=== horstle_ is now known as horstle
[04:32] <tjaalton> woohoo, the new stack has moved, and 15 video drivers purged
[04:34] <RAOF> Woo!
[04:35] <bryceh> yay
[04:35] <bryceh> yay is one of the two words Zoë knows
[04:35] <bryceh> the other word being 'dada'
[04:35] <bryceh> (yay!)
[04:35] <tjaalton> :)
[04:36] <Sarvatt> bryceh: which was first? :)
[04:41] <bryceh> Sarvatt, of course they came together ;-)
[06:23] <ripps> okay, latest xserver-xorg updates prevent me from logging in now
[06:25] <tjaalton> using nvidia?
[06:25] <tjaalton> teh blob
[06:25] <tjaalton> heh
[06:27] <RAOF> Blobitty blob blob.
[06:27] <RAOF> BOOM!
[06:27] <tjaalton> yeah it'll do that :)
[06:28] <tjaalton> lo and behold, --enable-gl-osmesa breaks xa-vmwgfx build?
[06:29] <tjaalton> nice
[06:29] <RAOF> Win!
[06:30] <ripps> I've managed to login using gnome classic (no effects), can someone help me figure this out?
[06:30] <tjaalton> using the nvidia blob?
[06:30] <ripps> tjaalton: yes, only choice, nouveou doesn't work with my card
[06:30] <tjaalton> you're screwed then
[06:31] <tjaalton> sorry
[06:31] <ripps> that wasn't the answer i wanted... any workarounds, maybe i can help figure out what's going wrong
[06:32] <tjaalton> the driver is broken
[06:32] <tjaalton> not much we can do about it
[06:32] <tjaalton> than wait for a newer version
[06:33] <tjaalton> nvidia is aware of it AIUI
[06:33] <tjaalton> you can roll back of course
[06:33] <tjaalton> though it's kinda tedious
[06:36] <tjaalton> i'll send a note to ubuntu-devel
[06:38] <ripps> well, things seem to work well enough in gnome-classic, I can even run glxgears, so it must only be in the compositing. I can survive here for a little while
[06:38] <tjaalton> iirc it's related to glx
[06:39] <tjaalton> breaks with the current xserver
[06:46] <ripps> ha, I'm even able to use minecraft, so if it's glx, then it's pretty specific part of it that's causing it
[06:46] <tjaalton> ok maybe it wasn't that
[07:19] <tjaalton> hm it's the swrast targets that make the build fail, damn threads making the output nonlinear :)
[07:19] <tjaalton> mesa build
[07:20] <tjaalton> maybe we can drop it anyway
[07:25] <tjaalton> probably best to break the build more on the ubuntu branch :P
[07:25] <tjaalton> RAOF: quick ack; drop swrast, we have llvmpipe anyway?
[07:26] <RAOF> Does that work on arm?
[07:26] <RAOF> Or !{x86_64,i386}?
[07:26] <tjaalton> it's not built elsewhere
[07:27] <tjaalton> ie. not tested either i guess
[07:27] <RAOF> So we'd no longer have a software fallback on !{x86_64,i386}?
[07:27] <RAOF> That seems bad
[07:28] <RAOF> (Sorry for slow replies; baby)
[07:29] <tjaalton> sure
[07:29] <tjaalton> I'll check the actual failure then :)
[07:30] <tjaalton> /usr/bin/ld: cannot find -l-O2
[07:30] <tjaalton> eh
[07:30] <tjaalton> typo somewhere?
[07:33] <tjaalton> oh it was reported already
[07:33] <tjaalton> on mesa-dev
[07:36] <tjaalton> --disable-shared-glapi should do the trick
[07:37] <tjaalton> well, it got further
[07:44] <RAOF> We don't want to build with --disable-shared-glapi, though.
[07:48] <tjaalton> it doesn't make sense for swx11
[07:51] <RAOF> Oh, we're still building swx11?
[07:52] <tjaalton> that's what I meant yes :)
[07:52] <tjaalton> sorry, not swrast then
[07:52] <tjaalton> ln -f .libs/libGL.so.1.6.0 ../../../../x86_64-linux-gnu/libGL.so.1 CC     fakeglx.lo
[07:52] <tjaalton> ln: accessing `.libs/libGL.so.1.6.0': No such file or directory
[07:52] <tjaalton> how it fails now
[07:54] <RAOF> We should just dump swx11 like a month-old tuna
[07:55] <RAOF> As far as I can tell it's entirely pointless.
[07:55] <tjaalton> yeah
[07:55] <RAOF> If you'd said swx11 in the first place, rather than swrast, I'd have been all for it :)
[07:56] <tjaalton> i'll do it on the ubuntu branch though
[07:56] <tjaalton> hehe
[08:03] <mlankhorst> morning
[08:06] <tjaalton> howdy
[08:06] <tjaalton> mlankhorst: enjoy the new x in quantal :)
[08:07] <mlankhorst> I saw :)
[08:07] <mlankhorst> \o/
[08:07] <tjaalton> i hope to get mesa ready for testing today
[08:07] <mlankhorst> 804 or 8.1?
[08:08] <tjaalton> probably not a good idea to upload it before the weekend though
[08:08] <tjaalton> 8.1
[08:08] <tjaalton> 8.0.4 is there already
[08:08] <mlankhorst> indeed not, just leave it for monday :)
[08:09] <mlankhorst> although I can't argue against swx11 being gone
[08:10] <tjaalton> yeah it was the diff from edgers that should make it build
[08:10] <tjaalton> lots of packaging changes needed though
[08:14] <mlankhorst> no surprises there, they're finally fixing the make procedure :)
[08:14] <tjaalton> ooh, a single-pass build comes as a result of dropping swx11 and --enable-gl-osmesa
[08:15] <mlankhorst> yay
[08:16] <RAOF> WIN!
[08:16] <mlankhorst> thanks raof :p
[08:48] <tjaalton> RAOF: should 116_use_shared_galliumcore.diff or the equivalent be upstream?
[08:48] <RAOF> tjaalton: Bits of it are, I think.
[08:48] <RAOF> I forget precisely what happened to it.
[08:49] <tjaalton> ok just wondering since it failed to apply but can't find where it should apply
[08:49] <tjaalton> I'll probably disable it for now
[08:50] <RAOF> Yeah, should be fine.
[08:51] <RAOF> Upstream would ideally like that to be turned into “rejig things so that *all* the dri drivers are built into a single, monolithic so”
[08:55] <tjaalton> oh
[09:02] <RAOF> But has been moderately happy with what has actually gone upstream: ie - the dricore & co patches.
[09:03] <tjaalton> yeah
[09:08] <tjaalton> ok, build passed, now on to debian/*.install.in..
[09:32] <seb128> tjaalton, hey, do you have a bug reference for the nvidio,xorg 1.13 issue?
[09:33] <tjaalton> seb128: bug 1037896 seems like a good master bug
[09:33] <ubottu> Launchpad bug 1037896 in nvidia-graphics-drivers (Ubuntu) "Starting Firefox kills xserver immediately" [Undecided,Confirmed] https://launchpad.net/bugs/1037896
[09:34] <seb128> tjaalton, thanks
[09:35] <seb128> tjaalton, you don't have any bug with a stacktrace or details?
[09:35] <seb128> how come we pushed a stack which we knew would break nvidia?
[09:35] <tjaalton> seb128: no, it's nvidia so no point getting one
[09:35] <tjaalton> -release wanted to get it in
[09:36] <seb128> who in release?
[09:36] <tjaalton> -proposed is not meant for testing, but for getting the archive in shape
[09:36] <seb128> right
[09:36] <tjaalton> can't remember
[09:36] <seb128> so, why was that stack pushed to proposed when it was knew to break nvidia? ;-)
[09:37] <tjaalton> should blobs prevent us moving forward?
[09:37] <tjaalton> fglrx will not work until closer to release..
[09:37] <seb128> not that discussion again ;-)
[09:37] <tjaalton> most likely
[09:37] <tjaalton> hehe
[09:37] <tjaalton> ok
[09:38] <seb128> the issue is that breaking most of product stategy boxes a week before feature freeze is not a smart move for us
[09:38] <seb128> it means they are not able to work on compiz,unity anymore
[09:38] <seb128> and that screw us all
[09:38] <tjaalton> just put xserver on hold
[09:38] <seb128> you need to reach the people and told them before they press their morning upgrade button
[09:39] <seb128> they have been told Ubuntu was to be usable every day
[09:39] <tjaalton> I've sent an email to ubuntu-devel
[09:39] <tjaalton> meh
[09:39] <seb128> right, not everybody reads ubuntu-devel in the world :-(
[09:39] <seb128> we have lot of users out of -devel subscribers
[09:39] <tjaalton> "don't upgrade a week before FF" is what the small print should say :)
[09:39] <seb128> they upgraded a long time ago
[09:39] <seb128> on the promise that Ubuntu was usuable every day
[09:39] <tjaalton> well I can't reach out to everyone
[09:40] <seb128> and they just got screwed
[09:40] <seb128> yeah, I'm not blaming you
[09:40] <seb128> but we need to stop breaking users this way
[09:40] <seb128> we need to figure a way to put xserver on hold for a class of users
[09:40] <tjaalton> it should've been copied there last week
[09:40] <seb128> like nvidia users
[09:40] <tjaalton> that's easy
[09:41] <tjaalton> tseliot: upload a new nvidia-current that doesn't provide the new abi
[09:41] <tjaalton> why didn't I think of this before..
[09:41] <tjaalton> now lunch ->
[09:41] <seb128> why is the driver providing the abi if it's not working?
[09:41] <seb128> tjaalton, thanks
[09:41] <tseliot> tjaalton: I'm about to upload 304.37 which should work with the new abi
[09:41] <tjaalton> well it's _supposed_ to work with it
[09:42] <tjaalton> tseliot: fixing the issues?
[09:42] <seb128> tjaalton, and sorry the discussion landed on you, good work for the transition, it's nothing personal ...  it's just creating issues and it would be nice to figure how we can do better about that
[09:42] <tjaalton> no problem
[09:42] <tseliot> tjaalton: I have to test it on quantal first
[09:42] <tjaalton> well the current one is broken, you can't break it any more :)
[09:42] <tseliot> heh
[09:43] <tjaalton> anyway, now that nvidia-* (and in future fglrx?) provide the static list of abi's they work with, we can drop the latest one if there are issues..
[09:43] <tjaalton> to prevent this from happening again
[09:43] <tseliot> right
[09:44] <tjaalton> it would still allow us to move forward, just keeps the sorry users from upgrading
[09:44] <tjaalton> of course some would still say "screw it, I'm installing it by hand!" and come here later..
[09:47] <tjaalton> besides, it would be great for the devs to use vesa+llvmpipe for a change, and fix it ;)
[09:51] <om26er> Hi! there have been a bug in nvidia which have caused X to crash and return users to login screen. People are not reporting that the latest version of  nvidia driver have fixed the issue for them
[09:51] <om26er> bug 973096
[09:51] <ubottu> Launchpad bug 973096 in nvidia-graphics-drivers (Ubuntu) "Nvidia driver causes xorg crash" [Critical,Triaged] https://launchpad.net/bugs/973096
[09:51] <tjaalton> eh, this bug is old
[09:51] <om26er> how will we get it fixed in 12.04?
[09:52] <tjaalton> oh I thought you meant quantal
[09:53] <om26er> tjaalton, yeah, I am talking about 12.04
[09:59] <om26er> is there any way to provide 12.04 users with latest nvidia driver as updates? or do we just update point releases of nvidia drivers as well?
[11:10] <tjaalton> hm, glsl fails to build properly
[11:12] <tjaalton> ie. no .so
[11:13] <tjaalton> ah, airlied had noticed it as well
[11:14] <tjaalton> might be something else though, it's not a fatal error here
[12:14] <agrester_1> Is anyone available to give some emergency advice
[12:16] <tjaalton> if you're on quantal + nvidia, no
[12:17] <agrester_1> Okay, no Precise 12.04 LTS, basically heres my problem, I manually installed the Nvidia 304.37 drivers manually through the *.run from Nvidia, and each time Ubuntu pushes software updates and packages some file gets changed and the Nvidia installation gets corrupted
[12:17] <tjaalton> tseliot: had time to test the driver yet?
[12:17] <tjaalton> agrester_1: you get to keep both pieces
[12:17] <tjaalton> once you touch the upstream installer, all bets are off
[12:18] <agrester_1> If I was to use X-SWAT PPA and install drivers from there, would this endless --sanity check and reinstall dance end?
[12:20] <agrester_1> tjaalton, would the X-Swat PPA driver packages solve this problem?
[12:21] <tjaalton> i don't know, but i'm sure it would bring other issues
[12:22] <agrester_1> This is hopeless, I have to switch back to Windows
[12:22] <agrester_1> Im so demoralized by this
[12:23] <agrester_1> Most of what I do in Ubuntu these days is just fixing stuff constantly...
[12:23] <agrester_1> Rather than actually getting work done
[12:27] <tjaalton> such is life
[12:27] <agrester_1> Then Shuttleworth's idea is impossible...
[12:30] <Prf_Jakob> Btw, is it possible to test the no Unity2D quantal stuff?
[12:31] <tjaalton> Prf_Jakob: yes, just move your dri driver aside
[12:32] <tjaalton> gone ->
[12:32] <Prf_Jakob> Well that or turn off 3d in the guest :)
[12:33] <agrester_1> Heres the messed up thing...according to the Nvidia sanity checker a recent Ubuntu update has "The installed symbolic link '/usr/lib/libOpenCL.so' no longer exists.", I checked apt-file and the only possible update that could have removed that file is either Nvida-Current (which I don't have installed) or FGLRX which is part of an ATI driver package...
[12:53] <tseliot> tjaalton: I had to reinstall Quantal but I'll test the driver today
[13:05] <bjsnider> tseliot, what do you think of: https://launchpadlibrarian.net/112810892/buildlog_ubuntu-precise-amd64.nvidia-settings_304.37-0ubuntu1~precise~xup1_FAILEDTOBUILD.txt.gz
[13:07] <mlankhorst> sigh from time to time this whole reservation business makes me crazy..
[13:16] <tseliot> bjsnider: I'll investigate the issue
[13:16] <bjsnider> thanks. keep me posted
[13:17] <bjsnider> the i386 versions all build fine
[13:33] <tseliot> tjaalton: lightdm starts but then Unity dies
[13:54] <tjaalton> tseliot: ok, so please upöoad a version that drops the abi providws for the xserver
[13:55] <tseliot> tjaalton: yes, that's exactly what I plan to do. I'll have a chat with Nvidia today
[16:32] <agrester> Having gone out for a light walk and cleared my mind...if I use X-SWAT PPA to install the Nvidia 304.37 drivers will I have any problems with Nvidia dependencies and libraries being messed with by other updates or will they be left alone by apt-get?  And are there major advantages to doing X-SWAT versus using the NVIDIA-304.37.run scripts? 
[16:33] <agrester> If anyone has any advice or suggestions feel free...
[16:34] <bjsnider> https://help.ubuntu.com/community/BinaryDriverHowto/Nvidia?action=show&redirect=RestrictedDrivers%2FNVIDIA
[16:35] <bjsnider> if you used the nvidia-installer you may have damaged your system
[16:35] <bjsnider> you need to re-run the installer with --uninstall
[16:37] <agrester> Okay, then add 'ppa:ubuntu-x-swat/x-updates', update and then install nvidia-current to obtain the 304.37 drivers?
[16:39] <bjsnider> install through jockey so it works right, yes
[16:50] <agrester> Okay...
=== seb128_ is now known as seb128
[19:27] <agrester> bjsnider: thanks for the help appreciate it