More Tru64 errors

Sun Dec 9 13:55:00 GMT 2001

Jeff Sturm wrote:
>Or the collector didn't build for threads at all, in which case it has
>zero chance of working with a threaded libgcj. (The "limit" address shown
>in the gdb stack trace belongs to the primordial thread stack, not the
>current thread.)
>>Some time ago I'd suggested the patch below for 3.0. It would help
>prevent these confusing bug reports when a threaded gc isn't available.
>
Unfortunatly it would cause the build to break on such targets, which 
for 3.0 is arguably worse that building a broken java library.
Maybe it should just define GC_LINUX_THREADS (which should really be 
renamed GC_POSIX_THREADS) and hope that it works?
regards
Bryce.