoo!!
[20:38] <Q-FUNK> intel voodoo incantation required.
[20:38] <bryce> think I got it working...
[20:38] <bryce> yup :-)
[20:39] <bryce> tjaalton: ok with my greedy patch, borders are now correct, and I've verified greedy is turned on for 965
[20:39] <bryce> tjaalton: what else do we need to check?
[20:43] <tjaalton> that should be enough for now, if it forces EXA as well?
[20:43] <bryce> no I still need to code up that patch
[20:43] <bryce> first want to finish up this one and get it in
[20:44] <bryce> I figure I'll do the two things as separate patches so it'll be easier to drop one or the other independently in the future if we need to
[20:44] <tjaalton> sure
[20:46] <bryce> hmm, doesn't fix the glxgears issue but I guess that's separate.  Movie playback is a lot better
[20:46] <tjaalton> what issue?
[20:47] <tjaalton> btw, someone said that setting EXNoComposite true would improve the performance, but I couldn't tell the difference
[20:50] <bryce> the issue is, run glxgears, then move the window, and it doesn't clear the screen properly
[20:51] <tjaalton> isn't that true for all free drivers atm?-)
[20:52] <tjaalton> maybe the cleanup works with xaa though
[20:52] <bryce> oh is it?  I didn't see it on gutsy with -ati, but maybe it's a 1.4 issue?
[20:52] <bryce> oh and I'm not running compiz, duh
[20:52] <tjaalton> ah
[20:56] <mvo> bryce: that is unfortunately a known issue with compiz :(
[20:57] <mvo> we tried to come up with a workaround, but its (unsurprisingly) pretty difficult
[20:59] <bryce> hmm, this is odd - I rebooted my 965 laptop without installing any greedy patched versions, which had the tearing issue and so forth, and now I cannot reproduce the issues
[21:02] <tjaalton> using the defaults?
[21:03] <bryce> I think so... doing some more testing...
[21:12] <ubotu> New bug: #193458 in xinit (main) "Xsession fails due to an incorrect COLUMNS value" [Undecided,New] https://launchpad.net/bugs/193458
[21:13] <bryce> btw, here are the deb's I just tested, which seem to work:  http://people.ubuntu.com/~bryce/Testing/Greedy/
[21:15] <bryce> so far testing is confirming it, but would be nice for independent verification
[21:17] <bryce> performance seems about the same either way, hm
[21:19] <tjaalton> try scrolling in firefox
[21:21] <bryce> ok
[21:21] <bryce> ah, nastiness
[21:22] <bryce> aha, ok all issues reproduced on test hw without patch... now trying with patched driver...
[21:23] <bryce> ahhh, yes that's much better
[21:23] <bryce> cool even the screen tearing's gone
[21:33] <soren> Did something get updated in gutsy relatively recently that could cause my wife's laptop's X to die a few times of day?
[21:41] <tjaalton> only the security patches a month ago
[21:41] <tjaalton> hm, unless the intel driver got in updates
[21:43] <bryce> don't think that's happened
[21:47] <bryce> tjaalton: ok, the xserver patch portion of the fix is ready to be uploaded:  http://people.ubuntu.com/~bryce/Uploads/
[21:48] <tjaalton> 3 is uploaded already ;)
[21:51] <bryce> hmm, apt-get source only pulled a 2 for me
[21:55] <tjaalton> I uploaded it earlier today
[21:57] <bryce> ok, here is an xserver 1ubuntu3, and -intel 2ubuntu5: http://people.ubuntu.com/~bryce/Uploads/
[21:57] <bryce> could you upload both of those?  Next I'll do the force-exa-on-for-965 patch
[22:05] <tjaalton> sure
[22:15] <bryce> hrm
[22:22] <Ng> tjaalton: ok, that patch is needed, it stops the weird 16-bit look-a-like pixmaps
[22:22] <Ng> (which seems to affect 2d and 3d, weirdly)
[22:23] <Ng> I see it on the ubuntu logo in gdm and on the photo icons on my desktop without the patch
[22:23] <Ng> (I also tried with/without the migration config bits in xorg.conf)
[22:25] <tjaalton> Ng: thanks, I'll add that to the upstream bug
[22:26] <bryce> tjaalton: oh, hmm I forgot, but maybe the depends for -intel should be increased to require that newer xserver
[22:26] <Ng> tjaalton: is there likely to be another upstream drop into hardy before release?
[22:26] <bryce> I've got the xaa-default-except-i965 patch ready to go, I'll roll that depends update into it too if that sounds good to you
[22:26] <tjaalton> Ng: likely
[22:27] <Ng> groovy
[22:29] <Ng> hmm, I don't remember if I actually filed that as a bug
[22:29] <bryce> tjaalton: could you remind me why we prefer XAA for <i965? stability or performance?
[22:30] <tjaalton> performance
[22:30] <tjaalton> Ng: yep you did
[22:30] <tjaalton> Ng: I'll apply the patch on our intel for now
[22:31] <Ng> k
[22:31] <Ng> wrt XAA/EXA, subjectively i think EXA is as fast on 855 with the greedy migration doodad
[22:31] <tjaalton> hmm
[22:32] <tjaalton> maybe it doesn't hurt to let alpha5 be with EXA+greedy for all
[22:34] <bryce> could be, although iirc there are some hardware specificness for EXA performance
[22:35] <tjaalton> it would be nice to not go back to xaa
[22:35] <tjaalton> if possible
[22:36] <tjaalton> I mean staying as close to upstream as possible
[22:38] <bryce> ok, well I'm almost done with the revert back to xaa by default except i965 patch...  let me complete that
[22:39] <bryce> ...ok done - http://people.ubuntu.com/~bryce/Uploads/ 2ubuntu6
[22:39] <bryce> hmm
[22:40] <bryce> oh, btw I built the above but didn't test to ensure it goes to xaa on <965
[22:40] <bryce> but the patch looked very straightforward, I'd be surprised if it didn't do exactly what's intended
[22:41] <bryce> oh crap, I forgot to update the depends... one sec
[22:42] <tjaalton> of the driver?
[22:42] <tjaalton> I can do it locally
[22:42] <bryce> yeah
[22:42] <bryce> oh okay cool
[22:43] <tjaalton> although
[22:43] <bryce> if we even want to do that... the driver will build and work against older 1.4 drivers, just will not be using greedy in that case
[22:43] <tjaalton> right
[22:43] <bryce> so yeah, probably no need to change depends I guess
[22:43] <tjaalton> yep, so I'll just upload this one
[22:48] <tjaalton> man, I hope all those drm fixes are backported to the hardy kernel.. should fix a number of intel suspend/resume/hibernation issues
[22:59] <bryce> tjaalton: ok!  I've uploaded a third patch which would do greedy for ALL chipsets, to bug 177492
[22:59] <ubotu> Launchpad bug 177492 in xserver-xorg-video-intel "EXA is balls-achingly slow" [Critical,Confirmed] https://launchpad.net/bugs/177492
[23:01] <bryce> tjaalton: all else being equal, I think shipping all-EXA on would be ideal, however I'm not sure what effect turning greedy on for everyone would have
[23:02] <bryce> tjaalton: but all the patches now exist to take whichever route seems best
[23:02] <tjaalton> at least 945 suffers from performance issues
[23:02] <bryce> with greedy?
[23:02] <tjaalton> I uploaded ubuntu6, so 965 should be covered now
[23:02] <tjaalton> ah no, with plain EXA
[23:03] <bryce> ok cool; will those bugs auto-close or should I mark them fix released myself?
[23:03] <tjaalton> they'll auto-close
[23:04] <bryce> okie.  --> lunch
[23:23] <tjaalton> hmm, I wonder if 'dput -u foo' was enough, I can't see those packages on the list yet
[23:24] <tjaalton> instead all vdr plugins are there, and those were uploaded a few minutes ago
[23:25] <tjaalton> anyway, time to get some sleep.. night!
[23:25] <Ng> cya tjaalton 
[23:30] <ubotu> New bug: #69082 in gnome-power-manager (main) "Screen goes blank regardless (dup-of: 30969)" [Undecided,New] https://launchpad.net/bugs/69082