(j3.2006) (SC22WG5.5057) [ukfortran] WG5 vote on draft TS on further coarray features

[Malcolm Cohen]

Nick Maclaren writes:
>Passim.  The specification is messy and restrictive, and should be
>changed.  For example, it is not possible to reduce INTENT(IN) examples.

I agree.

>A far better solution would be either two procedures or a generic
>interface, of the following forms (CO_REDUCE is used as an example):
>
>    CO_REDUCE (SOURCE, OPERATOR [, STAT, ERRMSG])
>    ! SOURCE would be INTENT(INOUT)
>
>    CO_REDUCE (SOURCE, OPERATOR, RESULT [, RESULT_IMAGE, STAT, ERRMSG])
>    ! SOURCE would be INTENT(IN)

Unfortunately this particular scheme does not work when SOURCE is a scalar 
integer, because
   CO_REDUCE(integer,procedure,integer)
is ambiguous as to whether it is the first form or the second.

It would work if they had otherwise-gratuitously-different argument orderings, 
e.g.
    CO_REDUCE (SOURCE, OPERATOR [, STAT, ERRMSG])
    ! SOURCE would be INTENT(INOUT)

    CO_REDUCE (OPERATOR, SOURCE, RESULT [, RESULT_IMAGE, STAT, ERRMSG])
    ! SOURCE would be INTENT(IN)
(Swapping OPERATOR and RESULT also works, but this keeps RESULT and RESULT_IMAGE 
together.)

The same is true of CO_SUM, but there is no OPERATOR to swap so that's a bit 
harder to tweak.

It's probably clearer to have CO_SUM and CO_SUM_SELF.  Or something.

Cheers,
-- 
................................Malcolm Cohen, Nihon NAG, Tokyo.