/srv/irclogs.ubuntu.com/2010/06/05/#ubuntu-x.txt

=== pinkrobotking is now known as apachelogger
brycehSarvatt, what kind of help do you need exactly?01:00
Sarvattcore-dev :P01:01
Sarvatthttps://wiki.ubuntu.com/X/PackageNotes01:02
brycehare there debdiffs prepared for the merges?01:05
Sarvattnope but I can come up with that before server goes in on monday, been doing what I can asking debian to update things and merging things in git01:06
brycehthat would be great01:06
Sarvattbut even things like nvidia-current need rebuilding now since xserver-xorg-core breaks every package providing the old abi01:06
Sarvattand virtualbox01:06
brycehif it's just down to package sponsoring I can probably carve out a couple hours monday to lend a hand01:06
Sarvattyeah its just sponsoring really01:07
brycehdon't worry about the proprietary drivers at this point01:07
Sarvatti do because xserver wont upgrade and will remove input/video-all if they have the prop drivers installed :)01:07
brycehit's traditional that they all break once xserver is updated, and stay broken until we receive compatible versions01:07
brycehwon't upgrade?01:08
Sarvattif it was just the drivers broken it'd be one thing but xserver wont even install now01:08
brycehhmm01:08
brycehbut that affects only users with proprietary drivers installed?01:08
Sarvattupgrading the safe upgrade stuff which didn't include -core made nvidia-current get installed on  my netbook01:08
brycehI would be highly doubtful that there are many people running proprietary drivers on maverick right now, but if there are some, then they can clean up manually01:09
Sarvattmaybe we should just not have nvidia-current provide an abi, it doesn't make sense thinking about it because the ones in maverick are compatable wth multiple video abi's01:09
brycehif you want to be nice, maybe include some hints in an email01:09
brycehhistorically, we've always had a broken period from when the new xserver appeared until the drivers built, which could last a couple days01:11
bryceha good goal would be to minimize that by getting xserver and the main foss drivers in as close to each other as possible01:12
brycehthe outlier drivers, including proprietary drivers, probably can wait for the second round after that01:12
Sarvatti'll start gathering up debdiffs :)01:12
brycehexcellent01:12
KangaroooSarvatt: ive got 3 more crashes different with Xorg. should it post to the same bug?02:19
Sarvatthttps://edge.launchpad.net/ubuntu/+ppas -- must beat language pack builders! :D03:55
KangaroooSarvatt: I added more dgb to bug 58771004:26
ubot4Launchpad bug 587710 in nvidia-graphics-drivers (Ubuntu) "if FF wit lot Gif & overloading CPU then X-boom to login (affects: 1) (heat: 6)" [Undecided,Confirmed] https://launchpad.net/bugs/58771004:26
ripps*sigh* I remember when gmpc-trunk used be in the Most Active list. That was before the developer went on hiatus.05:39
rippsWhen is dri2 sync + swap coming?05:43
rippswhat is this? Do I see a 2.6.35 kernel building for maverick? Only a matter of time til one is backported to either xorg-edgers or kernel-ppa07:58
* hyperair growls at x-x-v-i08:33
ricotzSarvatt, i am building 2.6.35-1.1 for lucid in my ppa and will copy it over to edgers if it works08:59
lucazadeadding XAANoOffscreenPixmaps to poulsbo xorg.conf fix 3D but icons/pixmaps are broken.... grrr10:40
lucazadeany hint?10:46
lucazadeadding also RENDER "disable" fix both 3D and icons.. but not optimal workaround11:06
Sarvattheads up if you are using the dynpm module option for radeon its gone in that .35 kenel so it wont load11:53
ricotzSarvatt, :( i said it is building in my ppa, so there wasnt a need to jam more builders12:02
Sarvatti  didnt look at irc when i woke up, sorry12:04
ricotzSarvatt, alright12:05
Sarvattricotz: are you patching mutter? it's pretty much useless in maverick without http://sarvatt.com/downloads/patches/0001-Use-gdk_screen_get_system_colormap-instead-of-gdk_sc.patch12:46
Sarvatti need to figure out how to get collab-maint packages working with auto-xorg-git..12:48
=== apachelogger is now known as pinkrobotking
rippsgeez, everybody and their sister is building 2.6.35 in launchpad13:07
=== pinkrobotking is now known as apachelogger
Sarvatttheres the 2048x2048 compiz fix for intel \o/ - http://cgit.freedesktop.org/mesa/mesa/commit/?id=add3260157368458501709d08a3f913ed448234f13:42
Sarvatthmm randrproto installs docs to /usr/share/doc/randrproto instead of /usr/share/doc/x11-proto-randr-dev now13:49
ricotzSarvatt, yes, i am patching mutter with it13:50
ricotzSarvatt, yesterday you mentioned a shell script for updating edgers ppa, could i have a look at it?13:53
jcristauSarvatt: it does?13:55
Sarvattyeah13:56
Sarvattin git13:56
Sarvatti just made a script that build the world automatically in debian packages and that and libxt were the failures, fixed libxt in unstable13:57
ricotzok13:58
Sarvatthttp://cgit.freedesktop.org/xorg/proto/randrproto/commit/?id=68f8fbe50792e0525ba767d854b18db4acda07ff13:58
jcristaui guess i never test built libxt13:59
jcristauah right13:59
jcristauhmm but dh_installdocs randrproto.txt should still put it in the right place14:00
Sarvattdigging up the build logs14:00
Sarvattricotz: will upload it to xorg-pkg-tools soon, working out some kinks14:01
Sarvatthttps://edge.launchpad.net/~xorg-edgers/+archive/ppa/+build/1774308/+files/buildlog_ubuntu-lucid-i386.x11proto-randr_1.3.1+git20100604.6ecbca5e-0ubuntu0sarvatt~lucid_FAILEDTOBUILD.txt.gz14:01
ricotzSarvatt, thanks14:02
Sarvattdh_install: x11proto-randr-dev missing files (usr/share/doc/x11proto-randr-dev/*), aborting14:03
jcristauwtf14:03
jcristauyeah.  wonder what's going on there14:03
jcristauhum14:04
jcristau../configure: line 4078: PKG_PROG_PKG_CONFIG: command not found14:04
jcristauthis is also broken :)14:05
jcristaupkg-config not installed?14:05
Sarvattoh sheesh i fixed that in the upload before it and forgot, need to make a hook too add it to build depends14:06
Sarvattor should it be that way in git in the first place? :)14:08
SarvattBuild-Depends:14:08
Sarvatt debhelper (>= 5.0.0),14:08
Sarvatt automake,14:08
Sarvatt xutils-dev (>= 1:7.5~1)14:08
jcristaumay be easier to just give up and add it to xutils-dev Depends14:08
jcristaubut yeah, adding it to randrproto build-deps in git sounds good14:09
Sarvattokie pushing it14:10
Sarvattand depends of xutils-dev sounds good to me :)14:11
Sarvattoh noes, input-citron doesnt build against 1.8 either http://launchpadlibrarian.net/49692666/buildlog_ubuntu-lucid-amd64.xserver-xorg-input-citron_1:2.2.2%2Bgit20100604.6341ecaa-0ubuntu0sarvatt~lucid_FAILEDTOBUILD.txt.gz14:14
jcristausurprise..14:14
Sarvattbig old list here if you ever want to see build logs for anything - https://edge.launchpad.net/~xorg-edgers/+archive/ppa/+builds?build_text=&build_state=all14:18
Sarvattricotz: http://sarvatt.com/sarvatt-update-ppa.sh --  just dont run it on edgers anytime because it relies on only one person running it stores the latest commit in .lastcommit so it knows to only update the package if its been update :)14:23
Sarvattif you change the ppa to an undefined one it'll make source packages for the world14:24
ricotzthanks, will have a look14:26
ricotzSarvatt, i would be nice if the date in version-string is actually the date of git-commit14:27
ricotzLASTCOMMITDATE=$(git log ${LASTCOMMIT:7} --pretty=%ci | head -1 | awk '{ print $1 }' | sed 's/-//g')14:28
Sarvattricotz: oh but its ok to use for mesa, thats the CUSTOM_ONE target15:15
Sarvatti put the hacky hook that replaces debian/ in hooks-sarvatt15:17
Sarvattargh found yet another problem in fglrx-installer, patch pointing at the wrong file. i'm not getting any feedback on it and cant test it so i asked people to come here if they have a problem with it :)15:55
Sarvatti hate uploading things i cant test :)16:17
BernardoSarvatt: did you manage to advance in the psb drivers?17:27
jcristauSarvatt: do you remember when/how the 'Failed to submit batchbuffer: Input/output error' intel errors got fixed?17:35
=== Sinnerman is now known as Cobalt
Sarvattjcristau: there were so many different bugs causing that, it depends.. most were the "dont release *something* at load detect time" commit in the kernel that i dont think ever made it upstream17:58
Sarvattlooking for the commit, i know debian had it though17:59
Sarvatti remember bgoglin pinging it on intel-gfx saying you were carrying it17:59
Sarvattah crap, script moved mesa to a +17:59
Sarvattlibgl1-mesa-dri-gallium_7.9.0+git20100605.6d741627-0ubuntu0sarvatt_i386.deb17:59
Sarvatt :(17:59
=== albert231 is now known as albert23
Sarvattjcristau: if it was happening at suspend/resume or when idle and doing a dpms off this is the commit - https://patchwork.kernel.org/patch/80474/18:13
Sarvattthat one had really specific triggers18:13
=== |Alexia_Death| is now known as Alexia_DEath_
=== Alexia_DEath_ is now known as Alexia_Death_
Sarvattbut there are so many bugs that cause that error message18:14
* hyperair gapes at the amount of upgrades from the xorg-edgers ppa18:21
jcristauSarvatt: okay, thanks18:22
=== AlanBell is now known as ilikechickens
=== ilikechickens is now known as AlanBell
rippsdammit, the new 2.6.35 is incompatible with linuxwacom. I have Bamboo CTL-460, and it requires building a new wacom module, but I'm unable to build a module for this kernel.20:42
Sarvattit's not in 2.6.35?20:53
Sarvattis it just the kernel side that doesnt support it or xf86-input-wacom? wacom module load?20:54
Sarvattyeah i dont see the pci ids in the kernel :(20:57
Sarvattwhy haven't they pushed that to the mainline kernel yet21:08
Sarvatti mean i see it in here over a year ago21:08
Sarvattoh ok here's where they added support for your d4 device in linuxwacom, not as old as i thought - http://linuxwacom.git.sourceforge.net/git/gitweb.cgi?p=linuxwacom/linuxwacom;a=commit;h=2683f4f52f4e43f4bbabb168d1a31e78200d9d8c21:14
rippsSarvatt: I fixed the wacom issue. The module didn't work, and the code in linuxwacom-0.8.6-2/0.8.7-2 was out of date. I had to change the name of a few functions in order for the the module to build21:51
rippsin wacom_sys.c21:52
rippsyeah, but this crap should be in mainline already. No reason I should be rebuilding wacom modules at this point.21:54
rippsif this doesn't go in, they better make a dkms or something21:55

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!