[00:12] <ubotu> New bug: #153857 in xserver-xorg-video-ati (main) "games crash X or freeze system if compiz is turned on" [Undecided,New] https://launchpad.net/bugs/153857
[00:13] <ubotu> New bug: #153844 in xorg (main) "Ubuntu doesn't install inside Parallels on the Mac" [Undecided,New] https://launchpad.net/bugs/153844
[00:51] <ubotu> New bug: #154104 in xserver-xorg-video-intel (main) "glxinfo: 'Direct rendering: no' on 915GM video chipset on Gutsy" [Undecided,New] https://launchpad.net/bugs/154104
[01:25] <bryce> tepsipakki: https://wiki.ubuntu.com/X/Debugging
[01:40] <ubotu> New bug: #122703 in linux-restricted-modules-2.6.22 "Upgrade Atheros drivers to snapshot/trunk to support AR5008" [Low,Confirmed] https://launchpad.net/bugs/122703
[02:57] <pcjc2> Bryce?
[07:06] <ubotu> New bug: #154194 in linux-restricted-modules-2.6.22 (restricted) ""link is not ready" when resuming from suspend with ipw3945 " [Undecided,New] https://launchpad.net/bugs/154194
[08:42] <ubotu> New bug: #154231 in xserver-xorg-video-mga (main) "G550 (dual) DVI does not work" [Undecided,New] https://launchpad.net/bugs/154231
[09:30] <ubotu> New bug: #81512 in xserver-xorg-video-intel (main) "Attempting to open terminal in Xubuntu cause gdm restart" [Undecided,New] https://launchpad.net/bugs/81512
[09:59] <bryce> hi seb128
[09:59] <seb128> hello bryce
[12:05] <ubotu> New bug: #153915 in scim (main) "Sometimes can't rename files in nautilus (dup-of: 66104)" [Low,Invalid] https://launchpad.net/bugs/153915
[14:25] <ubotu> New bug: #154380 in xorg (main) "Screen resolution does not change in live cd" [Undecided,New] https://launchpad.net/bugs/154380
[15:04] <ubotu> New bug: #154384 in xserver-xorg-video-ati (main) "Gutsy freezes when X starts with direct rendering on radeon 9200 (worked fine on Feisty)" [Undecided,New] https://launchpad.net/bugs/154384
[15:05] <ubotu> New bug: #154390 in xserver-xorg-video-ati (main) "Display DPI way off " [Undecided,New] https://launchpad.net/bugs/154390
[18:40] <immoT-> tepsipakki, xorg.conf: empty and xorg.0.log: empty
[18:40] <immoT-> status half-configured xserver-xorg
[19:46] <ubotu> New bug: #154522 in xserver-xorg-video-ati (main) "Changing virtual terminal back to X causes crash" [Undecided,New] https://launchpad.net/bugs/154522
[20:00] <ubotu> New bug: #154528 in xserver-xorg-video-ati (main) "[gutsy] ati driver not working in gutsy" [Undecided,New] https://launchpad.net/bugs/154528
[20:53] <ubotu> New bug: #146828 in xorg (main) "gdm uses incorrect screen resolution" [High,Triaged] https://launchpad.net/bugs/146828
[21:56] <bryce> heya tormod
[21:56] <tormod> hi bryce!
[21:56] <tormod> Just installed golden Gutsy on my laptop...
[21:56] <tormod> (my ati laptop)
[21:56] <bryce> did it get the right resolution?  :-)
[21:57] <tormod> no more feisty builds from me.
[21:57] <tormod> yes, but I think dpi is wrong!
[21:57] <bryce> haha
[21:58] <tormod> I see a lot of DPI issues. Have you seen this upstream?
[21:59] <bryce> a little, although I think a lot of the DPI issues are due to ubuntu-specific issues we've not yet fixed
[22:00] <bryce> well, debian-specific at least - I think it's due to bad heuristics in the xserver postinst code
[22:00] <bryce> but I don't have proof.
[22:00] <jcristau> the xserver postinst is bad. i think we agree on that :)
[22:00] <bryce> in any case, I have a patch around somewhere to hardcode xserver to 96 dpi
[22:00] <tepsipakki> gdm should do that?
[22:01] <bryce> jcristau: heya.  Btw, I don't know if you care much about ddcprobe/xresprobe, but we've put in a bunch of little fixes to it that you might be interested in
[22:01] <tormod> bryce: well most reports are about wrongly using 96dpi instead of the real one :)
[22:02] <bryce> jcristau: the xresprobe/ddcprobe/etc. code is appalling...  lots of little logic errors, out of date edid stuff... I'm looking forward to eliminating it
[22:02] <bryce> however it works a LOT better now than it did for Feisty
[22:02] <tormod> maybe not "most", there are some funny ones too. 96x6 and such.
[22:03] <bryce> tormod: ah ok, well for Ubuntu we're already forcing to 96 dpi.  I expect this does cause problems for various cases
[22:03] <jcristau> bryce: i suppose i should have a look at that.. it'd be nice to let the server itself handle that though
[22:03] <tormod> where is it forced?
[22:07] <bryce> jcristau: yeah, I figured since we weren't going to get that bold in Gutsy, I should put time in slapping ducttape and bailing wire on our existing stuff.  About 50% of the xresprobe bugs in LP got fixed for gutsy, which I feel pretty good about.
[22:07] <jcristau> nice.
[22:08] <bryce> but for Hardy I figure we'll be either ripping it out entirely, or relegating it to some fallback case, so don't plan to put much more effort into it
[22:08] <jcristau> i'll try to spend some time to update our xresprobe package
[22:09] <bryce> however I have definitely noticed - when I started with canonical, EVERY review of Ubuntu (even glowingly positive ones) complained about resolution detection issues, and so far (knock on wood), none of the reviews I've read make any mention of resolution
[22:09] <bryce> tormod: I think seb128 forced it in one of the gnome libs
[22:16] <tormod> bryce: so all gnome apps use 96dpi and don't ask the server?
[22:17] <tormod> I guess that explains that nothing changes when I set dpi in the server with xrandr...
[22:18] <tormod> I guess there is a gconf setting for this? gnome has this reputation of being configurable in all details. wait... no
[22:18] <bryce> yup, that's how I understand it
[22:21] <bryce> heya pcjc2!
[22:21] <pcjc2> hi
[22:21] <pcjc2> still no luck on 855 HW
[22:22] <pcjc2> its a real tricky one
[22:22] <bryce> yesterday I put work into getting the debugger handbook written and up in wiki, nicely formatted:  https://wiki.ubuntu.com/X/Debugging
[22:22] <pcjc2> Mostly because I don't understand the memory management for AGP and textures
[22:22] <pcjc2> I was going to comment on that... looks really good
[22:23] <bryce> tepsipakki, tormod, pcjc2, there's still some TODO's in there to be filled in, especially in how to analyze different kinds of bugs, so please add bits you know on that
[22:23] <bryce> pcjc2: have you replied to Chris Jones?  Sounds like he's willing to be your test monkey
[22:24] <pcjc2> GAH... hate computers so much.. managed to make epiphany start full screen every time it loads, with no obvious way to stop it doing that!
[22:24] <pcjc2> I've been in touch
[22:24] <bryce> great
[22:24] <pcjc2> Unfortunately I'm running out of ideas to test
[22:24] <pcjc2> I've got debug logs from the kernel with DRI debugging turned on from Claudio
[22:25] <pcjc2> All that tells me is which bits of the DRI driver are getting used on his card. Same with some very verbose MESA debugging options turned on
[22:25] <pcjc2> I still think this must be down to a register / memory mapping getting trampled across suspend though.. but have no docs on how the 855 works to know where to look
[22:26] <pcjc2> The Mesa drivers aren't a very nice read
[22:26] <bryce> yeah
[22:26] <pcjc2> Unfortunately(?), my card uses newer drivers, so we're not exercising the same code paths. Something I did wonder about was perhaps trying to persuede the various drivers I have an 855 chip
[22:27] <pcjc2> (But don't really expect the 3D core to be backwards compatible)
[22:27] <bryce> well, maybe just summarize as much as you do know, and list out whatever questions remain, and leave it for a bit?  Maybe then others can pick it up and do more research, or you might have more ideas in the meantime
[22:27] <pcjc2> indeed.. I think it might be a good point to ask some intelligent questions of the mesa developers about how textures are mapped in memory etc..
[22:27] <bryce> yup
[22:28] <pcjc2> One lead I should have pushed earlier... is a warning in the Xorg log
[22:28] <pcjc2> "(WW) intel(0): Failed to allocate texture space."
[22:29] <bryce> one thing I think would be helpful to include in the debugger guide is a list of common warnings/errors in logs, and what they imply
[22:29] <bryce> I listed a few that I've collected (but not that one)
[22:29] <pcjc2> yet textures appear in the memory allocation for the card
[22:29] <pcjc2> I don't know what it implies.. .will have to dig.
[22:30] <bryce> often it seems like serious sounding errors are just innocuous things, while simple warnings end up being signals of very bad stuff
[22:30] <pcjc2> I think it means adding a load of printk to the kernel
[22:30] <pcjc2> Evicting pixmaps looks an interesting one, and apparently that patch 120 breaks EXA (Which I'd wondered about trying)
[22:30] <pcjc2> Perhaps there is a way to run-time test for XAA / EXA for that behaviour
[22:31]  * bryce nods
[22:31] <pcjc2> I often get warning about Vblank, but haven't seen any adverse effects to it
[22:31] <bryce> for Gutsy I'd decided not to really worry about EXA users since that's an even newer tech than many of the other new tech's we've turned on
[22:31] <pcjc2> Have you heard many complaints about Compiz causing eye-strain / headaches?
[22:31] <bryce> so that's why I ended up leaving 120 in even though it causes problems for people
[22:32] <bryce> nope, not yet
[22:32] <pcjc2> I've seen a few, and have many times in the past tried Compiz, and turned it off because it hurt my brain whilst switching windows
[22:32] <pcjc2> (I sit typing with a nasty headache)
[22:32] <bryce> however for hardy, I think we should drop 120.  EXA should be more mature by then, and even if it isn't I think we need to remove barriers to using it
[22:32] <bryce> erf
[22:33] <bryce> I get headaches with crts with low refresh rates (60hz), but with lcds and high rates it's fine
[22:34] <pcjc2> http://www.j5live.com/?p=409
[22:34] <pcjc2> this LCD is 60hz, but there is probably more persistence on it than a CRT
[22:36] <bryce> during UDS in seville spain, I got extremely sick (baaad food poisoning), and at one point I turned on compiz with the wobbly windows, and it made me so queasy I had to shut it off right away
[22:36] <bryce> but with the effects set to minimum I've not had any issues
[22:37] <bryce> but the systems I spend extended periods of time with aren't running compiz at all
[22:37] <pcjc2> I've got the low level of effects on
[22:38] <pcjc2> It might even be something about the shadowing giving a 3D effect. brain gets confused, it looks raised, but displays like its flat
[22:39] <pcjc2> I've head a bit of kickback about compiz from people who find it unstable
[22:39] <bryce> yeah
[22:39] <pcjc2> on the other hand, it does seem to win people over, make people want to try linux. Sad really
[22:40]  * bryce nods
[22:40] <bryce> at least people can be predictable that way ;-)
[22:41] <bryce> honestly, I think the biggest *true* new X feature that will get newbs productively using linux is the xrandr automatic external monitor feature, for doing projectors with laptops
[22:41] <pcjc2> That was the one feature which kept me on the i810 driver for a while
[22:42] <bryce> it was the thing i was most embarrassed about when doing presentations on my linux box
[22:42] <pcjc2> (It broke my VGA out support for a bit, but its fixed now)
[22:42] <bryce> and I know there's a ton of business user types for whom that was a showstopper
[22:42] <pcjc2> its still not "right" yet.
[22:42] <bryce> yup, but we made good solid progress on it
[22:43] <bryce> it incurred a lot of new bugs and problems - like xinerama breakage, etc.
[22:43] <pcjc2> the i810 driver would happily send my 1400x1050 LCD size to the projector, which would downsize to its native LCD resolution
[22:43] <bryce> still, I think it will prove to be well worth it
[22:43] <pcjc2> I agree
[22:43] <pcjc2> The intel driver says ok.. projector is 1024x768... which either messes up your desktop
[22:43] <pcjc2> or causes that funny "short gnome panel" bug people are reporting with TV out
[22:44] <pcjc2> What it probably ought to do, is rescale the laptop LCD to 1024x768 (whatever) to match the presentation device.
[22:44] <pcjc2> You can achieve this with the xrandr command, but not the GUI at present
[22:45]  * bryce nods
[22:45] <pcjc2> and the usual load of bugs with mode probing ;)
[22:46] <pcjc2> I think in general we're moving in the right direction. With displayconfig-gtk I can bin the code I started ages ago to setup a xinerama config
[22:47] <pcjc2> (I didn't get far, just using cairo a bit to draw some screens and let you drag them about)
[22:48] <bryce> ah interesting
[22:48] <bryce> I wonder if that could be ported in, once we have xrandr?
[22:48] <pcjc2> I was trying to go for the "MS Windows" dual head config look, but was learning Cairo at the time
[22:48] <bryce> was that written in C, or?
[22:48] <pcjc2> there wasn't anything useful.. it was just a thought exercise about how it might look
[22:48] <bryce> ahh
[22:49] <pcjc2> I can't remember if I was wokring in C or python
[22:49] <pcjc2> xrandr is the enabling feature
[22:53] <pcjc2> I was wondering.. does it make sense to release the dri modules separately from the main kernel?
[22:53] <bryce> no idea, but that's a good question
[22:55] <pcjc2> I was just thinking it would be nice to rebuild a newer dri
[22:55] <pcjc2> and send someone the deb to try
[22:55] <bryce> yeah
[22:55] <pcjc2> but I've not the time, patience or hard drive space to rebuild the whole kernel
[22:56] <bryce> well, it's on my todo list for hardy to sort out doing this
[22:56] <pcjc2> (Have been building the dri and agp subdirs and sending claudio .ko files
[22:56] <pcjc2> Hardy is another 6 months right?
[22:56] <bryce> yup
[22:57] <pcjc2> (wasn't sure if there would be more -dev time to get a LTS release polished)
[22:58] <bryce> well, during the 6 months of development we'll be focusing less on new features, and I expect we'll be less aggressive at including cutting edge upstream releases
[22:58] <bryce> e.g. I don't even plan to think about xserver 1.5
[22:59] <pcjc2> I hope to see the electronics CAD I'm a developer for have a newer version
[22:59] <pcjc2> Should be doable... just have to bug the Debian maintainer
[23:00] <pcjc2> (Ok.. have to bug our maintainer to do do a minor release with a load of bug fixes first, then...)