[01:14] <Sarvatt> eww, so yeah, think I see what else might be causing some of the problems with vga16fb.. its making the agp modules load early and nvidia's agp module doesnt get used and causes problems
[01:15] <Sarvatt> just gonna get worse when agp is built into the kernel :)
[01:20] <RAOF> Sarvatt: Fortunately, we're going to be dropping the vga16fb patch that binds it to *
[01:51] <Sarvatt> i think tomorrow is gonna be transitioning edgers to the new xsfbs day after maintenance
[02:16]  * Sarvatt signed his life away for a vmware workstation trial to try this darn driver
[02:20] <RAOF> Thanks for giving that a whirl!
[02:23] <Sarvatt> i just went the and installed the 2D gallium component in libgl1-mesa-dri-gallium, the 3D side is in dri because i'm not using the same build setup you're working on
[02:24] <Sarvatt> the egl/gles stuff is all wacky and different in 7.9, debian/ from origin/ubuntu doesnt work
[02:25] <Sarvatt> plan on enabling everything and seeing whats installed in a few minutes when this is done downloading though, will probably finish compiling before i can even boot anything virtualized on this atom :)
[02:26] <Sarvatt> when is the 10.04.1 freeze? that xserver SRU really needs to go in
[02:30] <Sarvatt> guess I better build vmmouse in edgers first huh
[02:31] <Sarvatt> failed because of the xsfbs stuff forever ago and i forgot to fix it :D
[02:54] <Sarvatt> lol vmware no likey RGBA windows
[02:55] <Sarvatt> http://sarvatt.com/downloads/rgba.png
[02:59] <Sarvatt> at least i can watch irc while it loads
[03:19] <Sarvatt> heh i have to take screenshots of it to see whats going on
[03:27] <RAOF> Heh.
[03:34] <Sarvatt> saved by vnc
[03:43] <Sarvatt> well, an hour to fully install it
[03:43] <Sarvatt> i wouldnt have booted the livecd in qemu by now
[03:51] <RAOF> Lunch!
[04:41] <Sarvatt> need to make libgl1-mesa-dri-gallium depend on libkms1
[04:42] <Sarvatt> i'm doing it all over remote desktop to another machine running it in a VM since I cant use this one because of gtk in maverick.. plymouth segfaulted :)
[04:42] <Sarvatt> not sure if i'll even get 3D over RDP
[04:42] <Sarvatt> nice, it works! outside of plymouth :)
[06:21] <Bernardo> good morning
[07:35] <RAOF> Hoo.  Looks like there's a reasonable chance I'll get h264 decode acceleration on my intel chip in Maverick.
[09:59] <Bernardo> RAOF: which is your intel chip?
[10:00] <RAOF> GM45
[10:00] <Bernardo> nice. Too bad I've lost h264 decoding for now on my poulsbo, or I could brag on having it since hardy... ;)
[10:09] <Bernardo> But I do hope we get more and more hw decoding features supported
[10:28] <Dr_Jakob> Sarvatt: I'm guessing that you are not running XChat in the VM? :)
[10:52]  * Bernardo tried to get ahead on https://bugs.freedesktop.org/show_bug.cgi?id=28077, but got quickly lost in exa calls
[10:52] <ubot4> Freedesktop bug 28077 in Acceleration/EXA "X segfault in miCopyRegion / fbCopyNtoN" [Normal,New]
[11:27] <Dr_Jakob> Sarvatt: plymouth doesn't support vmwgfx?
[12:03] <alf__> RAOF: Hi!
[12:12] <RAOF> alf__: Ho!  How are those packages for you?
[12:13] <alf__> RAOF: Great, thanks!
[12:13] <alf__> RAOF: I notice that the gallium intel driver is not built by default. Is that on purpose?
[12:14] <alf__> RAOF: I built it my self and it was a bit unstable to be honest (at least as far as gles2 is concerned).
[12:14] <jcristau> considering that it doesn't work, not building it seems sane.
[12:15] <alf__> jcristau: It worked more or less for gles 1.1 examples that I tried, but segfaulted with gles2 progs.
[12:16] <alf__> jcristau: and the software rasterizer sucks :)
[12:16] <jcristau> *shrug*
[12:16] <RAOF> alf__: What intel card do you have.  I understand that i915g is less liable to instantly hang your GPU than i965g, which is what I was testing.
[12:17] <RAOF> Imagine a question mark in the first sentence :)
[12:18] <alf__> RAOF: i915. The driver actually segfaults with gles2 in userspace. Valgrind catches the invalid write.
[12:19] <RAOF> Well, that's one up on i965, which hung my GPU as soon as I tried anything - gles1, gles2, openvg.
[12:19] <alf__> :D
[12:20] <RAOF> It's possible this is better in git master, but I don't think so - Intel gallium looks pretty unloved.
[12:21] <alf__> RAOF: No problem, we actually needed the build mainly for -dev packages so others can build against mesa egl/gles*
[12:22] <RAOF> Yeah, that was what I was thinking.
[12:22] <alf__> RAOF: Is radeon support better? I am planning on buying a new desktop with an ATI/AMD card and it would be great if I could try gles* things out
[12:23] <RAOF> If you can aquire an r300-r500 card, absolutely.  There's a good chance that Maverick will ship with that gallium driver as DRI driver for those cards.
[12:24] <RAOF> r600+ is less good; the current gallium driver doesn't actually submit commands to the card yet :)
[12:25]  * alf__ goes to check what chip is in the card he is planning to buy
[12:25] <RAOF> (The current generation cards are Radeon 4xxx, which is r700, and Radeon 5xxx, which is r800/greenhorn or some other such codename)
[12:26] <RAOF> Possibly marmaduke.
[12:26] <alf__> So I am doomed :)
[12:27] <RAOF> ebay should set you up with a perfectly servicable r300-r500 for about $30
[12:27] <jcristau> i thought you could use egl_dri2 and classic dri drivers
[12:28] <RAOF> Only for opengl; they don't support gles
[12:28] <RAOF> egl_dri2 gets you OpenGL on a EGL context, as far as I understand it.
[12:28] <jcristau> k
[12:29] <alf__> RAOF: thanks for all the info!
[13:01] <lucazade> hi! new IVI xorg-x11-drv-emgd released... http://bugs.meego.com/show_bug.cgi?id=2205#c7
[13:01] <ubot4> bugs.meego.com bug 2205 in Graphics subsystem "request integration of EMGD into ivi/embedded distribution" [Major,Reopened]
[13:49] <Bernardo> hi lucazade
[13:49] <lucazade> hi Bernardo
[13:50] <Bernardo> Those are good news - if the driver at least builds. :)
[13:50] <lucazade> if it works could be a record
[13:51] <Bernardo> for now, I'd be glad if it builds. The moblin driver won't even build...
[14:00] <lucazade> Bernardo: don't know if useful, this repo contains moblin drivers for karmic http://fit-pc2.com/download/ubuntu/dists/karmic/source/
[14:10] <Bernardo> Weren't those using our drivers?
[14:11] <Bernardo> I'll check, but that seems suspiciously like our karmic drivers
[14:13] <Bernardo> bbl, time to enjoy the sun
[16:54] <trelayne> hi all, when in twinview using nvidia, when I maximize a window, it maximized between both monitors.. Anyone know how to make it maximize only on one?
[17:28] <Sarvatt> Dr_Jakob: yeah, VMware workstation doesn't like a maverick host with all of the gtk2 wackiness :)
[17:28] <Sarvatt> 7.1
[17:31] <Dr_Jakob> Sarvatt: hmm weird, have anybody filed a bug with vmware?
[17:31]  * Dr_Jakob checks the internal bugzilla...
[17:32] <Sarvatt> haven't looked yet, still messing around with it on a windows host remotely. turns out that first upload i did failed to build because it also needed xineramaproto
[17:32] <Sarvatt> and i just woke up, updating things now to see if it works
[17:33] <Sarvatt> (in the VM)
[17:34] <Dr_Jakob> ok
[17:34] <Dr_Jakob> you might need to blacklist vga16... I had problem with that whem I played around with 10.04.
[17:37] <Dr_Jakob> Hmm nope can't say anybody has filed a bug for that..
[17:37] <Dr_Jakob> but then again my bugzilla-fu is weak.
[17:53] <Sarvatt> seems to be working - http://sarvatt.com/downloads/vmwgfx.txt
[17:55] <Sarvatt> using swrast though, hmm
[17:56] <Sarvatt> ah acceleration disabled
[17:57] <Sarvatt> maybe i have to reinstall the tools or something, its enabled in the vmware settings
[17:58] <Sarvatt> -rw-r--r-- 1 root root 9768624 2010-06-01 23:51 /usr/lib/dri/vmwgfx_dri.so
[18:00] <Sarvatt> ah maybe because i'm logged into the host over remote desktop :)
[18:11] <Sarvatt> ah when it tries to use it it just dies with setting up EXA in the log
[18:12] <Sarvatt> [  114.333800] [drm:vmw_fb_setcolreg] *ERROR* Bad regno 16.
[18:12] <Sarvatt> [  114.333806] [drm:vmw_fb_setcolreg] *ERROR* Bad regno 17.
[18:12] <Sarvatt> [  114.333808] [drm:vmw_fb_setcolreg] *ERROR* Bad regno 18.
[18:12] <Sarvatt> repeated up to 255 in dmesg
[18:19] <Sarvatt> so it only works when i start it when logged into the host over RDP so the 3D isn't used, bummer
[18:30] <Sarvatt> ok gonna start transitioning xorg-edgers to the new xsfbs stuff so upgrades will probably be broken for a few hours :)
[18:31] <bryceh> Sarvatt, wishing luck
[18:31] <Sarvatt> last time i tried it in another PPA it forced nvidia-current and radeonhd to  be installed
[18:43] <Sarvatt> hmm the breaks: on  xserver-xorg-input-synaptics (<= 1.2.2-1), needs to be adjusted I think, ubuntu is at 1.2.2-1ubuntu4 thats broken by the new xserver
[18:45] <cwillu_at_work> -novtswitch is disabled these days?
[18:45] <Sarvatt> will push it just need to check the other breaks first
[18:46] <Sarvatt> "these days"?
[18:46] <cwillu_at_work> is that scare quoted because it should still work, or because it hasn't worked in forever?
[18:47] <Sarvatt> it's quoted because i dont think it was ever enabled
[18:47] <Sarvatt> don't know if you meant it was before
[18:47] <cwillu_at_work> let me restart
[18:47] <cwillu_at_work> the -novtswitch doesn't do anything;  is this recent, or has it been a noop for a while?
[18:49] <cwillu_at_work> I'm specifically looking at arm / omap
[18:50] <Sarvatt> how are you starting X?
[18:50] <Sarvatt> if its gdm you have to patch the source to change starting options afaik..
[18:51] <cwillu_at_work> Sarvatt, I'm starting x from a serial terminal:  Xorg vt2 -novtswitch
[18:53] <cwillu_at_work> looking at the source, there's a check on serverGeneration;  one way it checks VTSwitch, the other way it doesn't
[18:58] <cwillu_at_work> hw/xfree86/os-support/linux/lnx_init.c:229 is what I'm looking at (compare to :291)
[20:09]  * cwillu_at_work patiently waits for xorg to compile in a qemu-static chroot
[20:09] <Sarvatt> The following packages will be REMOVED:
[20:09] <Sarvatt>   xserver-xorg-input-all xserver-xorg-input-vmmouse xserver-xorg-video-all xserver-xorg-video-nv
[20:09] <Sarvatt> The following NEW packages will be installed:
[20:09] <Sarvatt>   dkms nvidia-current nvidia-settings xserver-xorg-video-radeonhd
[20:09] <Sarvatt> The following packages have been kept back:
[20:09] <Sarvatt>   xserver-xorg-core
[20:10] <Sarvatt> thats what i get with just xorg-server updated
[20:10] <Sarvatt> cant figure out why nvidia-current is in there
[20:18] <Sarvatt> ah thats just a dist-upgrade, upgrade is fine.. what the heck :)
[20:25] <Sarvatt> yeah i'll just do all this on maverick first and upload the lucid stuff once i work it out :)
[20:26] <Sarvatt> luckily xorg-server failed on lucid because of libpciaccess
[20:27] <cwillu_at_work> how long does a typical clean -j1 build of xorg take?
[20:27] <Sarvatt> already built?
[20:27] <cwillu_at_work> no
[20:27] <Sarvatt> xorg the metapackage?
[20:28] <cwillu_at_work> no, the big one
[20:28] <cwillu_at_work> fresh checkout, on a slow computer, but with lots and lots of memory and the build-tree on a ramdisk
[20:28] <Sarvatt> real	0m5.028s user	0m3.384s sys	0m0.872s
[20:29] <cwillu_at_work> liar
[20:30] <Sarvatt> you mean a fakeroot debian/rules clean?
[20:30] <cwillu_at_work> yes
[20:30] <Sarvatt> of xserver?
[20:30] <Sarvatt> thats my time..
[20:30] <Sarvatt> time fakeroot debian/rules clean
[20:30] <cwillu_at_work> no, I meant, building the package, _after_ a debian/rules clean :p
[20:30] <Sarvatt> oh lol
[20:31] <Sarvatt> about 20 minutes on my atom cpu
[20:31] <cwillu_at_work> trying to decide if it's worth it to cancel the build and rerun it with j4
[20:31] <cwillu_at_work> hmm
[20:31] <Sarvatt> 14 on a PPA
[20:31] <Sarvatt> probably 5 of that spent installing deps
[20:31] <cwillu_at_work> 2 hours and counting on a quad-core building it in an arm qemu
[20:33] <Sarvatt> lol yeah that sounds normal
[20:33] <Sarvatt> its about 3-4 hours for me in a arm chroot
[20:33] <cwillu_at_work> okay
[21:11] <cwillu_at_work> oooo, dh_install!  I must be nearly done
[21:12] <Sarvatt> merged xorg 7.5+6 in git. should I bump things for xserver 1.8 abi's?
[21:23] <cwillu_at_work> ding
[21:24] <Sarvatt> what xserver were you building?
[21:26] <cwillu_at_work> all of them? :p
[21:26] <cwillu_at_work> xorg-server-1.7.6; dpkg-buildpackage on armel
[21:26] <Sarvatt> ah I see why nvidia-current is getting pulled in, Provides: xserver-xorg-video-6
[21:27] <Sarvatt> and xserver-xorg depends on xserver-xorg-video-all | xserver-xorg-video-6, xserver-xorg-input-all | xserver-xorg-input-7 when the -all's were getting removed
[21:30] <Sarvatt> ah didn't know if you were building 1.8.1 from git
[21:33] <Sarvatt> -nv was build against xserver 1.7.x in edgers so its no wonder
[21:34] <Sarvatt> darn wacom and tslib not being in pkg-xorg, cant use auto-xorg-git
[21:48] <Sarvatt> doh yeah and this..
[21:48] <Sarvatt>   xserver-xorg-core: Breaks: xserver-xorg-input-7
[21:48] <Sarvatt>                      Breaks: xserver-xorg-video-6
[21:48] <Sarvatt> guess i'll downgrade the breaks some more
[21:48] <jcristau> that would be bad.
[21:49] <jcristau> unless i'm misunderstanding..
[21:56] <Sarvatt> oh what the heck, nvidia-current Provides: xserver-xorg-video-7 so that doesn't explain it
[21:56] <Sarvatt> i must have -vv and looked at the maverick one..
[22:03] <Sarvatt> just gotta update everything providing xserver-xorg-video-6 still I guess
[22:05] <Sarvatt> i should have prepared all drivers for upload before uploading the server, next time i'll remember that
[22:18] <Sarvatt> call me lazy but as often as I update stuff I like to just say all irrelevant drivers are broken and leave it at that :) but the vast majority of those crappy ones are just easy syncs from debian-unstable so i can just script all these
[22:21] <Sarvatt> heyo tormod, you around?
[22:21] <Sarvatt> how do I set options to pass to auto-xorg-git with ppa-update?
[22:22] <tormod> hi Sarvatt, for a short while :)
[22:22] <tormod> you can in a hacky way include them in the hooks argument
[22:23] <Sarvatt> need to pass update crap drivers noone cares about that provide xserver-xorg-video-6 still :(
[22:23] <Sarvatt> mass update rather
[22:23] <Sarvatt> might as well do it now because when i move to xserver 1.9 soon i'll have the sources handy to just rebuild with 
[22:23] <tormod> I sent you my "master" ppa update script once, can you find it?
[22:24] <tormod> it had some examples of passing options thru ppa-update
[22:24] <Sarvatt> oh?
[22:24] <Sarvatt> i have like 30k files in the xorg-pkg-tools directory so i probably missed it :)
[22:24] <Sarvatt> (tab completion takes awhile to say the least)
[22:26] <tormod> the script was "go-update.sh" sent Fri, Sep 18, 2009
[22:26] <Sarvatt>  AOPS=?
[22:26] <Sarvatt> AOPTS rather
[22:26] <Sarvatt> ah thanks
[22:27] <Sarvatt> guess I could just drop the packages from video-all/input-all in the meta, that'd be easier :)
[22:27] <tormod> wouldn't be the first time we do that in xorg-edgers :)
[22:31] <Sarvatt> goodbye apm, ark, chips, cirrus, geode, i128, i740, neomagic, siliconmotion, tdfx, openchrome and voodoo!
[22:31] <Sarvatt> well it still doesnt stop things from being broken actually
[22:31] <Sarvatt> xserver-xorg-core actually breaks anything with the old abi's :(
[22:38] <Sarvatt> its too much to work around from what I can see, going to just drop the breaks..  can't account for people that installed things not in the meta and not planning on updating things like virtualbox and crap
[22:39] <Sarvatt> virtualbox-ose-guest-x11 provides xserver-xorg-video-6
[22:45] <Sarvatt> The following packages have unmet dependencies:
[22:45] <Sarvatt>   xserver-xorg-video-nv: Depends: xorg-video-abi-7.0
[22:48] <Sarvatt> ahh i rebuilt one i hacked up at a bad point trying to revert the xsfbs changes, oops
[22:51] <Sarvatt> i hate to upload all drivers like this, using this numbering scheme stops debian updates from coming automatically later unless the version gets bumped upstream and i dont watch changes in most of these