[00:19] <bjsnider> nvidia has now promoted 195.36.15 from pre to official
[00:40] <LLStarks> what's all this talk about ati open drivers now suddenly becoming good?
[05:05]  * abhinav is away: Abhinav|away
[05:06]  * abhinav is away: breakfast
=== yofel_ is now known as yofel
[09:48] <knittl> weeeeeeee, working standby <3
[15:23] <abhinav> hi. would this be a good place to get help regarding lucid (10.04) issues ?
[15:24] <BUGabundo> abhinav: unless specific X problems, no
[15:24] <BUGabundo> you should try #ubuntu+1 
[15:24] <abhinav> BUGabundo: k, thanks
[15:52] <abhinav> What does the ubuntu-x-swat ppa provide that the non-ppa repos don't ? (in context of ATI radeon drivers)
[15:54] <tjaalton> nothing
[15:55] <tjaalton> there are no ati packages
[15:55] <abhinav> tjaalton: so I can comment it out ? I had it active during karmic, and it got disabled on update to lucid.
[15:55] <tjaalton> yes
[15:56] <abhinav> tjaalton: ok, I see the following on the wiki/ppa page https://launchpad.net/ubuntu/lucid/+source/xserver-xorg-video-ati
[15:56] <tjaalton> that's from the main repo
[15:57] <abhinav> tjaalton: ok, thanks :) 
=== apachelogger is now known as fedoralogger
=== verbalshadow is now known as verbalshadow_
=== verbalshadow_ is now known as verbalshadow
[16:34] <BUGabundo> how is nvidia CUDA support in lucid?
[16:37] <tjaalton> BUGabundo: included in nvidia-current
[16:37] <BUGabundo> cool
[16:38] <tjaalton> though only the driver portion of it
[16:38] <tjaalton> the sdk isn't packaged
[16:39] <tjaalton> nor is the toolkit
=== Sinnerman is now known as Cobalt
[17:07] <Sarvatt> xf86Msg(X_WARNING, "PreInit returned NULL for \"%s\". This is usually harmless.\n", idev->identifier); ? :D
[17:07] <Sarvatt> vs  xf86Msg(X_ERROR, "PreInit returned NULL for \"%s\"\n", idev->identifier);
=== yofel_ is now known as yofel
[17:44] <Sarvatt> oh boy, it'd be nice to have something similar to ppa-purge in the release upgrade process, it just disables the PPA but if that still has newer than whats in lucid it sticks around
=== Sinnerman is now known as Cobalt
[18:29] <Sarvatt> ok I have some issues with this - http://testcases.qa.ubuntu.com/Hardware/X/ProprietaryDrivers#Upgrade%20From%20Manual%20Install%20Testing
[18:29] <Sarvatt> that is *never* going to succeed
[18:29] <Sarvatt> the drivers from nvidia.com require manual reinstallation every kernel abi update
[18:33] <Sarvatt> I'm really confused about https://bugs.edge.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/522328 and that test case on the wiki because it should never succeed as its written
[18:33] <ubottu> Ubuntu bug 522328 in nvidia-graphics-drivers "[Test xpr-008] Lucid upgrade broke when upgrading from manually installed nvidia proprietary drivers" [High,Triaged]
[18:36] <Sarvatt> bryceh: what do you think about the PreInit error string change I posted above?
[18:36] <bryceh> heya
[18:36] <Sarvatt> heyo!
[18:37]  * bryceh looks
[18:38] <bryceh> hmm, well we could redo the test I guess (or mark it known-fail)
[18:39] <jcristau> Sarvatt: is that still necessary after we stop setting the driver to evdev for non-event devices?
[18:39] <bryceh> the basic thing it's trying to test is that if a user manually installs fglrx, then upgrades, the system is not going to end up upbootable
[18:48] <tjaalton> jcristau: where is it handled? the evdev rule?
[18:48] <Sarvatt> failsafeX should be the intended outcome and by all accounts it seems thats working correctly for ubuntu at least, I'm not sure about non-gdm though
[18:48] <Sarvatt> jcristau: hmm good point, will check that since we're already not setting evdev for non event devices in lucid
[18:49] <tjaalton> right, not seeing those errors anymore
[18:49] <tjaalton> so it's done :)
[18:49] <jcristau> Sarvatt: yeah i think i have that in git but not uploaded in sid
[18:50] <tjaalton> it is
[18:50] <Sarvatt> btw jcristau did you see that we missed setting evdev by default for ID_INPUT_TABLET?
[18:50] <jcristau> Sarvatt: ah?
[18:50] <bjsnider> Sarvatt, you mean xpr-008 on this page: http://testcases.qa.ubuntu.com/Hardware/X/ProprietaryDrivers#Upgrade%20From%20Manual%20Install%20Testing
[18:50]  * jcristau needs to run, can you fix it in git or file a bug?
[18:50] <bjsnider> would never succeed?
[18:51] <Sarvatt> i wasn't sure if you wanted that change in debian too but it seemed like an oversight, touchscreens that identify as tablet's (like resistive tablet overlays) that work right with evdev werent loading evdev
[18:51] <Sarvatt> sure jcristau 
[18:51] <tjaalton> doing
[18:51] <tjaalton> *fixing
[18:53] <tjaalton> pushed
[18:53] <Sarvatt> wacom needs locking down a bit to not load for every ID_INPUT_TABLET as well
[18:53] <tjaalton> yes
[18:55] <Sarvatt> https://bugs.launchpad.net/bugs/537801
[18:55] <ubottu> Ubuntu bug 537801 in xf86-input-wacom "Touchscreen doesn't work after karmic->lucid upgrade" [Medium,Confirmed]
[18:55] <tjaalton> I'll inform ron about it
[18:55] <Sarvatt> gotta run too, doh
[19:36] <tjaalton> heh, a colleague is pissed because his old laptop doesn't work with a newer X anymore (ISA bus in it) :)
[19:52] <bryceh> tjaalton, heh
[19:53] <bryceh> tjaalton, friday a bunch of us got together downtown, including kees and MostAwesomeGuy (Corbin)
[19:53] <bryceh> kees had an old laptop with an ati card which didn't work right with KMS
[19:53] <bryceh> so he spent the whole day with corbin banging his head on the bug
[19:54] <bryceh> by the end of the day I go, "You know... you could get a new laptop for just $500..."
[20:01] <tjaalton> bryceh: hehe :)
[20:02] <bjsnider> nvidia-settings control file in lucid still says "amd64 i386 lpia"
[20:03] <tjaalton> bryceh: had a quick chat with ron about wacom. I'll prepare the new version and he'll pull it (with our changes) on top of his tree and releases it sometime next week
[20:18] <bryceh> tjaalton, excellent
[20:19] <bryceh> bjsnider, feel free to post a bug report about that.  tseliot may take a look.  extra points for including a debdiff :-)
[20:21] <bjsnider> i doubt i can include such a thing, since i have no idea what it is
[20:22] <bjsnider> he has to release the new version anyway, so it's no big deal to change the control file to read "any" instead
[20:47] <Sarvatt> jcristau: you're right, we dont need to shut up the PreInit returns NULL messages now that evdev only handles event devices, I thought we would since stuff like lid switch event devices gave the error but I guess they don't anymore :)
[20:52] <Sarvatt> hmm xorg-server 1.7.6 failing to build, looks like that cursor patch ya threw in recently bryceh - http://launchpadlibrarian.net/41532730/buildlog_ubuntu-lucid-i386.xorg-server_2:1.7.6~git20100321%2Bserver-1.7-branch.c552ec12-0ubuntu0sarvatt_FAILEDTOBUILD.txt.gz
[20:52] <Sarvatt> are you sure thats still relevant?
[20:53] <jcristau> what patch?
[20:54] <Sarvatt> http://git.debian.org/?p=pkg-xorg/xserver/xorg-server.git;a=blob;f=debian/patches/110_fix-swcursor-crash.patch;h=ec2db45ab460c78305fdd2445b703654023f3f24;hb=dcd61b292cdcf379b911a7112425d59ad5a79ef7
[20:55] <Sarvatt> that was reported against hardy or intrepid, dont think its relevant anymore
[20:56] <Sarvatt>   * Add 109_fix-swcursor-crash.patch: Avoid dereferencing null pointer
[20:56] <Sarvatt>     while reloading cursors during resume.
[20:56] <Sarvatt>     (LP: #371405)
[21:08] <bryceh> 109_fix-swcursor-crash.patch is the corrected patch
[21:08] <Sarvatt> <Sarvatt> ah bryceh, ya commited the bad patch to git, it should be if (!cursor_screen_priv || !cursor_screen_priv->isUp)
[21:09] <Sarvatt> fixed it up in git
[21:10] <Sarvatt> 109 was still wrong in git, was the same as 110
[21:10] <bryceh> I fixed it before uploading it
[21:10] <Sarvatt> i dont think its still relevant on 1.7 branch though, looks like its already accounted for going by http://cgit.freedesktop.org/xorg/xserver/log/hw/xfree86/modes/xf86Cursors.c?h=server-1.7-branch
[21:12] <Sarvatt> dunno why it was messed up in git still then, no biggie though
[21:12] <Sarvatt> oops you got it too
[21:13] <Sarvatt> sorry about that
[21:14] <tjaalton> and now after all this, drop the patch?-)
[21:28] <bryceh> tjaalton, for the -ati version, apt-cache madison produces the following:
[21:28] <bryceh> http://pastebin.com/vKKmqJXc
[21:29] <bryceh> problem is that it's seeing two results for experimental and unstable
[21:29] <bryceh> These lines are being mixed up:
[21:29] <bryceh> xserver-xorg-video-ati | 1:6.12.6-1 | http://ftp.debian.org/debian/ unstable/main Sources
[21:29] <bryceh> xserver-xorg-video-ati | 1:6.9.0-1+lenny4 | http://ftp.debian.org/debian/ unstable/main Sources
[21:29] <bryceh> These are as well:
[21:29] <bryceh> xserver-xorg-video-ati | 1:6.12.192-1 | http://ftp.debian.org/debian/ experimental/main Sources
[21:29] <bryceh> xserver-xorg-video-ati | 1:6.12.99+git20100201.a887818f-1 | http://ftp.debian.org/debian/ experimental/main Sources
[21:29] <bryceh> it's not clear to me how to have it select between these
[21:30] <bryceh> prefer first one seen?
[21:33] <jcristau> largest version
[21:37] <tjaalton> yep
[21:39] <tjaalton> some archs are falling back, that's why there are several versions listed
[21:39] <tjaalton> http://packages.debian.org/search?keywords=xserver-xorg-video-ati
[21:40] <tjaalton> bryceh: btw, the list was last updated on 18th
[21:55] <Sarvatt> oh looks like that cursor patch IS still relevant on 1.7 branch (and master) http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=507916
[21:55] <ubottu> Debian bug 507916 in xserver-xorg-video-intel "xorg: Xorg segfault on swsusp resume" [Important,Open]
[21:55] <Sarvatt> i'm going to add http://sarvatt.com/downloads/patches/0001-dix-be-more-verbose-when-we-run-out-of-opcodes.patch to edgers too, it's got me curious
[21:58] <Consul_Falx> hello folks
[21:58] <Consul_Falx> am using 64bit Kubuntu Lucid beta1 on a centrino1duo laptop with ATI radeon mobility X1450 shit ...
[21:59] <Consul_Falx> actually, all things work fine, but the graphics from the very beginning is slightly but visibly distracted
[21:59] <Consul_Falx> any advice how to configure that?
[21:59] <Consul_Falx> (the system is on defaults, XOrg 7.5, using radeon open driver)
[22:00] <bryceh> tjaalton, manually updated for now
[22:01] <tjaalton> bryceh: okie
[22:01] <Sarvatt> i dont think anyone but you knows what you mean by distracted there Consul_Falx :D what do you mean?
[22:02] <Consul_Falx> Sarvatt: flickery, snowy or like that ...
[22:03] <Sarvatt> like while the ubuntu logo splash screen is on the screen?
[22:03] <Sarvatt> and goes away when the KDE loading screen comes up?
[22:04] <Sarvatt> it changes the display 3 times during boot and looks pretty fugly now especially if you boot fast and use KDE where you dont get the smooth plymouth-X transition, i dont know what you mean by flickery or snowy though
[22:05] <Sarvatt> VGACON comes up with the fsck messages, then console-setup changes the fonts on vgacon, then drmfb loads and some text goes on there sometimes before the splash is displayed on that
[22:06] <Consul_Falx> Sarvatt: it started upon the first splash used during the install process, now it flickers all time, even if a tty without xserver running is open ...
[22:06] <Sarvatt> and thats only up for a second or so unless you have to enter a password or wait for a fsck if you dont use GDM
[22:07] <Sarvatt> ok, run ubuntu-bug xserver-xorg-video-ati and put [RV515] in the title if you don't mind
[22:08] <Sarvatt> we need a bunch of logs and that'll add them to a bug report automatically
[22:08] <bryceh> attach photos to the bug to show what's going on
[22:08] <Consul_Falx> and, btw, the fsck occurs too and fails
[22:08] <bryceh> Consul_Falx, actually if you can reproduce it on regular ubuntu, not kubuntu, that would help too
[22:08] <Sarvatt> are you using multiple monitors?
[22:08] <Consul_Falx> bryceh: I suppose the flickering will not show in screenshots
[22:09] <Consul_Falx> no, I have only one set up by default during the installation
[22:09] <Consul_Falx> and I cannot afford installing another clean ubuntu distro, bryceh
[22:10] <tjaalton> just install ubuntu-desktop
[22:10] <tjaalton> it'll pull gnome
[22:11] <Consul_Falx> vu@ethereal-laptop:~$ ubuntu-bug xserver-xorg-video-ati
[22:11] <Consul_Falx> kfmclient(2094)/kdecore (KSycoca) KSycocaPrivate::openDatabase: Trying to open ksycoca from  "/var/tmp/kdecache-vu/ksycoca4"
[22:11] <Consul_Falx> Launched ok, pid = 2098
[22:11] <Consul_Falx> : Fatal IO error: client killed
[22:11] <tjaalton> kde fail
[22:11] <bryceh> Consul_Falx, well, understand my focus is ubuntu so bugs against other distros are not a priority for me to look at
[22:13] <Consul_Falx> bryceh: well, I do definitely understand... I will get down to playing with this later since there's no obvious simple solution to this and I have things to do -.-
[22:16] <Consul_Falx> so long, people
[23:05] <kklimonda> may problems with user switching be related to nouveau or should I search higher in the stack? When I login to the guest session everything is fine but when I get back I have dark screen with mouse cursor and I can't switch to ther VTs (or rather I have no login prompt on them)
[23:13] <RAOF> Logs might make it easier to determine what's happening.
[23:41] <Sarvatt> i guess mesa-common-dev should depend on libdrm-dev now? Requires.private: libdrm >= 2.4.15 in dri.pc installed there
[23:42] <jcristau> the only user of afaik is the server, afaik
[23:43] <jcristau> bah
[23:43] <jcristau> the only user of dri.pc is the server, afaik