CREAD buggy?

[Andrew Pinski]

> 
> Steve,
>    Thanks. Looking through the xplor-nih heap routines, it appears
> that they use first element of the heap array for the absolute
> address of a heap allocation and the second element as the
> length of that particular heap allocation.
>              Jack
> ps I'll try the  -fbounds-check and -fstrict-aliasing flags next.

I think he means -fno-strict-aliasing.  But you might as well test
-fargument-alias as it looks like you are passing the same variable
to the two different agruments and fortran standards says they don't
alias.

-- Pinski