xplor-nih segfaults

[Jack Howarth]

      Does anyone know if the -fstack-protector code works in 4.1 branch
for gfortran? I can build the cp_test.f90 with the -fstack-protector
flag with 20050728 gcc 4.1 cvs and the resulting binary seems to run
correctly but I have no idea if this means the stack protector code is
functional. Does any one have a quick fortran example that should
corrupt the stack that I can test with?
       My idea is that if gcc 4.1's gfortran has a functional stack
protector on Darwin 8 (Tiger), I can build xplor-nih with the
-fstack-protector flag for the fortran code compiled with 
gfortran and hopefully I should see a runtime error the moment
the stack is about to be corrupted. Makes sense, no? However I think
I might need to compile libgfortran itself with -fstack-protector
to catch any flaws in its io routines that might corrupt the stack.
Not sure how that should be done properly though.
          Jack