/srv/irclogs.ubuntu.com/2009/10/03/#ubuntu-x.txt

Duke`Hi10:57
Duke`I'm trying to build deb packages with xorg-edgers's xorg-pkg-tools10:58
Duke`I run "auto-xorg-git -t "~" -g -H hooks -d origin/ubuntu -p mesa", it creates the .orig.tar.gz, the diff.gz, the .dsc, the .build and the .changes10:59
Duke`but I have no .deb file at the end. Is it expected?10:59
Duke`do I have to make the .deb manually?11:00
Duke`what's next?11:00
tormodDuke`, yes these files are the source package. to build binary packages you can run debuild -b -us -uc11:21
tormodusually we upload the source package to a PPA11:22
Duke`ok, thanks, I'll look at this command11:23
tormoddon't forget apt-get build-dep mesa before building binaries11:24
Unggnuhi all11:57
UnggnuI just wanted to ask if the patch from this report could make it into Karmic: https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/43524111:57
ubottuLaunchpad bug 435241 in xserver-xorg-video-intel "Intel KMS ignores some monitor resolutions (SHP 31)" [Undecided,Confirmed]11:57
Unggnuhttp://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=blobdiff;f=drivers/gpu/drm/drm_edid.c;h=bbcb2e22675b127ce64454e69ba7a9abe55b6cc9;hp=80cc6d06d61b6ba53ae316a8de5112c8cf434a09;hb=5c61259e6c7290082108e93815f7f72f27da14f4;hpb=26bbdadad356ec02d33657858d91675f3e9aca9411:57
Unggnu* http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=commit;h=5c61259e6c7290082108e93815f7f72f27da14f411:58
Duke`hum, is there a way to make the building use more than one CPU core?12:02
UnggnuDuke`: sure -j CORENUMBER12:03
Duke`can I pass it to debuild?12:03
Unggnusome say you should use 1,5xCores12:03
UnggnuI don't know, but it works with dpkg-buildpackage12:04
Unggnuafaik12:04
Duke`I know the option for make, but I'm generating deb package12:04
Duke`ok12:04
Duke`I'll try next time I build one :)12:04
UnggnuImho the Debian-Script should recognize the core number on its own and pass it12:04
Duke`currently it's using one core of two, and it's slow :(12:05
UnggnuIt has a huge impact if you have a quad core and compile the kernel or wine or similar big things12:05
Unggnu+especially12:05
Duke`yeah, I already experienced these benefits with only a dual core :)12:07
Unggnuthat's why the build script should recognize its on their own, I mean the difference is even more severe with an i712:09
Duke`ok, I have all my deb packages \o/ It was a bit long.12:35
Duke`thank you for your help tormod12:35
tormodDuke`, np. you know you can just d/l debs from xorg-edgers also right?12:35
Duke`yes, I've done that for months, but I need to track some performance regression12:36
Duke`so I'll try to do some bisecting myself...12:36
tormodI see, recompiling the whole mesa is quite long12:36
tormodyes, you know the tricks with mesa env vars so you can just test the libs without reinstalling?12:38
Duke`aahhhh, I've heard about them once but I forgot their existence...12:38
Duke`I should look at it, yeah12:39
tormodto save you the whole debuild > install debs, rerun cycle12:39
tormodmaybe you can even git-bisect a built tree and only rerun make (or debian/rules build)12:40
Duke`without cleaning up? possible12:40
tormodfor env vars, see http://dri.freedesktop.org/wiki/TestingAndDebugging12:42
tormodhttp://dri.freedesktop.org/wiki/Building also mentions LD_PRELOAD12:47
lv_hi folks - i'm having a hard time getting my extended desktop to work with ubuntu jaunty jackolope using an intel 945GM --- if I am to go through the process of enabling UXA - should i solve this problem13:08
lv_?13:08
lv_as of now an output of lshw is showing my graphics controller to be 'unconfirmed'13:09
ScottKUrgh.  No compositing today all of a sudden on my mini 10v.  Bug report soon.18:13
ScottKMeh, of course as soon as I want to replicate the problem, it goes away18:50
Duke`__hehe19:04
=== Duke`__ is now known as Duke`
=== estan_ is now known as estan
mdeslaurhmm...I've updates karmic, and now Xorg uses 100% of the CPU all the time.21:06
mdeslaurs/updates/updated/21:06
mdeslaurthe only thing that looks related in dpkg.log is mesa21:08
pwnguinhmm21:11
pwnguinis compiz running?21:11
mdeslaurpwnguin: no, I turn off desktop effects.21:13
mdeslauralthough I tried turning compiz back on, same thing21:14
pwnguinmy guess was compiz was somehow running in software mode21:14
mdeslaurnot even touching the keyboard and Xorg is using 100% cpu21:15
pwnguinhave you browsed /var/log/Xorg.0.log?21:16
mdeslauryeah, but the only thing that looks odd is: (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)21:16
mdeslaurhold on, I'll try logging out and loggin back in21:17
mdeslaurhmm...logging out and logging back in, the problem is solved and I have "(II) Open ACPI successful (/var/run/acpid.socket)"21:20
mdeslaurlooks like X is starting before the acpid socket is there for some reason21:20
mdeslaurwhen I fresh boot21:20
jcristaushouldn't make a hugh difference though21:20
jcristauhuge, even21:20
mdeslaurhmm...I was on vt 9 after a fresh boot, and now I'm on vt 721:22
mdeslaursorry, opposite21:22
mdeslaur7 after a fresh boot, 9 now21:22
pwnguinyou might see if there are other xorg logs of recent origin21:24
mdeslaurI found one from a week ago, no difference except for the acpi failure21:25
mdeslaurbug 44170621:32
ubottuLaunchpad bug 441706 in xorg "100% cpu usage with latest updates, acpid socket missing" [Undecided,New] https://launchpad.net/bugs/44170621:32
pwnguinmdeslaur: how attached are you to 3d?21:37
jcristauoh, nvidia.  fun.21:40

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