This brings up a related question: Some of the "propagation of INTENT"
issues are addressed in a special way for PURE procedures. (I have
already gotten warned about some of these issues while trying to write
some code for a compiler that accepts the PURE procedure attribute.)
=
So: What would be the interaction of READONLY rules with those for
PURE ??
Loren P. Meissner
[log in to unmask]
> -----Original Message-----
> From: [log in to unmask]
> [mailto:[log in to unmask]]On Behalf Of Van Snyder
> Sent: Wednesday, September 02, 1998 10:59 AM
> To: comp-fortran-90
> Subject: READONLY
>
>
> Dick Hendrickson and Richard Maine have made some interesting points
> concerning a READONLY attribute.
>
> . . .
>
> Richard Maine touched on the other problem, which is a bit stickier:
> I couldn't find a constraint (this is essentially a compile-time
> restriction) that if an INTENT(IN) dummy argument is used as an actual
> argument it must be associated to an INTENT(IN) dummy argument or be
> part of an expression (and therefore in theory not assignable
> -- is (X) used as an actual argument guaranteed not to allow X to be
> changed?). It would be easy to require that a READONLY actual argument
> could only > be associated to an INTENT(IN) dummy argument, but the
> induction doesn't appear to hold.
>
> It's unlikely that these issues could be ironed out during the one
> remaining J3 meeting allowed for introducing new technical content.
>
> It looks like READONLY is going to miss the train, again.
>
> Best regards,
> Van Snyder
>
>
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|