[00:24] <Ng> is intrepid generally working for 965 folks?
[00:25] <Ng> 55fps from glxgears and compiz won't run
[00:26] <Ng> hmm, although it'll run now
[01:34] <bryce> bdmurray: your scripted -ati test request has proven quite fruitful
[05:22] <bryce> http://people.ubuntu.com/~brian/complete-graphs/xserver-xorg-video-ati/plots/xserver-xorg-video-ati-fullyear-open.png
[08:55] <tjaalton> Ng: upgrade your xserver-xorg-core
[09:06] <Ng> tjaalton: apt isn't offering me any updates for that
[09:08] <Ng> ii  xserver-xorg-core            2:1.4.99.906-2ubuntu2        Xorg X server - core server
[09:16] <Ng> also, I guess I'd still need mvo's package of evdev git to get scrollwheel emulation via hal?
[09:16]  * Ng really hopes that can be included somehow
[09:25] <tseliot> tjaalton: can you upload these files, please? http://albertomilone.com/ubuntu/newlrm/pitti/lista_ia.txt
[09:35] <tseliot> tjaalton: 173 now works with kernel 2.6.27 while 177.70 fixes a regression in text rendering
[10:06] <Ng> argh, laptop completely froze when I locked the screen
[10:07] <Ng> nothing in logs, and the highly computationally taxing screensaver I use is "Blank Screen" ;)
[11:30] <tjaalton> Ng: yep, works with the hardy kernel
[11:30] <tjaalton> don't let the screen to fall asleep
[11:30] <Ng> tjaalton: erk
[11:30] <Ng> tjaalton: any idea what it is?
[11:31] <tjaalton> seems like glxgears doesn't count the fps correctly, it's very smooth even though it's <50fps
[11:31] <Ng> I dunno, compiz feels a bit sluggish too
[11:31] <tjaalton> compared to?
[11:31] <Ng> before the upgrade
[11:31] <tjaalton> which upgrade :)
[11:32] <tjaalton> hardy -> intrepid?
[11:32] <Ng> hardy->intrepid. it's hard to be sure though, because all my compiz settings were eaten by the upgrade, so none of my tweaks of things like switching workspaces are the same now
[11:33] <Ng> but 3d oddness is a lot less important than hard lockups ;)
[11:34] <tjaalton> compiz sluggishness is probably because EXA doesn't preserve enough memory for offsceen pixmaps
[11:35] <tjaalton> around 12MB for my resolution
[11:36] <Ng> ah ok
[11:36] <Ng> do you know if there's a bug report for the screensaver lockup thing?
[11:37] <tjaalton> its not a screensaver problem
[11:38] <wgrant> Ah, is that why my laptop started hardlocking and sometimes panicking on resume today?
[11:38] <tjaalton> the screen doesn't know how to wake up when blanked
[11:38] <tjaalton> umm, driver/kernel, or whatever
[11:38] <tjaalton> anyway it seems to be a kernel problem
[11:39] <tjaalton> and I don't know if it has been filed yet
[11:39] <Ng> tjaalton: sure, I didn't mean to suggest gnome-screensaver was locking the machine
[11:39] <Ng> but whatever it is, it's pretty deep because it really does seem to hardlock - no capslock, no magic sysrq
[11:39] <tjaalton> but ssh works
[11:39] <Ng> that I haven't tried yet
[11:40] <wgrant> I've had a small amount of disk access while it has been otherwise apparently dead. So it's not a full kernel lock generally.
[11:40] <Ng> it's almost lunchtime, I'll poke at it a bit then (since I need to lock while I go out anyway) and I'll file something if I can't find it
[11:40] <Ng> hmm
[11:40] <tjaalton> works here. nothing in the xserver log or dmesg
[11:40] <tjaalton> its probably in the intel drm code
[11:40] <tjaalton> .26 is equally broken
[11:41] <Ng> I'm gonna try setting my screensaver to something other than blank screen and turn off all the display sleeping stuff
[11:41] <wgrant> Do people know about the Synaptics... issues?
[11:44] <tjaalton> wgrant: what issues? I don't have one
[11:44] <wgrant> tjaalton: Since upgrading about 10 hours ago, I can't click.
[11:44] <wgrant> And the vertical scrolling region is very thing.
[11:45] <wgrant> And the touchpad tab is gone from the Mouse capplet.
[11:45] <wgrant> (probably because the XInput device name is different)
[11:45] <wgrant> s/thing/thin/
[11:45] <wgrant> And I can click fine with the physical buttons, but not by tapping.
[11:45] <tjaalton> you had SHMConfig in xorg.conf?
[11:46] <wgrant> I didn't.
[11:46] <tjaalton> the driver has apparently changed some defaults
[11:46] <tseliot> tjaalton: did you see my message about the drivers?
[11:46] <tjaalton> tseliot: yes, downloaded
[11:46] <tjaalton> I'm at a seminar though :)
[11:46] <wgrant> It seems to be a "SynPS/2 Synaptics TouchPad" rather than "Synaptics Touchpad" now.
[11:47] <wgrant> The latter is hardcoded into g-s-d.
[11:47] <tseliot> tjaalton: ah, ok, I just wanted to be sure. No hurry
[11:47] <tjaalton> wgrant: what settings did the capplet allow you to change
[11:47] <tjaalton> ?
[11:48] <wgrant> tjaalton: Turning on and off (touchpad, tapping, horizontal scrolling, vertical scrolling).
[11:48] <tjaalton> you can do that with xinput
[11:48] <wgrant> I tried to work out how.
[11:48] <tjaalton> the driver (and xserver etc) now support input properties
[11:49] <wgrant> Oh.
[11:49] <wgrant> It's not in the manpage.
[11:49] <wgrant> But I see it on the command line.
[11:49] <tseliot> wgrant: I have the same problem: https://bugs.launchpad.net/bugs/262292
[11:50] <tseliot> wgrant: the fact that 2 touchpads show up depends on the fact that one is hardcoded in the xorg.conf
[11:50] <wgrant> tseliot: I've been subscribed since I noticed it.
[11:50] <wgrant> tseliot: Yep, I noticed that after I removed my xorg.conf earlier.
[11:50] <tseliot> oh, right
[11:50] <wgrant> tjaalton: How do I know which properties I can use?
[11:51] <tseliot> wgrant: xinput list-props <device> ?
[11:51] <wgrant> I get just one.
[11:51] <wgrant> 	Device Enabled:		1
[11:51] <wgrant> And I can set anything I want.
[11:52]  * wgrant checks the driver source.
[11:52] <wgrant> I also have some very strange graphical glitches with -intel on login and when new windows are created.
[11:53] <tseliot> wgrant: does the screen flicker?
[11:53] <tseliot> it does in my case with 2.4.1
[11:54] <tseliot> jcristau: did you say that you found what is causing the problem ^ ?
[11:55] <tjaalton> wgrant: it should work like that
[11:55] <wgrant> tseliot: It gets strange lines flashing up, and sometimes looks uninitialised for a fraction of a second.
[11:55] <tjaalton> wgrant: the props
[11:55] <wgrant> tjaalton: Yes, but brute-force isn't an effective way of working out what to set.
[11:55] <tjaalton> I mean there should be more props listed
[11:56] <tjaalton> my evdev is too old to test it with my mouse
[11:57] <wgrant> Am I looking for something like "MaxTapTime" or "Synaptics Tap Time"? I see too many different names.
[11:59] <Ng> ok, lunchtime, I'm going to lock the screen and see what happens! ;)
[11:59] <wgrant> "Connection reset by peer"
[12:26] <Ng> tjaalton: sooo, X immediately crashed when my screensaver activated ;)
[12:26] <Ng> I suspended from gdm when it restarted, and it then refused to resume
[12:28] <Ng> but I don't see anything in the Xorg.0.log.old, it says something about suspending AIGLX for a vt switch, but there's no information about a crash
[12:30] <wgrant> Aha.
[12:31] <wgrant> tjaalton: I applied the latest rev from upstream git, and I can now click.
[12:32] <wgrant> The scrolling region is still smaller, but at least it's mostly functional.
[12:34] <wgrant> Setting those through xinput didn't do any good - I can't get any effect with it.
[12:40] <wgrant> Oh, so it was deliberately turned off. I see.
[12:40] <wgrant> http://cgit.freedesktop.org/xorg/driver/xf86-input-synaptics/commit/?id=fb98432436c5e1cc69b5f4b84f625e3700e51e04, which was reverted in http://cgit.freedesktop.org/xorg/driver/xf86-input-synaptics/commit/?id=3d39926875446ef80dc7c23e1e90ce776c911f13.
[12:49] <Ng> tjaalton: filed it as bug 262605
[12:49] <Ng> I put it against -intel, but that's a guess ;)
[12:51]  * Ng puts screensaver and power settings back to normal.
[12:51] <Ng> time to see if the machine really does stay alive with ssh, from my iphone ;)
[13:01] <Ng> tjaalton: yep, the machine is still alive. I was able to kill -9 X and restart gdm and log back in
[13:08] <tjaalton> Ng: that's not the same I'm seeing
[13:08] <tjaalton> killing X doesn't bring the screen back
[13:10] <Ng> it wouldn't die from SIGTERM, but SIGKILL seemed to do the trick. It didn't actually come back to life until I restarted gdm though
[13:10] <Ng> which is a bit weird, but this whole thing is weird
[13:10] <Ng> and I noticed I get an oops on boot relating to IRQs
[13:11] <Ng> which can't possibly be a good thing ;)
[15:02] <Ng> tjaalton: I noticed something weird just now, I locked the screen on a typical "I'm walking away from my desk" reflex, then logged in and kill -9'd X, gdm came back of its own volition this time, I rebooted at gdm and instead of usplash I got a purple mess, hit ctrl-alt-f1 and instead of being a regular console I had some kind of higher resolution framebuffer thing going on
[15:02] <Ng> I dunno if that's a side effect of the modesetting jazz, or if something is trying to make me use a framebuffer console I don't want
[15:02] <Ng> I notice the boot messages are trying to run v86d for fb stuff
[15:06] <Ng> tjaalton: do you see bug 262600 too?
[15:49] <Ng> bleh, installing v86d just causes X to stop working
[18:54] <Ng> tjaalton: I disabled compiz and I can lock my screen now ;)
[18:54] <Ng> security > bling :)
[19:11] <kees> also, battery life > bling.  :)
[20:40] <bryce> tseliot: I've done the paperwork to get x-kit and screen-resolution-extra into main - #262775, #262778
[20:55] <tseliot> bryce: great, thanks a lot :-)
[20:55] <bryce> you might take a quick look at them and make sure I didn't list anything incorrectly
[20:56] <bryce> also, if you want to put out a new release with that time.time() bugfix, I can upload that today
[20:56] <tseliot> bryce: yes, please. The new release is in my bzr branch
[20:57]  * tseliot has a look at the 2 reports
[21:01] <bryce> thanks uploaded
[21:03] <tseliot> bryce: in both reports: Packaging system (debhelper/cdbs/dbs) ? cdbs
[21:04] <bryce> tseliot: ahh, ok please update
[21:14] <tseliot> bryce: ok, done
[21:19] <bryce> patched gnome-control-center is uploaded
[21:19] <bryce> tseliot: once the packages enter main, then I'll put a Recommends on it
[21:20] <bryce> but at least now if people manually install screen-resolution-settings they can test it
[21:20] <tjaalton> tseliot: umm, you've changed the 173 tarball?
[21:20] <tseliot> tjaalton: yes, because I had to add a patch
[21:21] <tseliot> and that broke the orig tarball
[21:21] <tseliot> since patches are in debian.binary
[21:21] <tjaalton> tseliot: somehow I got two tarballs, same name, different size
[21:22] <tseliot> tjaalton: one has a -1 added to the version
[21:23] <tseliot> otherwise debuild would have complained about a missing orig tarball
[21:23] <tjaalton> you can add stuff outside debian/
[21:23] <tjaalton> no need for a new tarball
[21:24] <tjaalton> or am I missing something?-)
[21:25] <tseliot> tjaalton: debuild complained about not being able to represent the changes or something like that
[21:31] <tjaalton> ok, the other tarball was because of wget failing to finish the download
[21:31] <tjaalton> uploaded
[21:32] <bryce> yesterday I'd taken the goal of just upstreaming one -ati bug, but unfortunately hardly any of the remaining -ati bugs are ready to upstream.  So I ended up triaging through almost the full list.  Knocked out quite a few invalid/fixed/misfiled though
[21:33] <bryce> http://people.ubuntu.com/~brian/complete-graphs/xserver-xorg-video-ati/plots/xserver-xorg-video-ati-halfyear-open.png
[21:33] <tjaalton> heh, that's impressive
[21:33] <tjaalton> just like it should be
[21:37] <bryce> I think we can easily achieve <100 by Intrepid release
[21:38] <tseliot> tjaalton: ok, thanks
[21:38] <bryce> I noticed tormod had asked for info on 8/8 for a bunch, so on 9/8 whichever ones received no response can be expired
[21:38] <bryce> similarly, bdmurray did a request for testing against intrepid on 8/14, so unanswered ones can be expired 9/14 
[21:39] <bryce> I've been upstreaming a bunch, so I'll go through those and pull in fixes so we can close as many of those as possible
[21:41] <bryce> tjaalton: btw there's been a purge request on xserver-xgl to remove it from universe
[21:42] <tjaalton> bryce: woohoo \o/
[21:42] <bryce> :-)
[21:59] <tjaalton> Ng: btw, try the older intel driver if it crashes too
[22:04] <Ng> tjaalton: ok, right after I figure out why 2.6.27 has killed my ethernet port ;)
[22:05] <tjaalton> heh
[22:05] <bryce> tjaalton: is read-edid not available for amd64?  (bug #242043)
[22:06] <bryce> (doesn't seem to be installable on my amd64 so I'm going to confirm it)
[22:08] <tjaalton> so it seems.. last change 2.5y ago
[22:08] <bryce> pitty, oh well
[22:08] <tjaalton> and the one before that in 2002
[22:08] <tjaalton> when there was no amd64..
[22:10] <tjaalton> Currently, get-edid builds against a customized lrmi (Linux Real Mode Interface - homepage), and thus can only be used on x86 systems (not even 64-bit). 
[22:10] <tjaalton> http://www.polypux.org/projects/read-edid/
[22:14] <tjaalton> heh, a new version was released on 25th, but still doesn't build on !x86
[22:37] <bryce> I like read-edid a lot.  I wish it were more actively maintained.
[22:53] <bryce> hmm, 1.4.2 doesn't have any significant improvements.  Basically just changing maintainer, adding an exit code on failure, and a lot of autoconf fussing
[22:53] <bryce> so not really worth merging
[22:54] <bryce> bdmurray: btw, tormod and I de-blacklisted compiz for ATI-gfx laptops
[22:54] <bryce> bdmurray: in Hardy and up until now, compiz would refuse to work if you had a laptop with an ATI graphics card
[22:54] <bryce> due mainly to a lot of bugs in the -ati driver
[22:55] <bryce> bdmurray: anyway, we expect that for *most* cards this change will be fine; people can switch compiz back on, happiness.
[22:55] <bryce> bdmurray: however note that there are probably some remaining bugs, and we may start seeing a pickup of bugs about compiz + ATI laptops, so keep an eye out for them