possible mmap() problems

[Joe Buehler]

Christopher Faylor wrote:

> Well, you shouldn't really get *any* thread terminations.

I added some debugging.  The thread terminations are terminations
of the "read_pipe" thread during cygthread::detach().  I presume
there is a boundary case in detach() that I will need to figure out.

Curiously, sometimes the name of the thread (__name) is a null pointer
at the time of terminate_thread().

I'll keep looking at it.  I'm trying to track down disappearing virtual
memory crashing ksh93 at the moment.
-- 
Joe Buehler