[02:06] <ubotu> New bug: #30388 in linux-restricted-modules-2.6.15 "content in appwindows gets unreadable when moving a window" [Medium,Invalid] https://launchpad.net/bugs/30388
[02:07] <jcristau> wtf 220036.
[02:15] <bryce_> pesky feature requests
[02:22] <ubotu> New bug: #219970 in compiz (main) "Compiz (still!) freeze on ATI X300SE (dup-of: 195051)" [Undecided,New] https://launchpad.net/bugs/219970
[09:06] <ubotu> New bug: #220135 in xorg (main) "Hard lock booting/logging out/restarting" [Undecided,New] https://launchpad.net/bugs/220135
[09:35] <ubotu> New bug: #220150 in linux-restricted-modules-2.6.24 (restricted) "Mouse cursor corruption on dual-head ATI Radeon Mobility X1300 (hardy+fglrx)" [Undecided,New] https://launchpad.net/bugs/220150
[12:55] <ubotu> New bug: #220203 in xserver-xorg-video-intel (main) "Enabled compiz cauzes Xorg to freeze when CPU under load on HH" [Undecided,New] https://launchpad.net/bugs/220203
[13:11] <Q-FUNK> guys, could it be that dexconf doesn't know about -amd or -geode?  launching without xorg.conf systematically reverts to -vesa.
[13:12] <tjaalton> dexconf doesn't need to know
[13:12] <tjaalton> it's the server
[13:12] <Q-FUNK> server misses magic to pick -amd?
[13:12] <tjaalton> yep
[13:13] <Q-FUNK> can this be patched in for Hardy?
[13:13] <Q-FUNK> cause right now, the driver fails to be found as -amd or -geode.
[13:13] <tjaalton> should be possible
[13:14] <tjaalton> patches welcome
[13:14] <Q-FUNK> ajax added this to git (and I made a quick patch in bug #214119) but I'm not sure if the magic number is correct.
[13:14] <ubotu> Launchpad bug 214119 in xserver-xorg-video-geode "[HARDY] Koolu W.E. (ION A603) does not detect higher resolution than 800x600" [High,Triaged] https://launchpad.net/bugs/214119
[13:17] <Q-FUNK> ﻿(15:15:50) ogra: why would it work with X -cofigure then ?
[13:17] <Q-FUNK> (15:16:03) ogra: test it yourself :)
[13:17] <Q-FUNK> (15:16:16) Q-FUNK: beats me
[13:17] <Q-FUNK> (15:16:52) Q-FUNK: Hardy's X doesn't have the amd vendor magic to associate the chip with the driver.
[13:17] <tjaalton> X -configure apparently writes the Driver line to the config?
[13:18] <Q-FUNK> yup
[13:18] <Q-FUNK> but why would it succesfully match the driver when generating the config and not when launching without any xorg.conf?
[13:19] <tjaalton> dunno, check the source
[13:22]  * ogra waves
[13:23] <tjaalton> ogra: howdy
[13:23] <Q-FUNK> ogra: so LTSP just calls X -configure once and then copies the resulting config to the appropriate place, before launching ldm?
[13:24] <ogra> right
[13:24] <ogra> Xorg -configure -novtswitch :1
[13:24] <ogra> is the actual call
[13:24] <Q-FUNK> ok
[13:24] <ogra> it creates xorg.conf.new which then gets handled by sed scripts for replacing of values with overrides from the user
[13:24] <Q-FUNK> which currently succeeds at finding -amd 2.7.7.7 (or my patched version 2.7.7.8 with libDDC support)
[13:25] <ogra> does it work if you try it manually as well ?
[13:26] <Q-FUNK> however, it fails to find -geode completely, even with the symbolic link in place to link  amd_drv.so to geode_drv.so
[13:26] <Q-FUNK> ogra: making Xorg -configure and then copying the xorg.xonf.new to /etc/X/xorg.conf ?  yes, it works, but only with -amd.
[13:27] <ogra> just read it, no need to copy it :)
[13:27] <ogra> is Driver "amd" in there ?
[13:28] <ogra> if so, then dexconf or dpkg-reconfigure's magic suppresses something 
[13:28] <ogra> ... something plain X finds
[13:28] <tjaalton> dexconf doesn't write Driver at all
[13:28] <ogra> it rewrites the config 
[13:28] <ogra> or did that change ?
[13:28] <tjaalton> yes, but a minimal one
[13:28] <ogra> hmm
[13:28] <jcristau> yes, dpkg-reconfigure xserver-xorg rewrites the config
[13:29] <jcristau> if you don't want the config rewritten, don't do that
[13:29] <ogra> then i dont see any logic in the fact that it works on ltsp but not on packages X
[13:29] <ogra> *packaged
[13:30] <tjaalton> pretty simple. the server xfAutoConfig.c doesn't know how to map the vendor id to "amd"
[13:30] <ogra> but why does it know that with "Xorg -configure -novtswitch :1" ?
[13:31] <ogra> if its known in he C code and no scipts fiddle wit it it  needs to generae the same Driver line in both cases
[13:31] <ogra> the only logical explanation to me would be that the scripts fiddle with the file or the preseed avlues
[13:31] <jcristau> because that's two different code paths
[13:32] <jcristau> iirc
[13:32] <ogra> jcristau, X is the same binary in both cases
[13:33] <ogra> which is where both methods should get their core information about PCI ID -> driver assignment
[13:35] <Q-FUNK> ogra: ﻿"Xorg -configure -novtswitch :1" correctly finds -geode and writes that as Driver
[13:37] <Q-FUNK> http://launchpadlibrarian.net/13589055/171_xf86AutoConfig_geode_addition.diff
[13:37] <jcristau> ogra: 'Xorg -configure' and 'start X without xorg.conf' are two different code paths in the server
[13:37] <Q-FUNK> so this is what would be missing, assuming that it's the right vendor id?
[13:37] <ogra> jcristau, im not talking about 'start X without xorg.conf'
[13:38] <tjaalton> ogra: but that's the case where it fails, which this patch should fix
[13:38] <ogra> i'm talking about the two different xorg.xonf files we get from the different methids
[13:38] <jcristau> ogra: well or 'without Driver in xorg.conf'
[13:38] <ogra> tjaalton, oh, i didnt know you guys are experimenting without any confi here 
[13:39] <tjaalton> ogra: it uses autoconfiguration for all the "missing" parts of xorg.conf..
[13:39] <tjaalton> like Driver
[13:39] <Q-FUNK> so Gutsy wrote the config including Driver name but Hardy only writes non-default values like keyboard map?
[13:39] <tjaalton> right
[13:40] <tjaalton> this has been the case since November :)
[13:40] <tjaalton> or early December
[13:40] <ogra> it would then default back to vesa, right ? 
[13:40] <tjaalton> on x86 yes
[13:41] <tjaalton> (actually there's a bug that it fails to fall back to fbdev on ppc)
[13:41] <Q-FUNK> so adding the above magic patch should allow this to work without a config or Driver line?
[13:42] <tjaalton> yes
[13:42] <jcristau> there's nothing magic about it...
[13:42] <Q-FUNK> how do I verify that I have the right ID number?
[13:43] <jcristau> lspci
[13:43] <Q-FUNK> and, ok, if we merge the above patch in Hardy, we'd probably need to call the driver geode instead of amd.
[13:43] <jcristau> why?
[13:44] <Q-FUNK> or would you rather leave it as amd, despite the driver rename?
[13:44] <jcristau> i think it doesn't matter
[13:45] <Q-FUNK> jcristau: which line in lspci's VGA section?
[13:45] <tjaalton> Subsystem
[13:45] <jcristau> the one that tells you the vendor id, which is 1022
[13:46] <tjaalton> bryce: please push your xorg-server changes to git :)
[13:46] <tjaalton> hmm, and asac did an upload too
[13:46] <Q-FUNK> jcristau: subsystem doesn't list this here.  should I paste the result of lspci -v-v-v somewhere?
[13:47] <tjaalton> -vn should be enough
[13:47] <jcristau> no, it's not subsystem
[13:47] <tjaalton> the first part of it
[13:47] <tjaalton> ?
[13:47] <jcristau> lspci -n -> 01:00.0 0300: 10de:018a (rev a2)
[13:47] <jcristau> the 10de there says it's nvidia
[13:48] <Q-FUNK> ok, just a second. lemme output this to a file
[13:49] <tjaalton> oh right, not subsystem
[13:49] <tjaalton> that was for something else :)
[13:49] <jcristau> Q-FUNK: not needed. http://pci-ids.ucw.cz/iii/?i=1022
[13:50] <Q-FUNK> 1022:2081 
[13:51] <Q-FUNK> yup
[13:51] <Q-FUNK> so we have the right number
[13:51] <jcristau> (who doubt that?)
[13:52] <Q-FUNK> so it's just a matter of merging http://launchpadlibrarian.net/13589055/171_xf86AutoConfig_geode_addition.diff except that we'd have it load "geode" instead of "amd" then?
[13:53] <tjaalton> bbl ->
[13:53] <Q-FUNK> that would make X work without Driver line and also in LTSP with Xorg -configure?
[13:54] <jcristau> that wouldn't change anything for -configure
[13:54] <Q-FUNK> jcristau: having the right number and X that actually does what it's intended to are two separate issues.
[14:00]  * jcristau gives up
[15:41] <tjaalton> :P
[16:10] <ubotu> New bug: #213265 in xkeyboard-config (main) "Keyboard layout forgotten (hardy)" [Undecided,Confirmed] https://launchpad.net/bugs/213265
[16:16] <ubotu> New bug: #219835 in ubuntu "Keyboard settings revert following restart (dup-of: 213265)" [Undecided,New] https://launchpad.net/bugs/219835
[16:16] <jcristau> those sound like gnome bugs rather than xkeyboard-config
[16:27] <ubotu> New bug: #220143 in xorg (main) "update manager crashed while updating from Gutsy to Heron" [High,Incomplete] https://launchpad.net/bugs/220143
[16:35] <tjaalton> yep..
[17:18] <Q-FUNK> komputes: reply sent :)
[17:47] <ubotu> New bug: #36978 in linux-restricted-modules-2.6.15 "Dual Head Monitor sleep gives a black screen" [Medium,New] https://launchpad.net/bugs/36978
[17:53] <pochu> I have a quick question... I'm going to buy a new graphic card, but I'm not sure whether to buy nVidia or ATI... given that ATI is going to open the closed source driver (because they haven't yet, right?) what would you suggest me to do?
[17:54] <jcristau> no, they aren't going to open the closed source driver
[18:00] <federico1> pochu: get an ati; AMD is releasing documentation for their chips
[18:07] <bryce> lmao - http://www.youtube.com/watch?v=sPv8PPl7ANU
[18:07] <pochu> oh, is it only the specs? I thought it was the driver itself, but re-reading this I see I misunderstood it: http://www.linux.com/feature/119049
[18:15] <jcristau> pochu: they're working on the free radeon driver as well, and paying novell for their radeonhd work, but fglrx will remain closed
[18:33] <bryce> pochu: AMD/ATI has been a lot more active in the X / Linux community than nVidia.
[18:46] <ubotu> New bug: #218184 in ubiquity (main) "Kubuntu 8.04 beta not finish the installation in Brazil locale (dup-of: 184651)" [Undecided,New] https://launchpad.net/bugs/218184
[18:53] <pochu> I see, thank you all for your answers
[18:54] <pochu> OTOH, I can't find a mid-range low-profile ATI card... :/
[19:10] <bryce> Initialized empty Git repository in /home/bryce/src/xorg-server/xorg-server/.git/
[19:10] <bryce> ssh: connect to host alioth.debian.org port 22: Connection timed out
[19:10] <bryce> fatal: The remote end hung up unexpectedly
[19:13] <jcristau> bryce: weird. seems to work for me
[19:15] <bryce> tjaalton: anyway, I'd like to get my last xorg-server commit into git, but it seems to be inaccessible to me
[19:15] <jcristau> bryce: ssh works?
[19:16] <bryce> jcristau: yes
[19:18] <jcristau> i just did 'git clone git.debian.org:/git/pkg-xorg/xserver/xorg-server'
[19:19] <bryce> I was trying git clone ssh://bryce-guest@alioth.debian.org/git/pkg-xorg/xserver/xorg-server.git
[19:19] <jcristau> should be the same
[19:19] <jcristau> and ssh shouldn't timeout anyway. is it reproducible?
[19:19] <bryce> yep
[19:19] <jcristau> hmm
[19:20] <bryce> aha, that time it seems to be going
[19:46] <mario_limonciell> tjaalton, so slangasek shot down fglrx-8-04 for deferring to hardy-8.04.1 right?
[19:48] <tjaalton> mario_limonciell: that's right
[19:49] <tjaalton> bryce: ok, cool
[19:49] <mario_limonciell> okay. 
[20:07] <pochu> why, oh why did I buy a slim desktop?
[20:51] <ubotu> New bug: #220345 in xserver-xorg-video-ati (main) "[hardy] kernel 2.6.24-16-generic + radeon driver = system freezes" [Undecided,New] https://launchpad.net/bugs/220345
[20:56] <ubotu> New bug: #220344 in xorg (universe) "Transparency function in Xlib API for programmers" [Undecided,New] https://launchpad.net/bugs/220344
[21:02] <jcristau> someone should reject that one... (220344)
[21:03] <bryce> ok I'll take a look
[21:07] <bryce> hmm, xcb sounds like the wrong level for such a functionality
[21:23] <tjaalton> ok, running F9 live from a usb stick.. wanted to see kernel modesetting in action
[21:24] <tjaalton> funny that it used 1360x768 by default, native is 1024x768
[21:26] <ubotu> New bug: #220361 in xorg (main) "8.10RC install fails on Dell Latitude D600" [Undecided,New] https://launchpad.net/bugs/220361
[21:37] <bryce> jcristau, seb128: there is a patch for xserver-xgl at https://bugs.launchpad.net/ubuntu/+source/xserver-xgl/+bug/215396/comments/11 that fixes xgl to report the correct version of xrandr it supports
[21:37] <ubotu> Launchpad bug 215396 in gnome-control-center "gnome-display-properties crashed with SIGSEGV in _start() [when running xgl]" [Medium,Fix released] 
[21:42] <ubotu> New bug: #220364 in xserver-xorg-video-nv (main) "Nvidia card 7150 not recognized in Hardy RC Live CD" [Undecided,New] https://launchpad.net/bugs/220364
[21:45] <jcristau> bryce: better do something like http://cgit.freedesktop.org/xorg/xserver/diff/randr/rrdispatch.c?id=3b71b0f89f1db837da91650baa0ef4bb7ef2e98f imo
[21:45] <jcristau> (with 1 as SERVER_RANDR_MINOR)
[21:45] <jcristau> rather than redefine stuff from randrproto
[21:52] <bryce> jcristau: *nod*
[22:04] <jcristau> bryce: btw debian doesn't ship xgl :)
[22:04] <bryce> ahh
[22:04] <bryce> didn't know that; ok so nevermind on that one
[22:04] <bryce> (I wish we didn't ship xgl either, but anyway)
[23:30] <bryce> federico1: hey, so how did last week's xrandr gui hacking session go?  Got many good patches?