Finger info for

 Work officially begins. Time to get this building under gcc again. Since
 we changed up the source tree layout, the Makefiles from UE2 are basically
 useless. Rather than rewrite them, I'm going to spend some time exploring
 SCons, which came highly recommended by TTimo, the Doom3/linux guy. First
 Unreal steals their colored lighting, then their build system! :)
 Obviously, there's a lot to be done at this point, but best to start now
 so I'm not scrambling to port a whole engine when UE3 games get closer to
 shipping. Updates as I have them.

Spider-Man 2:
 Gone gold, baby.

 The installers on MacSoft's Neverwinter Nights expansion pack discs are
 actually a modified version of MojoPatch. Go order your copies!

Shrek 2:
 Now shipping!

 For those that weren't at WWDC, Apple gave out preview discs of MacOS X Tiger.
 One of the things Tiger installs by default? OpenAL.framework. No kidding.
 That's basically awesome. It's Apple's version, which is open-sourced and
 residing in's CVS repository.

 Likely I'll move my implementation over to Linux and stop further Mac
 development, so that there is a clear technology path on the Mac. I'll
 devote further Mac development and debugging to the Apple implementation.
 After all, it was partially a stop-gap solution (remember when ut2003 took
 25% of the CPU mixing audio? It was a needed fix, no doubt!), and partially
 a technology proof-of-concept to show Apple what works well in terms of game
 development. No doubt it has served me well.

 In the short term, I'll have to decide what we ship on the disc with
 Unreal-based games. For ut2004, my implementation is the only one with
 ALC_EXT_capture support for VoIP...this could be added to the other
 implementation, but hasn't been as of yet. My version is apparently a
 little faster, but it's stereo only (but the subversion repository doesn't
 crash on M-Audio 5.1 and 7.1 cards anymore), so Apple's tech is probably
 more attractive for further development by default.

 All of the missing functionality in Apple's implementation could be fixed
 with some elbow grease, which I'm sure will show up one way or another in
 the near future. Overall, this is a very good step forward, and I applaud
 Apple for giving game developers something they really need.

 Latest CVS builds and runs on Solaris/x86 (and presumably Solaris/sparc, too).
 I get a lot of questions about MacOS X: the game _does_ run on OSX, there
 just isn't a nice installer or anything at the moment, so you have to
 compile it yourself. When there's time, I'll put together a shareware-based
 installer, and, if I can find a copy, one that works with the Mac retail disc.

 (Yes, this is still being worked on.)

Unreal Tournament 2003:
 There's an exploit in the ut2003 network code, so here's a new build.


 The Linux one has about a million changes over the stock 2225, since it's got
 all the MacOSX work on top of it. Consider it beta. The Mac version has one
 or two fixes, so it's worth updating.

Unreal Tournament 2004:
 If Mac retail installer crashes on you, use this:

 Linux (x86 and amd64) official 3236 patch (new build with load times fixed):

 MacOS X (un?)official 3236 patch (YES, this is newer than 3229):

Call of Duty:
 1.4 is out, now with PunkBuster support:

 This is a 1.4 server with an exploit closed. Admins should all upgrade:

Postal 2 Share the Pain:
 Linux demo:
 Linux retail: In beta testing (apply at
 Mac retail: In beta testing (no more applications, please!)

America's Army:
 2.1.0 is out for Linux and Mac:

Other stuff:
 So I sat down and cleaned up my callstack-processing programming challenge.
 Overall, the code turned out to be sane, but there were some off-by-one
 errors in what I posted here.

 That being said, it turned out to be a huge win, much more than I
 anticipated for my purpose...since my eventual goal was to track callstacks
 where memory was allocated and freed, you not only find the usual frequency
 of repeated callstacks, but that the differences are even fewer since many
 programs only do memory management in a few choice places.

 Case in point: a run of GCC to compile some C code...of all stack frames
 (not just individual callstacks), less than one percent of them were
 unique...0.8%, specifically, so the storage tradeoff obviously paid off,
 especially when you're looking at a data set of about 250,000 callstacks
 from compiling a small C program.


When this .plan was written: 2004-08-21 18:04:42
.plan archives for this user are here (RSS here).
Powered by IcculusFinger v2.1.27
Stick it in the camel and go.