Sarvattlol man, I feel like an idiot00:54
Sarvatti've been trying to get that ati gallium patch working right, but every time i changed it i only installed xserver-xorg-video-ati so i wasn't even testing the new changes :)00:54
Sarvattso the little one line change works, here i was rewriting it all01:00
Sarvattyep patch looks good to go, can't seem to break it. sent it to dri-devel and xorg-devel01:25
m3gai'm trying to boot lucid on a tablet device but something during boot screws up the graphics. i've tried booting with 'single' and 'text' on the kernel command line but that doesn't help.01:43
m3gai suspect something in the kernel initialisation is tweaking something it shouldn't.01:43
m3gaif i put init=/bin/bash on the command line i get a prompt and can remount the usb-key rw. but i'd like to fix this.01:44
m3gaany clues on how to move forward?01:44
RAOFm3ga: Try with “nomodeset” on the kernel command line.  That'll disable kernel modesetting, which should stop the kernel touching your display hardware quite so much.01:45
m3gaRAOF: bingo. I have a login prompt. thanks!01:47
RAOFm3ga: What hardware is that?  Could you please file a bug?01:47
m3gaits a 'hanvon B10' tablet. what package should i log a bug against?01:48
m3ganice tablet btw. 1.3G Celeron. all intel devices. pretty sure the graphics is *not* gma500.01:49
RAOFRun “ubuntu-bug linux”.  That'll file a bug against the kernel, which is the problem here.01:49
m3gacool. will do.01:49
m3gaRAOF: false alarm. i was booting the karmic kernel by mistake. the lucid kernel does the right thing even without 'nomodeset'.02:11
Sarvattpretty sure its because it has an lvds without a lid so the old karmic kernel just thought there were no screens, they fixed that in .3202:13
m3gaah, that makes sense02:14
Sarvatti was writing patches for an insane amount of similar chinese knockoff intels with the same problem and was glad they changed the default.. crappy bioses :)02:15
m3ganow getting a kernel message '[drm:edid_is_valid] *ERROR* EDID checksum is invalid, remainder is 27'02:15
m3gais that a crappy bios? :-)02:15
Sarvattcan just ignore it unless things are broken?02:16
m3gaignoring for now02:16
m3gaRAOF: this is the hardware : http://www.hanvon.com/en/products/touchpad/touchpad_B10.html02:23
Sarvattm3ga: can you sudo lspci -vvnn | pastebinit ?02:56
Sarvattjust curious, looks like a fun toy :)02:56
m3gaSarvatt: give me half an hour or so02:57
Sarvattmost likely the wufu just isnt supported yet without external drivers, vendor id 05E1 prodict id 0100 are the important parts, looking now03:02
Sarvattm3ga: the drivers are pretty nasty, overwrites policykit, disables your sources and makes you only use their mirror in taiwan, and they only have modules for karmic even though it says its for lucid03:24
Sarvatti went over the blobs in the there but it doesnt look like any wifi chipset i know about03:27
Sarvattm3ga: just google 3dsp wifi linux and see if you can find any guides, better off just replacing it with anyther card though :)03:28
SarvattThanks a lot for you using our products. We would like to provide source code to you! but it will be very hard to maintain and support due to its complicacy. and I believe the day we open source code will come soon.03:29
Sarvatti like that response :)03:29
NafaiHi guys :)03:42
Nafaiso something I upgraded in the last couple of days is causing nvidia-settings to segfault on startup03:42
Nafaibut even after installing the dbgsym package, the gdb backtrace may not be that useful :(03:43
Nafaihttp://gist.github.com/438631 03:43
brycehNafai, lynx or meerkat?03:51
brycehalright, couldn't even begin to guess, sorry03:52
brycehI guess, figure out what it was you upgraded and try downgrading it.  ;-)03:52
Sarvatt nvidia-settings --no-config work?04:02
NafaiNope, same problem04:03
Nafaialready have that installed :(04:05
Sarvattcan ya grab debug packages for all of the rest of the depends too and get a bt full?04:11
Sarvatti can't reproduce on stock lucid, got any wacky gtk related ppa's enabled? :)04:12
Sarvatt** (nvidia-settings:3303): CRITICAL **: menu_proxy_module_load: assertion `dbusproxy != NULL' failed04:12
Sarvattyep its gtk04:12
Sarvattiget the same segfault it looks like with geany if i have indicator-appmenu open04:13
Nafaiah, that probably is it04:14
NafaiI can do a ppa purge on that04:14
Sarvattnvidia-settings is open source btw if you want to figure out why - http://cgit.freedesktop.org/~aplattner/nvidia-settings/tree/04:17
m3gaSarvatt: what kind of crack must they be smoking to think that replacing the normal sources.list is acceptable?04:27
Sarvattits to stop you from upgrading away from their policykit replac...err rootkit04:30
m3gai'm currently evaluating this for the company i work for. if it works out we'd be buying them in lots of 100004:33
m3gai'll come back to wifi after i get X working04:34
Sarvattx isnt working?04:36
Sarvattwell the touchscreen most likely has problems04:36
SarvattRAOF: i'm kind of confused by mesa/LLVM04:37
Sarvatti mean, it looks like llvm support is all or nothing04:37
Sarvatti see the llvm symbols in the libs and stuff, but llvm isn't getting pulled in as a dep?04:37
RAOFWell, it certainly wanders all through the code.04:37
RAOFThat seems odd, unless a lot of llvm is inline or something.04:38
Sarvattnot to mention it looks like enabling llvm makes the gallium stuff x86/SSE2 only so not really compatible with i686 we have going? i dunno04:39
RAOFReally?  Wow.04:39
RAOFWell, we could always… build mesa _yet_ another time :)04:39
Sarvatti enabled llvm in edgers because it was supposed to speed up swtcl chipsets with r300g but i think i'll back that out for sure.. :D04:41
Sarvatti think i hate the nvidia-graphics-drivers build system more than mesa's and it's nothing but shuffling around files...05:08
superm1Sarvatt, i thought tseliot spent a lot of time cleaning it up the last cycle05:42
superm1what's janky about it now?05:43
=== Bernardo is now known as Bernardo|Away
=== Bernardo|Away is now known as Bernardo
=== Bernardo is now known as Bernardo|Away
=== Bernardo|Away is now known as Bernardo
=== Bernardo is now known as Bernardo|Away
=== Bernardo|Away is now known as Bernardo
Sarvattoh I don't know, the 53 nested variable maze in the rules is fun to maintain when they completely redo the package layout upstream :)07:16
* Sarvatt tries to figure out what PIPE_CAP_DEPTHSTENCIL_CLEAR_SEPARATE is why its busted on nouveau07:17
RAOFSarvatt: How far through the fun of forward-porting the intel 2.7 DDX to 1.9 did you get?07:31
Sarvatt/usr/lib/dri-gallium totally breaks glx over ssh07:32
Sarvattgood question, i dont remember07:33
Sarvattlet me try to find where i was working on it07:33
RAOFBecause I was talking to (I think!) ickle in #intel-gfx, and he suggested that an essentially-equivalent task was to add a non-GEM codepath for the current DDX, and that he wasn't averse to participating in such an enterprise.07:34
Sarvattsheesh i've got 2663 directories in my git packaging directory, i know i was working on it in a stupid subdirectory like temp8/ or something too07:36
RAOFYay organisation! :)07:36
RAOFAlso, there's some more nvidia-graphics-drivers goodness in the xorg apport script now, attempting to detect manually-installed nvidia drivers and grabbing the appropriate logs for forwarding them upstream.  Feel free to suggest improvements.07:37
Sarvattwhy not just make it run nvidia-bug-report and attach those logs? :)07:38
RAOFBecause nvidia-bug-report is a little bit evil.  Particularly - it'll always dump the log in the current directory, whatever that might be.  Also, it's less easy to get at the data in nvidia-bug-report.log.gz07:43
Sarvatti can't find anything, guess I didn't do more than 2.8.0 :(07:47
Sarvattor i trashed it already07:47
Sarvattah i remember now, i was going through the list then i found http://cgit.freedesktop.org/~airlied/xf86-video-intel/ and got caught up messing with that07:48
RAOFIs that a 2.2 DDX + patches?07:48
Sarvattyeah from RHEL, backported a crapload of kernel changes for new hardware to UMS07:51
Sarvatti got kind of discouraged about 2.7.1 because I remembered that was the peak of intel driver crappiness :)07:51
Sarvattyeah jaunty!07:52
Sarvatti mean both EXA and UXA completely blew and it wasn't hardly usable07:52
RAOFI don't think I _had_ an intel graphics card for Jaunty.07:53
Sarvattoh jaunty was 2.607:53
brycehright, 2.6.  *shudder*07:54
RAOFRight.  Let's crash i965_dri.so and see if I can't get a better backtrace this time…07:57
RAOFGah.  Alternatively, it could be a bug that only appears when not built with -O007:58
brycehSarvatt, mm - http://people.ubuntu.com/~fta/ppa-dashboard/ubuntu-mozilla-daily--ppa.html08:03
Sarvattoh he's been changing that, looked different last i saw it08:03
Sarvattoo he released the source! https://code.edge.launchpad.net/~fta/+junk/ppa-dashboard.trunk08:06
RAOFSo, gaussian blur on a GM45 does not provide the world's most responsive desktop environment.08:10
hyperair965 doesn't like it either08:12
Sarvatti dont even like any blur with ambiance/radiance RGBA08:12
hyperairrather, it doesn't like any kind of blurring.08:12
RAOFSo, now lets see if using the version built with -O2 segfaults.08:15
Sarvattpython-launchpadlib in jaunty doesn't support anonymous login :(08:18
brycehSarvatt, jaunty?08:22
brycehSarvatt, might be a newer version in a ppa08:22
Sarvattyeah latest ubuntu I can run on my VPS since upstart doesn't work in openvz08:23
brycehoh bummer08:23
brycehyay lp #589811 is fixed09:08
ubot4Launchpad bug 589811 in xorg-server (Ubuntu Maverick) (and 2 other projects) "My email address is in Xorg.0.log so users email me directly for support (affects: 1) (heat: 281)" [High,Fix released] https://launchpad.net/bugs/58981109:08
Sarvatti dont even see email that isn't labeled anymore because i've got so much like that, got too trigger happy subscribing to bugs for awhile there too :)09:29
Sarvattdont think i'll get around to refreshing the lcdfilter patch in cairo, 13 hunks failed09:31
RAOFIsn't it 1am or somesuch ungodly time where you are?09:32
Sarvattand it doesn't build with gcc-4.5 apparently, boo!09:32
Sarvattdebian isn't packaging cairo-perf with this, think i'll add that09:33
Sarvattpeople are going to complain if i upload cairo without the lcdfilter patch though :(09:34
Sarvattmaybe not so much since firefox bundles cairo09:34
seb128our firefox bundles the lcdfilter patch though09:34
seb128we got lot of complain when we didn't have it09:35
Sarvattyeah thats what I mean, wont be that big a deal since crappy firefox was the big complaint :)09:35
Sarvattwhat was the trick you guys use to regenerate 99_autoreconf.patch?09:43
hyperairautoreconf -vfi?09:44
* hyperair doesn't like autoreconf patches, they're bulky.09:44
Sarvattyeah same here, just all the darn desktop packages use it. i found it in an old log, cdbs-edit-patch09:49
seb128Sarvatt, cdbs-edit-patch 99_autoreconf.patch09:51
seb128rm -r autom4te.cache09:51
seb128Sarvatt, we are moving to run autoreconf in the rules this cycle09:52
=== 17SAAXETA is now known as gomyhr
=== apachelogger is now known as fluffymaster
=== go is now known as gomyhr
Sarvattso source format 3 packages get their patches applied at source extraction time? this build kept failing because git checkouts don't work since they weren't applying the patches, plus the clean rule completely doesn't actually clean it10:22
Sarvattclean rule does         rm -f *-stamp, stamps are *-stamp-*10:22
Sarvattok no lcdfilter patch is not an option, this is horrible :)10:58
maxbSarvatt: The (slightly odd to me) convention that bzr-builddeb is encouraging is to keep the *patched* source package under version control11:44
Sarvattoh joy...11:45
=== 50UAAE0BQ is now known as gomyhr
tseliotSarvatt: have you tested nvidia 195 with kernel 2.6.35? (note I haven't had the time to test and rebuild the driver yet)13:34
Sarvattno, i've only tested 256.29 but people are saying 195.24 works with a no change rebuild against the new server. i've noticed a bug where nvidia-current was getting installed on non nvidia with a dist-upgrade once it's providing xserver-xorg-video-7 though, i think the only reason it's not happening to people now is because xserver-xorg-core breaks nvidia-current13:37
Sarvatti dropped the provides in x-updates13:37
tseliotSarvatt: I fail to see why that would happen13:38
Sarvattit's easy to test, boot an alpha 1 livecd and sudo apt-get update && sudo apt-get dist-upgrade on non nvidia once it's rebuilt in the archives and it'll offer nvidia-current13:38
Sarvatti cant figure it out either13:39
Sarvattbeen trying to for weeks13:39
=== fluffymaster is now known as apachelogger
Sarvattit happened when i moved xorg-edgers to 1.8 too13:39
tseliotI'll have to check the xserver package then13:40
Sarvattwhen it's rebuilt i'll boot the livecd and do a simulated install and try to figure it out some more13:40
Sarvattalso nvidia_supported no longer works with nvidia-256, i had to hardcode the modaliases from the last working one :(13:41
tseliotSarvatt: another thing: do you know if we have this patch in Maverick? http://bugs.freedesktop.org/attachment.cgi?id=3538313:41
tseliotI'll rework nvidia_supported too, as soon as I have some time to look at the nvidia package13:42
Sarvattyes we do13:43
Sarvattthere's an ordering issue somewhere screwing up nvidia-current but i am completely stumped about it13:44
tseliotordering issue?13:45
Sarvattyeah the order its trying to resolve packages with dist-upgrade is wacky with xserver-xorg-core having breaks on the old abi's13:46
tseliotaah, ok13:46
Sarvattif you have a machine that hasn't updated to xserver 1.8 yet, you can add the x-updates ppa and try a dist-upgrade and see what it offers to see what i'm talking about13:47
Sarvatti just went through it all again yesterday but i guess ya weren't in the channel13:48
Sarvatt<Sarvatt> The following NEW packages will be installed:13:48
Sarvatt<Sarvatt>   baobab gnome-dictionary gnome-screenshot gnome-search-tool gnome-session-common gnome-system-log libprotobuf6 libprotoc6 libx11-xcb113:48
Sarvatt<Sarvatt>   libxcb-dri2-0 linux-headers-2.6.35-2 linux-headers-2.6.35-2-generic linux-image-2.6.35-2-generic nvidia-current nvidia-settings13:48
Sarvattthats on a fresh install of alpha 1 with fglrx installed doing a dist-upgrade13:49
tseliotthat definitely doesn't look good13:49
Sarvattwith x-updates enabled13:49
Sarvattwith edgers I pinned it down to xserver-xorg-video-nv being screwed up and if i purged that it stopped trying to offer nvidia-current13:50
tseliothow did xserver-xorg-video-nv cause the system to install nvidia-current?13:51
tseliotshouldn't it be replaced by nouveau instead?13:51
tseliotnote: I haven't installed maverick or created any maverick chroot yet13:52
Sarvatti wonder if its because xserver-xorg-video-nv provides xf86-video-driver-riva128 and it thinks that has video abi 6 early in the dependency resolution14:00
Sarvattthats the only thing i can think of, kees had a similar problem earlier where he couldn't upgrade xserver-xorg-core because it thought something was providing xserver-xorg-video-6 still until he purged -nv14:01
Sarvattif you dist upgrade while something provides video-6 the first thing it does is try to remove video-all14:02
Sarvattso xserver-xorg-video-all | xserver-xorg-video-7 can get fulfilled for xserver-xorg14:03
Sarvatti bet thats it actually, i'll test it out today14:04
tseliotgood, let me know how it goes14:04
rippsI'm having trouble using my webcam with any gstreamer based client18:04
rippsI have a backtrace, would the fact I'm using xorg-edgers interfere?18:04
=== Bernardo is now known as Bernardo|Away
=== Bernardo|Away is now known as Bernardo
johanbrripps, there are apparently xv issues with recent gtk18:18
johanbryou can try launching cheese with "XLIB_SKIP_ARGB_VISUALS=1 cheese"18:19
rippsjohanbr: ah that works. I've got a ton of wrappers for that in my bin/. There needs to be a better way to handle the argb incompatibilities18:22
johanbrI'm sure people are working on it18:22
rippsI guess I'll create some wrappers for cheese and empathy18:25
=== Amaranth_ is now known as Amaranth
=== Bernardo is now known as Bernardo|Away
rippswoot! looks like wacom-dkms successfully worked when upgrading kernel.18:53
rippsLet's reboot and see if it really worked.18:54
=== Bernardo|Away is now known as Bernardo
rippslol, it looks like rbga is being disabled again for gtk19:30
=== Bernardo is now known as Bernardo|Away
=== apachelogger_ is now known as apachelogger
Sarvattkinda funny, I was actually enjoying RGBA themes now :) don't understand why they weren't enabled by default though, the only thing we got from all that seemed to be crashes?20:28
* Ng wonders if there is any visibility of issues relating to USB input devices locking up20:37
NgI've not seen it myself, but someone mentioned it to me in context of http://ubuntuforums.org/showthread.php?t=149724720:38
SarvattNg: you're evil, i actually read that.. :)20:40
Nghow is that evil?20:41
NgI never go to the forums other than when people are talking about issues to me20:41
Sarvattit's scary how much bad info is contained in one thread20:41
Sarvatt"my mouse doesn't work" "reinstall nvidia" "intel 8xx is broken change to vesa fix your mouse"20:42
Sarvattnoone even gave any details of their hardware :)20:42
Sarvattlike is it locking up after suspend, or after multiple plug ins or anything?20:44
Sarvattits almost guaranteed the problem will be in dmesg when it happens20:45
SarvattNg: did that flickering screen problem ever get fixed?20:47
seb128Sarvatt, "the only thing we got from all that seemed to be crashes?"20:48
seb128Sarvatt, we needed to land the change to know about those20:48
seb128Sarvatt, we also allow other teams to pick up and land themes etc 20:49
NgSarvatt: I don't think so, but I've not been paying lots of attention to it20:49
Sarvattoh, I mean the CSD/RGBA stuff was completely transparent to me from a user's perspective, the only thing I noticed with it enabled was the crashes until I found out I could manually enable RGBA in the theme 20:50
Sarvatti think less people would be complaining if they could actually see it in action :)20:51
Sarvattbecause it shut me up from complaining once i did20:51
seb128well we need both, we need to know about issues and we need to get it used20:52
seb128seems to design team doesn't have a theme ready to use yet20:52
seb128and we collected technical issues we needed to know about now20:52
seb128so we turn it off until we have time to work on those20:52
seb128it's likely going to be back on after alpha220:53
Sarvatti've been using ambiance with rgba and it looks great20:53
* Sarvatt nods20:53
NgSarvatt: I'll see if I can get the affected person to check dmesg when it happens and file a bug20:53
seb128in practice bratsche does our gtk work and is busy on the appmenu work now20:54
Sarvattpoke me if you can, I'll help :)20:54
seb128that should land in alpha2 for UNE20:54
seb128once that's working he will be back on rgba work20:54
seb128then we will enable it again20:54
seb128we don't win anything out of crashes right now20:54
Sarvattcairo 1.9.8 checks out good on R600 at least - http://sarvatt.com/downloads/cairo/tests/20:59
Sarvattcouldn't run the pdf tests because they need poppler >= 13.220:59
=== Bernardo|Away is now known as Bernardo
Sarvattapw: do you know if 586 is going to continue being the default target for the i386 kernel in maverick?21:35
Sarvatti imagine our binutils/gcc using i686 are already making userspace incompatible with the few processors that 586 supports that 686 doesn't21:36
apwSarvatt, i would have thought so21:40
apwif it hasn't already of course21:40
rippswhat exactly is the reason for the switch to i686?21:42
apwripps, as i understand it its time for a clean break so we can take advantage of the newer features21:43
rippsapw: what kind of features?21:44
apwripps, we are optimising for h/w that is almost non-existant, the majority case is i686 and above... they are paying a penaltiy for a very very few systems21:44
apwand as those systems are very old already and can stay on lucid for 3 years, now was deemed the time to make a change if we were21:45
apwbut as for the politics, it was not my decision, nor was i present when it was made21:45
Sarvatti for one am super happy about the change :)22:01
=== apachelogger is now known as fluffymaster
geserRAOF: do you know if moving glx.h from /usr/include/GL/glx.h to /usr/include/glx.h in maverick was intended? this seems to break some other software which checks for GL/glx.h22:53
jcristausounds like a bug22:54
geserthen the last mesa upload broke it22:55
gesermesa 7.8.1-1ubuntu has it installed in /usr/include/GL/glx.h, the same 7.8.1-2 from Debian experimental but 7.8.1-3ubuntu1 in maverick installs it now in /usr/include/glx.h22:56
Sarvattindeed it does22:57
Sarvattwant to yell at me yet jcristau? :)22:58
jcristaunot unless you insist22:59
jcristaushit happens22:59
Sarvatthmm, maybe that wasnt the most recent file list23:00
Sarvattdri.pc is screwed up in origin/ubuntu too23:00
geserlooks like the changes to mesa-common-dev.install dropped the GL directory:23:00
geser+dri/usr/include/GL/gl.h usr/include23:01
geserand so on for the other files23:01
Sarvattyeah that should be dri/usr/include/GL/gl.h usr/include/GL :(23:02
Sarvattthe merge was all screwed up, didnt fix up the new locations23:03
Sarvattyay more mesa test builds incoming :)23:05
Sarvattugh, yeah I see why, there was something strange with the build in that other things from debian/tmp/dri needed the extra path in the .install but mesa-common-dev was the exception23:20
Sarvattdri/usr/include/EGL usr/include works for libegl1-mesa-dev but mesa-common-dev needs dri/usr/include/GL/gl.h usr/include/GL23:20
Sarvattadding dri/usr/include/EGL usr/include/EGL to libegl1-mesa-dev made it install to usr/include/EGL/EGL23:21
Sarvatttest building to be sure and will fix up the ubuntu branch's bad merge after if noone else gets them :D23:22
shadeslayerSarvatt: bug 594863 regarding the above issue23:26
ubot4Launchpad bug 594863 in mesa (Ubuntu) "glx.h moved, causes FTBFS in other packages (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/59486323:26
Sarvattit'll be fixed soon if you didn't read the scrollback :)23:27
Sarvattthere's a lot more messed up in the ubuntu package than just glx.h23:28
brycehgood morning raof23:31
RAOFbryceh: Good morning23:31
* RAOF adds “test that things can build against it” to the set of pre-mesa upload checks.23:32
Sarvattdid you even merge anything?23:33
Sarvattit doesnt look like you did, weird23:33
jcristaui try to debdiff against the previous changes file usually.  but sometimes i don't have that, and sometimes i miss stuff23:34
RAOFI use git for that; I diff against debian-experimental and origin/ubuntu.23:35
Sarvatthow did you merge?23:41
RAOFWith git.23:42
RAOFYou wouldn't have seen it, because I didn't push until just then.23:42
RAOFI'll be back in ~15 minutes.  Need to shift to to my brother's house which _isn't_ being rewired today.23:43
Sarvatti mean what commands did you use? whatever you did didn't actually go through?23:43
Sarvattoh ok23:43
Sarvattlooks right to me23:49

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