| Richard Maine 2005-04-06, 12:21 pm |
| "Pat" <L_Pat_S@hotmail.com> writes:
> 4) I thought of a possible shortcut compared to 3, namely changing
> Cray-pointer statements of the form:
> pointer (pointer_variable, pointee)
> to
> equivalence (pointer_variable, pointee)
....
That seems unlikely to be of any use.
> longer being an actual pointer variable. Now it is quite possible
> that, for example, "pointer_variable" and/or "pointee" could be
> involved in the dynamic allocation of memory within the GraspVU code.
I'd say that "quite possible" is an understatement. More like
extremely likely. Equivalence won't help with that.
The pointer and pointee don't even need to take the same amount of
space.I'd say it was an odd coincidence if they happened to.
> In short, can anyone find
> something functionally wrong with replacing "pointer (" with
> "equivalence (" in Fortran code?)
All in all, I can't think of any context in which this would do
anything useful. It isn't a matter of finding what might be wrong
with it. I can't think of any way in which it would be right.
They just don't do anything even similar. Equivalence makes 2 or more
variables share the same storage. A Cray pointer makes one variable
contain the address of another.
--
Richard Maine
email: my last name at domain
domain: summertriangle dot net
|