c133.org/blog : tech/fdo
Fri, 24 Feb 2006
.: for the record ::

user infoairlied: I'd be a lot less annoyed with companies like ATI if they shipped drivers (even proprietary ones!) that actually supported their latest cards.

The R5xx-series cards have been out for MONTHS now and there's still no acceleration, 2D or 3D, Open Source or proprietary. Good job, ATI!

Anyway. I'm excited about the way that things are headed, and I'm really glad that other people are getting excited, too.

To the people who were wondering - the "AI" in "AIGLX" stands for Accelerated Indirect - it's basically required to do a lot of neat things like using the local card to render OpenGL from a remote application. It also enables all sorts of kick-ass effects like the new bling in metacity and compiz (which doesn't work yet, but it's only a matter of time).

[07:02] | [] | # | G
Fri, 30 Dec 2005
.: #if 0 ::

There is a lot of really useless code sitting in Xorg today.

I was playing with a CVS checkout and noticed a large chunk of one of the source files was commented out with #if 0; so, this piqued my curiosity and I decided to figure out how many blocks of code are commented out with #if 0 in the server, driver, and lib trees.

The current count, in CVS HEAD, as of about five minutes ago:

moduleblocks of #if 0 code

All combined, that's 1112 blocks of code completely disabled at compile time just sitting there in the tree.

Haven't bothered checking any of the other modules; I'm sure there's more in apps but I don't really think it matters for docs or fonts. Still. That's a lot of useless crap.

[01:21] | [] | # | G
Tue, 05 Apr 2005
.: wtf xda ::

[@ajax] so working on DRI for i128, i turned off XAA
[@ajax] in the spirit of experimentation, i ran xcompmgr -a
[@ajax] it feels _much_ faster now
[@ajax] like, firefox doesn't tear when scrolling
[@ajax] it's kinda weird
[@daniels] hmm
[@daniels] maybe we should rename it to XDA :P
[@ajax] totally unaccelerated, but just fast enough to feel useful
[@ajax] the only unpleasant part is switching desktops

[22:59] | [] | # | G
Mon, 21 Mar 2005
.: spin spin ::

Donnie: Planet KDE, and every ten minutes. I'm thinking about changing it to update every fifteen, though. We're getting enough feeds that we're spending more time updating than idling.

[17:27] | [] | # | G
Sat, 12 Feb 2005
.: xdevconf version 2.0 ::

Zack Rusin flew in and Adam Jackson drove up last night and both are crashing at my place.

We're at the X Developers' Conference in Cambridge. ajax is taking notes, so I'll link to them once he uploads them.

Anyway. Back to the discussion about freedesktop and current level of desuckification.

[10:41] | [] | # | G
Tue, 07 Dec 2004
.: you've laid your brood of eggs in my navel! ::

Found out something mildly amusing today.

Apparently nobody has bothered building anything from the xlibs or xserver trees on an AMD64 system yet.

These modules contain such gems as:
# if defined(__AMD64__) || defined(AMD64)

Which doesn't actually *work* on these machines, as gcc defines __amd64__ and not AMD64. (Yeah, kids, case matters.)

Daniel added me back to the xlibs and xserver groups, so I should be able to commit some fixes for this stuff; however, there are still a lot of other issues on AMD64 that haven't been fixed yet. ::sigh::

[14:18] | [] | # | G
Sat, 05 Jun 2004
.: I kick ass ::

Tonight is rocking hardcore.

I spent about twenty minutes autotooling the 'nv' driver from the old monolithic X server tree, and convinced it to build under Daniel's Xizzle/Xorg tree.

After a bit more hacking around with the build system (to get it to properly generate the Makefiles in the nv dir, and to get the new NV driver module loaded into the libdrivers.a), it actually worked!

It loads the driver, and it sets the proper resolution, and everything!

Now the only thing missing is composite support. Which is a little out of my league, but I have faith that the guys at Sun working on it (yeah, I'm looking at you, Stu and Deron) will pull it off.

Damn, this has been a good night.

[22:21] | [] | # | G