(j3.2006) restricted pointers?

[Van Snyder]

On Thu, 2013-08-15 at 17:33 -0500, Bill Long wrote:
> Hi,
> 
> It has been pointed out that several compilers have command line 
> switches to tell the compiler that, in any given scope, there is no 
> aliasing of pointers.  Sort of like "restrict" in C.   Has some 
> attribute, along the lines of CONTIGUOUS, ever been proposed for this? 
> Note that I am NOT suggesting this as a F2015 feature - I'm just looking 
> for J3 history on the topic.

I have a recollection of something along these lines called a VIEW.  It
was more like an "automatic" equivalence than a pointer.  It was a
specification, not executable.  I thought it came and went around about
S/8-99 or so, but I can't find it.  I looked from S/8-98 until S/8-107.
It couldn't change type or kind, but could view a section of another
array.  Unlike equivalence, the shape of the view was automatic.

It explicitly limited aliasing to the view and viewed objects, unlike
TARGET, which simply says "watch out!"

There have also been proposals to allow a list of pointer names in the
TARGET attribute, and a list of target names in the POINTER attribute.

> Cheers,
> Bill
>