[00:00] <jbarnes> on the tiled rendering section it should probably refer to pre-965 platforms not just 945
[00:00] <jbarnes> the other section looks good
[00:01] <cwillu> i.e., into the 800's?
[00:01] <cwillu> or just 910+?
[00:01] <jbarnes> yeah would have affected 8xx as well
[00:01] <bryce> got it
[00:03] <albert23> bryce: the part on /dev/dri/card0 permissions doesn't seem correct. We use hal+acl to set the permissions, not udev.
[00:04] <albert23> note the + in crw-rw----+
[00:10] <bryce> albert23: ok, was just copying hearsay off one of the performance bugs
[00:11] <bryce> albert23: I've s/udev/hal+acl/ but could you update that section of the page if it's still not correct?
[00:11] <albert23> bryce: I can take a look tomorrow. It's getting late here
[00:12] <bryce> ok
[01:32] <bryce> tjaalton: heh, I just got ready to upload the fix for 337926 and see you've beaten me
[01:38] <bryce> ...or maybe not.  Looks like you've gone to bed.  I'll go ahead and put it in.  Hopefully I am not stepping on toes too hard here...
[02:03] <bryce> http://fedora.nicubunu.ro/webcomics/ctrl-alt-backspace.svg
[02:11] <cwillu> bryce, disabled it by default just in time.  I figure we got another 18 months before alt-sysrq-k becomes well known enough to be the butt of bad webcomic jokes :p
[02:11] <bryce> cwillu: hehe
[02:11] <bryce> cwillu: hey, I want to help sort out why you're not able to set Triaged; you're still having that problem?
[02:15] <cwillu> sec
[02:17] <cwillu> no, triaged still doesn't show up on the list
[02:18] <cwillu> new/incomplete/invalid/confirmed/inprogress/committed/released
[02:18] <cwillu> (looking at bug #174427, which is on xserver-xorg-video-ati)
[02:18] <ubottu> Launchpad bug 174427 in xserver-xorg-video-ati "[RV280 9250] [AGP] Tremulous causes X server crash (signal 11) at depth 16" [Medium,Invalid] https://launchpad.net/bugs/174427
[02:18] <cwillu> (and invalid, because I've been ignoring it for years now :p)
[02:20] <jbarnes> bryce: so looks like lp 348428 is fixed but the fdo bug hasn't been updated?
[02:20] <ubottu> Launchpad bug 348428 in xserver-xorg-video-intel "Switching to another user and then to anything else causes freeze in drm_intel_bo_unreference ()" [High,Fix released] https://launchpad.net/bugs/348428
[02:20] <bryce> hmm, you're definitely on the team
[02:21] <cwillu> there was some launchpad trouble around when you confirmed me though, do things ever get out of sync like that?
[02:22] <bryce> jbarnes: well, we "fixed" it by dropping patches we think caused the regression
[02:22] <bryce> those are upstream patches so the issue still needs solved there afaik
[02:22] <jbarnes> ah ok
[02:23] <bryce> actually I just dropped one of the two patches, the other seemed less suspicious
[02:23] <bryce> weirdly, the issue that the patch was supposed to fix did not reappear on dropping it
[02:23] <jbarnes> funky
[02:24] <bryce> -intel is like magic to me
[02:24] <jbarnes> it's a twisty maze
[02:24] <cwillu> filled with cavernous bugs, all alike
[02:24] <jbarnes> something like that
[02:28] <bryce> jbarnes: btw dunno if you saw but I've switched off DRI for i865 and i810 due to freeze problems.
[02:28] <jbarnes> yeah I saw that...
[02:28] <bryce> some day it would be nice to fix that
[02:28] <jbarnes> yeah 8xx has rotted a bit, we hope to fix a bunch of 8xx bugs after 2.7.0 comes out but nothing solid yet
[02:29] <bryce> that would be sweet; roughly a quarter to a third of our -intel bugs are for 8xx chips
[02:30] <jbarnes> yeah there are a ton of centrino platforms out there
[02:31] <jcristau> bryce: intel wants people to buy new laptops, so they break the old ones ;)
[02:32] <bryce> jcristau: sometimes I wonder ;-)
[02:33] <bryce> I really lost it on one guy that complained about support for SVideo on his i845, that he'd never buy intel graphics, yada yada
[02:33] <bryce> I've heard the same complaint from nvidia and fglrx users about hardware support dropping
[02:33] <bryce> at least with -intel, users *can* contribute the effort to help maintain the old stuff
[02:34] <bryce> they just *don't*
[02:34] <bryce> (afaik)
[02:34]  * cwillu pokes bryce with a "hey, I'm a user" stick
[02:34] <cwillu> definitely can't mark bugs as triaged though
[02:34] <bryce> a rare gem you are!
[02:34]  * cwillu beams
[02:35] <bryce> yeah I'm trying to figure this out...  I wonder if you have to be an Administer for the project
[02:35] <jcristau> every once in a while a non-intel and non-redhat patch makes it upstream :)
[02:35] <cwillu> can't set importance either, which isn't terribly surprising
[02:35] <bryce> however I don't seem to be able to bless people into Adminship (I can demote people though...)
[02:35] <cwillu> I tried logging out and back in yesterday, on the off chance that it was something in my session
[02:37] <bryce> cwillu: possibly it requires being on the ubuntu-qa team... 
[02:37] <bryce> (if my reading of 118708 is correct)
[02:39] <cwillu> Should I poke ubuntu-qa then?
[02:39] <cwillu> probably not really a huge deal for now
[02:40] <bryce> one sec
[02:40] <cwillu> I mean, I can always add a triaged tag and poke you with a list
[02:40] <bryce> here try this - https://launchpad.net/~bugsquad/+join
[02:41] <cwillu> <waiting for launchpad>
[02:41] <jcristau> cwillu: that goes without saying :)
[02:42] <cwillu> oh, no, it needs to be said :p
[02:42] <bryce> we sync launchpad's refresh rate to the performance of your video driver
[02:42] <cwillu> oh, let me switch from my laptop (i945) to my desktop (nvidia :p)
[02:43] <jcristau> i thought it was to make sure you could finish your coffee break while it's loading a page
[02:43] <cwillu> I usually keep a coffee maker 2 feet to the left of my left-most monitor
[02:43] <cwillu> maybe moving it to the next room would improve launchpad's perceived performance :)
[02:44] <cwillu> bryce, no, still can't touch importance or set triaged
[02:44] <bryce> hrm
[02:44] <bryce> ok, I'll check with brian murray and figure out how to get you set up
[02:45] <cwillu> k, thanks
[02:45] <cwillu> really, as long as launchpadlibs works I can still be productive
[03:21] <bryce> cwillu: aha brian pointed me to the guidelines - https://wiki.ubuntu.com/UbuntuBugControl
[03:40] <cwillu> okay, 
[03:40] <cwillu> I'll hit the enter key instead of shift
[03:40]  * cwillu mutters
[03:40] <cwillu> I'll get that started tomorrow.  Heading off to the coffee shop with laptop in tow right away :)
[03:41]  * cwillu will poke again in 30 minutes or so
[03:52] <bryce> ok cool; movie time
[05:00] <tjaalton> bryce_: no problem, I noticed that my dev laptop was not here, so had to postpone it :)
[05:05] <bryce_> hidy ho
[05:05] <bryce_> hidie?
[05:05] <bryce_> tjaalton: how are your children doing?
[05:08] <tjaalton> bryce_: sitting on the couch, watching telly ;)
[05:09] <tjaalton> ie. doing well, still coughing though
[05:09] <tjaalton> and we finally found a place we could call "home", so it's bargaining time..
[05:11] <tjaalton> hmm, I'm sure I heard my phone beep somewhere..
[05:12] <cwillu_clone> bryce, bug #328484 strikes me as something that could be upstreamed.  Reporter is responsive, bug is reproducible, etc
[05:12] <ubottu> Launchpad bug 328484 in xserver-xorg-video-intel "[igc4] displayport monitors not detected in Intrepid" [Undecided,Confirmed] https://launchpad.net/bugs/328484
[05:13] <jcristau> keithp is already working on DP
[05:13] <cwillu_clone> He's tried the jaunty beta and a more recent jaunty daily, and confirmed that it still exists
[05:13] <cwillu_clone> ah, k, know the bug # offhand?
[05:14] <jcristau> 19995
[05:15] <jcristau> (not offhand, had to find it first :) )
[05:15] <cwillu_clone> :)
[05:15] <cwillu_clone> that's on freedesktop?
[05:15] <jcristau> yeah
[05:34] <tjaalton> bryce_: could you ack bug 355340, it'd enable proper wacom hotplug and there are people confirming that, even with a serial tablet
[05:34] <ubottu> Launchpad bug 355340 in wacom-tools "[FFe] Please allow a new version of wacom-tools" [Wishlist,Confirmed] https://launchpad.net/bugs/355340
[05:35] <tjaalton> or nack, whichever you prefer :)
[05:37] <bryce_> sure
[05:39] <bryce_> done
[05:40] <bryce_> cwillu_clone: good find.  Yes, link the upstream bug to ours, and encourage the user to collaborate with upstream on the issue
[05:41] <cwillu_clone> already done
[05:41] <cwillu_clone> I've been talking to him in #ubuntu+1 as well, although he hasn't poked me yet today :)
[05:42] <cwillu_clone> bryce, is launchpadlib known for being broken randomly?
[05:43] <jcristau> i guess i'm going to have to coordinate somewhat with ron when i finally get the new x to unstable...
[05:44] <cwillu_clone> Launchpad.get_token_and_login('just testing', STAGING_SERVICE_ROOT, cache) is giving me a type error (float is required) in an egg which I _think_ was just pulled in when I installed it
[05:45] <jcristau> tjaalton: is there anything in wacom/hal that's not upstream yet, for the hotplug stuff?
[05:53] <cwillu_clone> and I was gonna get all pythonic on launchpad's api too :(
[06:00] <cwillu_clone> ah, there we go.  Never trust easy_install, even if you didn't call it yourself in the first place
[06:01] <tjaalton> jcristau: no, it's all in 0.8.3.*
[06:01] <jcristau> tjaalton: ah, nice. thanks.
[06:02] <tjaalton> 0.8.3 also supports the new Intuos4 devices.. had to check what they look like and it made me drool :P'''
[06:04] <tjaalton> bryce_: thanks
[06:09] <bryce_> cwillu_clone: yes launchpadlib has a lot of bugs in it
[06:09] <bryce_> cwillu_clone: I've been accumulating workarounds in a utility library for stuff
[06:09] <cwillu_clone> was a 2.6 bug in httplib2 that's fixed in our repository, but not on pipy
[06:14]  * cwillu_clone 's coffee is wearing off, time to sleep
[06:17] <bryce_> night
[06:29] <cwillu> well, time to leave denny's at least :p
[06:32] <cwillu> wow, it took that long to time out?
[07:49] <Unggnu> hi all
[07:49] <Unggnu> I wanted to add a passage at the top of the Backtracing wiki for apport. Is that Ok?
[07:50] <Unggnu> Afaik Apport can only track Xorg crashes in INtrepid oder later
[07:51] <Unggnu> But it is much more easier since no second computer is needed, no extra packages and so on
[07:51] <Unggnu> and all the needed information is posted
[08:01] <bryce_> heya Unggnu
[08:01] <Unggnu> hi bryce_
[08:01] <bryce_> Unggnu: yep apport usage is a good idea
[08:02] <Unggnu> I wanted a short passage at the top.
[08:02] <bryce_> in fact mdz had suggested doing it a while back, so he'll be happy to see your work
[08:02] <Unggnu> bryce_: btw. bug #345796 has a patch from upstream which works for me and a Gentoo guy :)
[08:02] <ubottu> Launchpad bug 345796 in xserver-xorg-video-intel "[i855 i915] Xorg crashed with SIGSEGV in drm_intel_bo_unpin()" [Medium,Confirmed] https://launchpad.net/bugs/345796
[08:02] <Unggnu> ok
[08:04]  * bryce_ looks
[08:07] <bryce_> Unggnu: ok thanks, I put it on my todo list.  I'll try to get it in tomorrow
[08:07] <Unggnu> thx
[08:19] <Unggnu> hm, jumping to an anchor with spaces doesn't seem to work in the Wiki
[08:20] <bryce_> nope
[08:38] <Unggnu> https://wiki.ubuntu.com/X/Backtracing - I hope it is alright
[08:38] <Unggnu> Btw. the anchor space problem should be fixed. Makes much more work.
[08:38] <Unggnu> if you can't to a specific part of a wiki page
[08:39] <Unggnu> +point
[08:39] <Unggnu> Apport works fine with X in Jaunty and according to pitti with Intrepid also
[08:52] <bryce_> yeah
[08:52] <bryce_> probably can do it with &32; or whatever the space html symbol is
[08:53] <bryce_> looks good
[09:01] <Unggnu> no, %20 doesn't work
[09:02] <Unggnu> I guess that's the reason why Wikipedia converts them to _
[09:17] <bryce> yay, wacom-tools updated.  congrats timo
[09:39] <tjaalton> bryce_: yeah :)
[09:56] <bryce> ok, enough bug damage for one day.  bedtime.  night.
[10:00] <tjaalton> night!
=== seb128_ is now known as seb128
=== ubott2 is now known as ubottu
[17:26] <bryce> morning
[17:27] <jbarnes> hi
[20:20] <sbeattie> hrm, is there a reason /etc/X11/xinit/xinitrc is not executable?
[20:22] <bryce> sbeattie: maybe it is sourced rather than executed?
[20:22] <bryce> heya albert23
[20:23] <albert23> hi bryce
[20:23] <bryce> sweet, finally a debugging method for GPU hangs on -intel:  http://bugs.freedesktop.org/show_bug.cgi?id=17638#c28
[20:23] <ubottu> Freedesktop bug 17638 in Driver/intel "[945GM] intermittent crashes (Ring of Death) if ExaNoComposite unset" [Critical,New]
[20:24] <sbeattie> bryce: the reason I ask is that the vnc4server's default .vnc/xstartup script does "exec /etc/X11/xinit/xinitrc" by default which fails; so I'm curious whether that should be fixed in in X or vnc4server.
[20:24] <bryce> hmm, that's such a common file that I would guess if it is not executable, it's for a very good reason (security perhaps?)
[20:25] <bryce> if its non-executable state were a problem, I'd have expected to see more widespread bug reports on it
[20:25] <bryce> maybe jcristau has better wisdom on this
[20:26] <bryce> btw I've written up https://wiki.ubuntu.com/X/Troubleshooting/Freeze for helping in triage freeze / gpu lockup bugs
[20:33] <sbeattie> bryce: one thing to note is that in all the hangs I've seen, the mouse still would move, but otherwise X was frozen. You may wish to incorporate that into the Narrowing Class section.
[20:36] <bryce> I mention that in the symptoms section, but I could mention that there too
[20:37] <bryce> done
[20:46] <cwillu> sbeattie, should be sourced, xinitrc can set environment vars to configure things, which doesn't work if you exec it afaik
[20:48] <cwillu> at least, so says redhat, gentoo and debian :p
[20:53] <sbeattie> cwillu: on a fork/exec, the parent won't get the env variables, true; but a direct exec() replaces the calling processes memory with the exec()ed one.
[20:53] <sbeattie> cwillu: amusingly, bug 257724 claims its executable on fedora.
[20:53] <ubottu> Launchpad bug 257724 in vnc4 "vnc4server startup script error" [Undecided,New] https://launchpad.net/bugs/257724
[20:56] <sbeattie> anyway, sourcing in vnc4server should be fine.
[20:56] <jbarnes> bryce: so on lp #344740 it sounds like you've definitely narrowed things down to the Xv patch?
[20:56] <ubottu> Launchpad bug 344740 in xserver-xorg-video-intel "[i855] xserver-xorg-video-intel-2.6.3 : Only green window when playing movies with XV extension" [Undecided,Fix released] https://launchpad.net/bugs/344740
[20:57] <jbarnes> (upstream fdo #20956)
[20:57] <albert23> bryce:  I have updated the /dev/dri/card0 permissions part on IntelPerformance
[21:25] <tseliot> federico1: did you review my patch? http://bugzilla.gnome.org/show_bug.cgi?id=568160
[21:25] <ubottu> Gnome bug 568160 in libgnome-desktop "Gnome Settings daemon causes high CPU usage with an expensive call" [Major,Unconfirmed]
[21:33] <bryce> jbarnes: yeah once that was reverted, people experiencing the crash said the problems suddenly went away
[21:33] <bryce> the only thing I'm scratching my head over is why the green window issue when playing Xv did not come back
[21:34] <bryce> albert23: excellent :-)
[21:34] <jbarnes> what was the patch you reverted exactly?
[21:34] <mnemo> there was a fix for green video included in the mesa7.4 upload
[21:34] <mnemo> a mesa fix
[21:34] <mnemo> for people seeing green video in elisa
[21:35] <bryce> mnemo: aha
[21:35] <bryce>   * Disable 114_fix_xv_with_non_gem.patch: At the time we accepted it, it
[21:35] <bryce>     sounded a little risky, so I took it on the condition that it didn't
[21:35] <bryce>     cause regressions, which apparently we have proof that it does.
[21:35] <bryce>     (LP: #348428) (Reopen 344740)
[21:37] <mnemo> bryce: this is the bug I was thinking about --> https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/349127
[21:37] <ubottu> Ubuntu bug 349127 in mesa "Elisa displays video with greenish/purplish colors" [High,Fix released]
[21:37] <mnemo> i was able to repro that bug on my intel G45 before the mesa7.4 upload
[21:38] <bryce> jbarnes: Eric's comment "...hopefully not anger tiny-aperture systems too much" made me uncertain about the patch to begin with.
[21:38] <jbarnes> yeah
[21:38] <bryce> mnemo: kewl, so that's one less loose end, that probably explains it.
[21:41] <bryce> jbarnes: btw I wrote up a page on troubleshooting freezes - https://wiki.ubuntu.com/X/Troubleshooting/Freeze - if you know any tricks worth listing there, I'd love to add more
[21:42] <bryce> jbarnes: cworth told me about the new batchbuffer dump technique, which sounds pretty cool for the future
[21:42] <jbarnes> yeah that can be helpful
[21:43] <jbarnes> bryce: the biggest thing with freezes (and most bugs but especially freezes) is figuring out a way to reliably reproduce it
[21:44] <jbarnes> might be good to add that to the page... create a series of steps that causes the freeze reliably
[21:44] <jbarnes> if you can do that we'll probably be able to fix it
[21:44] <jbarnes> if not, it'll take awhile (like the 945 hang)
[21:44] <jbarnes> bryce: so reverting the one line change to i830_video.c allows fast user switching to work?
[21:45] <bryce> ok will add
[21:45] <bryce> jbarnes: correct
[22:51] <seb128> is xvfb-run known to be broken?
[22:51] <seb128>  $ xvfb-run ls
[22:51] <seb128>  [: 182: Illegal number: 
[22:51] <seb128>  xvfb-run: error: display :99 already in use
[22:55] <seb128> hum
[22:55] <seb128> it seems to works fine using bash
[23:01] <seb128> ok, bryce broke it with his upload a week ago
[23:01] <seb128> +    if [ "$XVFBPID" -eq "$(</tmp/.X${SERVERNUM}-lock)" ]; then
[23:02] <seb128> the $() is "" using dash
[23:02]  * seb128 opens bug
[23:17] <seb128> bug #357338
[23:17] <ubottu> Launchpad bug 357338 in xorg-server "xvfb-run broken in jaunty using dash" [High,New] https://launchpad.net/bugs/357338
[23:18] <seb128> the change also breaks multiple runs, ie pygtk call it for python 2.5 and 2.6, it detects a running instance and breaks
[23:38] <jbarnes> bryce: just updated fdo #20956 (the fast user switching thing) in case you want to notify anyone
[23:40] <bryce> sweet
[23:41] <bryce> jbarnes: btw a kernel fix is going in tonight for some of the freezes on -intel that have been reported recently.  dunno if those reports made it up to the upstream tracker
[23:41] <jbarnes> the pat related stuff?
[23:46] <bryce> the BCHBAR patch
[23:46] <jbarnes> ah ok
[23:46] <bryce> it's being discussed on #ubuntu-kernel presently if you'd like to listen in
[23:46] <jbarnes> I need to respin that one for licensing reasons... I unintentionally relicensed that code as GPL :p
[23:48] <jbarnes> bryce: I guess they must be done talking about it now :)