it was like 40° today so i went running. i hadn't gone since around thanksgiving, and i really felt it. hopefully we'll get another cold front and i won't have to run again. ever.
with my success in os x addressbook hacking, i thought i'd write a backend for evolution to use it. it's probably pretty easy for all i know, since i couldn't get evo HEAD to be able to link itself.
that took a lot of time, since i am stubborn, and don't like being defeated by software.
i managed to catch the bruins game. they lost, but thornton got an amazing 1-on-3 goal. hopefully, someone else on the team will figure out how to score, and they might be able to make the playoffs.
after that, joe and i played frankenstein with some computers. my k6 300 (formerly my main development box in college, then a brief stint as bug-buddy.org, and recently the router at our apartment) is now sitting in an ATX case. it's old AT case now houses our "new" router: joe's k6 200 (but with the old router's hard drive). these routers have a tendency to kernel panic when we first install them, then be fine. we'll see.
i was going to lustrefy my k6 box (this is how much i love my friend phil), however that's not going so well. first, i tried the "chaos" kernel rpm. it turns out to be for P4s (which my K6 is not). then, i tried just about every possible patch they had except for the right one (well i tried 2). and since building a kernel is too easy, i of course have to cross compile it, which requires some "tweaks" to the Makefile to get things working on non-gnu/non-linux machines.
my guess is that i'll never get this working.
ok, ignoring the fact that i'm cross compiling (remember, i said ignoring), what's up with this:
handler.c: In function `mds_getattr_name': handler.c:683: `IT_SETATTR' undeclared (first use in this function) handler.c:683: (Each undeclared identifier is reported only once handler.c:683: for each function it appears in.)
this is with lustre 0.5.20.1, portals 0.5.20.1, linux 2.4.18-18.8.0-l12.