(j3.2006) finalizer infelicity
Wed Nov 21 15:59:36 EST 2007
Aleksandar Donev wrote:
> On Wednesday 21 November 2007 11:11, Michael Ingrassia wrote:
>> do our semantics really require us to call a
>> finalizer ([59:5] "it is called") when that
>> finalizer could not have been called explicitly ? Do they forbid us
>> to bind such finalizers?
> I agree with Michael that this seems ill-defined. Maybe this again points out
> to the conclusion that Jim's objections also raised: The dummies should be
> required to be assumed-shape arrays. Or better, the whole non-elemental
Assumed-size would be a better choice in this case. Seems inefficient
to create a dope vector just to call a final routine.
But this seems just a special case of the more general problem of
calling a subroutine with actual arguments that have a size that varies
from one call to the next. If that's what you anticipate happening, it
is reasonable to declare the dummy argument as assumed-size. If you
declare the dummy as explicit-shape, expect a "not-so-good-programmer"
hat and a chair in the corner facing the wall.
> finalizer business should be deleted. I am not even kidding :-\
> J3 mailing list
> J3 at j3-fortran.org
Bill Long longb at cray.com
Fortran Technical Support & voice: 651-605-9024
Bioinformatics Software Development fax: 651-605-9142
Cray Inc., 1340 Mendota Heights Rd., Mendota Heights, MN, 55120
More information about the J3