bootstrap (sic) failure on sparc-sun-solaris2.8

[lucier@math.purdue.edu]

Kaveh has already reported this in

http://gcc.gnu.org/ml/gcc-bugs/2001-11/msg00994.html

but it's causing a full bootstrap failure on sparc, which I think should
elevate it above the run-of-the mill bugs in gcc-bugs.  And it's
preventing me from seeing whether the recent register changes fix the
64-bit problems I've always had with the -m64 option, even when people
claim that this option "works" in 3.1.

Brad