Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!munnari!murdu!u3369429
From: u3369429@murdu.OZ (Michael Bednarek)
Newsgroups: comp.os.vms
Subject: Re:  Language question
Message-ID: <1256@murdu.OZ>
Date: Sun, 24-May-87 23:35:32 EDT
Article-I.D.: murdu.1256
Posted: Sun May 24 23:35:32 1987
Date-Received: Tue, 26-May-87 01:30:42 EDT
References: <8705221905.AA01082@sdag.cs.umd.edu>
Reply-To: u3369429@murdu.UUCP (Michael Bednarek)
Followup-To: comp.os.vms
Distribution: world
Organization: I.A.E.S.R., Melbourne University
Lines: 34

In article <8705221905.AA01082@sdag.cs.umd.edu> ogud@SDAG.CS.UMD.EDU (Olafur Gudmundsson) writes:
>The simple solution to your problem is to use equivalence statements
>this way there is no overhead because of copying.
>The only overhead is that you will have to use IF statments to pick the
>right data to work with
>
>Example :
>	subroutine foo( arr, typ)
>
>	int arr( *)   , typ 
>	real r4( 1)
>	integer*2 i2(1)
>	byte	c1(1)
>	eqivalence (arr(1), r4(1) , i2(1) ,  c1(1))
>
>c    and in the subroutine you do
>	if( typ.eq. INT) then
>
>	else if( typ .eq. REA4) then
>
>	...
>
>	endif

This is not only simple, it also doesn't work.
I suspected that immediately, but I compiled the fragment above ...
>Spelling mistakes make life more colorful
[This is not funny]
and got -F-MULDECNAM Multiple declaration of name, A name appears in two or more
inconsistent declaration statements.

Of course, the same approach within a main program will work.

Michael Bednarek (u3369429@murdu.oz  or  U3369429@xvax.dn.mu.oz.au)