(j3.2006) Comment 2 from TS Ballot

[Bill Long]

On 9/18/12 4:29 AM, Malcolm Cohen wrote:

>> "..., a pointer that is computed from information in that C descriptor..."
>
> Sorry, but this is both vague and incorrect!
>
> The information in that C descriptor includes lots of stuff that is perfectly ok
> to use when computing pointers inside other objects, like the element size and
> the extents.  Indeed, this is likely to happen for procedures that make copies
> of array arguments for example.

OK, that is a problem.

>

> The point is, a pointer value computed according to the C rules can have a lot
> of things going into it, including multiple pointers, but it can only be
> "derived from" or "based on" a single pointer (any extra ones must be subtracted
> out again along the way).
>

OK.

>
> I understand that you would prefer more specific wording, but getting more
> specific makes it makes it much more likely to be wrong.  Let not the perfect be
> the enemy of the good: vague but correct is better than precise and wrogn.
>

So,

"..., a pointer derived from the base_addr in that C descriptor..."

or

"..., a pointer based on the base_addr in that C descriptor..."


making (with the first choice) the whole sentence

"If a formal parameter that is a pointer to CFI_cdesc_t corresponds to
a dummy argument in an interoperable procedure interface, a pointer
derived from the base_addr in that C descriptor shall not be used to
access memory that is not part of the object described by the C
descriptor."

Any strong preference for "derived from" versus "based on"?

Cheers,
Bill




> Cheers,
>

-- 
Bill Long                                           longb at cray.com
Fortran Technical Support    &                 voice: 651-605-9024
Bioinformatics Software Development            fax:   651-605-9142
Cray Inc./Cray Plaza, Suite 210/380 Jackson St./St. Paul, MN 55101