Question on variable_size and function-at-a-time

[Geoff Keating]

> X-Original-To: geoffk@foam.wonderslug.com
> Date: Wed, 14 Apr 2004 22:27:08 +0200
> From: Toon Moene <toon@moene.indiv.nluug.nl>
> Organization: Moene Computational Physics, Maartensdijk, NL
> X-Accept-Language: en
> Cc: Geoff Keating <geoffk@geoffk.org>, Richard Henderson <rth@redhat.com>,
>         Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, gcc@gcc.gnu.org
> X-OriginalArrivalTime: 14 Apr 2004 20:22:51.0968 (UTC) FILETIME=[4301D400:01C4225E]
> 
> Robert Dewar wrote:
> 
> > Geoff Keating wrote:
> > 
> >> I believe it's wrong that we use SAVE_EXPR for array sizes for C.  We
> >> should instead create temporary variables to hold the size
> >> declarations, and set them in the appropriate point in the instruction
> >> stream.
> > 
> > I agree. In fact my feeling is that we should get rid of the use
> > of SAVE_EXPR wherever possible.
> 
> That's OK with me, as long as we have a way of expressing that in
> 
>        SUBROUTINE SUB(A, N)
>        DIMENSION A(N)
>        N = N + 1
>        DO I = 1, N - 1
>           PRINT*, A(I)
>        END DO
>        PRINT*, N
>        END
> 
> still works correctly and both
> 
> 1. Prints A(1 .. original N)
> 2. Prints original N + 1 at the end.

Sure, like this (in pseudo-F77):

        SUBROUTINE SUB(A, N)
	ASIZE = N
        DIMENSION A(ASIZE)
        N = N + 1
        DO I = 1, N - 1
           PRINT*, A(I)
        END DO
        PRINT*, N
        END

-- 
- Geoffrey Keating <geoffk@geoffk.org>