Richard E Maine wrote:
> On Sep 22, 2004, at 9:32 AM, Dan Nagle wrote:
>
>> Ordinary English words have their ordinary meanings
>> within the standard, unless ISO or the standard redefine them.
>
> ...
>
>> The standard defines the standard, implementations do not.
>
>
> Agree. But I know of no precise ordinary English meaning of
> "associated with the whole of an allocated target object." Nor
> is it clear to me that it is well-defined in the standard. My
> best guess at interpretation is that
>
> allocate(Q(10))
> P => Q(:)
>
> does not associate P with the whole of the allocated target object,
> but I could understand a different interpretation.
>
[snip]
Actually, the standard might be fairly precise. In F95 (the
one I can find on my desk), section 6.3.3.2 on deallocation
of pointer targets, says on 83:38
"If a pointer is currently associated with a portion
(2.4.3.1) of a target object that is independent of any
other portion of the target object, it shall not be
deallocated."
Doesn't that define what we mean by "whole object"? It's
the things that don't have any independent portions.
2.4.3.1 talks about portions of arrays, etc. and seems
to do the right thing.
Dick Hendrickson
|