Hmmm - it happens at random, and I've only seen it once on my UDB.
I got a core image out of cc1, and the stack trace is totally innocent,
and all the variables and arrays that are in the bottommost frame
and in the one up are fine. I also really had plenty of swap.

It's not clear to me what caused the SIGSEGV, but it really
doesn't appear to me to be cc1, nor does it appear to be a C library
issue (like I was wondering earlier).

It's possibly a hardware fault, but I guess I'll need to put in
some instrumentation to try and tell.

-matt