[10:11] <tjaalton> bugbot <3
[10:11] <tjaalton> http://www.bryceharrington.org/X/Reports/ubuntu-x-swat/totals.svg
=== thopiekar_ is now known as thopiekar
[15:20] <sits> hi, is it possible to build pieces of X (i.e. the intel X drivers) just by themselves?
[15:30] <tjaalton> sure
[15:37] <sits> is it as simple as copying a single binary if I just want to change the 2D driver to an older version?
[15:38] <tjaalton> why would you want to do that?
[15:40] <sits> Basically
[15:41] <sits> to try and bisect at which point in the intel driver git a feature was added that causes the glyph font cache on my EeePC 900 to become corrupted
[15:43] <sits> tjaalton: what do you think?
[15:44] <tjaalton> hum, it doesn't even use the intel driver
[15:44] <tjaalton> no wait
[15:44] <sits> tjaalton: I'm 100% it does :)
[15:44] <tjaalton> ah, GMA 900 translates to 915
[15:44] <sits> indeed
[15:44] <tjaalton> 950 was poulsbo
[15:44] <sits> yup
[15:44] <tjaalton> well I'd rather try different kernel versions first
[15:45] <sits> :D
[15:45] <sits> tjaalton:  https://bugs.freedesktop.org/show_bug.cgi?id=36326#c17
[15:45] <ubot4> Freedesktop bug 36326 in DRM/Intel "[915GM] Characters sometimes have horizontal lines through them (glyph font corruption)" [Major,New]
[15:45] <sits> tjaalton: so would I :)
[15:46] <tjaalton> there you go then
[15:46] <sits> tjaalton: ?
[15:46] <tjaalton> you've found the problem?
[15:46] <sits> nope
[15:46] <sits> see comment 8
[15:47] <tjaalton> so a kernel change that broke things?
[15:47] <sits> see comment 7
[15:48] <tjaalton> yes, confirms that
[15:48] <tjaalton> i just don't see your point, mind telling that?
[15:48] <sits> sure
[15:48] <sits> it is not just a kernel change
[15:48] <sits> it appears to be both a kernel and a driver change
[15:49] <sits> what I should have said is that I can use a 2.6.39-rc kernel with a 10.04 Xorg userland and not see the problem
[15:49] <tjaalton> yeah, should have said that
[15:49] <sits> using an 11.04 userland with a 2.6.34 kernel does show the problem
[15:49] <sits> tjaalton: I did on comment #6
[15:50] <tjaalton> but you didn't point that out
[15:50] <tjaalton> to me
[15:50] <tjaalton> so yes, you can compile the driver separately
[15:50] <sits> tjaalton: a very good point - sorry about that
[15:50] <tjaalton> but if there are xserver api changes, you need to take care of those
[15:51] <tjaalton> though you'll find out if the driver doesn't compile
[15:51] <sits> tjaalton: are drivers API backwards compatible?
[15:51] <sits> tjaalton: that's where things get tricky
[15:51] <sits> I can't compile stuff on the EeePC (just too small)
[15:51] <tjaalton> newer ddx driver usually needs newer libdrm & kernel
[15:52] <tjaalton> so compile on another 32bit box
[15:52] <sits> right
[15:52] <tjaalton> assuming the eeepc is 32bit
[15:52] <sits> it is
[15:52] <sits> I'm actually cross compiling for what its worth
[15:54] <sits> is the ABI backwards compatible
[15:54] <sits> hmm perhaps that doesn't make sense
[15:54] <tjaalton> newer driver should compile against an older xserver
[15:54] <sits> ah ok
[15:55] <sits> wish me luck!
[15:56] <tjaalton> though it probably needs a newer libdrm, so you need to build that too
[15:56] <tjaalton> or build an older -intel with some abi-patches for newer xserver
[15:57] <sits> doh! I guess I'll see what happens. I think I might have install with a Xorg newer than 1.7 around
[15:57] <sits> tjaalton: thank you!
[15:57] <tjaalton> np
=== Cobalt is now known as EvilCobaltX
=== EvilCobaltX is now known as Cobalt