(j3.2006) restricted pointers?

[Dick Hendrickson]

My recollection is that Jon Steidel and Larry Rolison proposed something
along this line when they were both involved with J3 from Cray.  Certainly
before the turn of the century (but in Las Vegas).  I think the proposal
was something along the lines of adding an optional list to a pointer
declaration that listed the things that this particular pointer would/could
point to.  Sort of like the optional hint list for the assigned goto.  My
recollection is that the proposal failed because compilers would still have
to deal with unrestricted pointers and  that most "important" codes/users
weren't currently having problems with poor optimization strictly because
of pointer ambiguities.  Sort of the "Dr., my code runs slow when I do
this.  Hmmm, don't do that."  solution.  I may misremember.

Dick Hendrickson


On Thu, Aug 15, 2013 at 7:47 PM, Dan Nagle <dannagle at verizon.net> wrote:

> Hi,
>
> On Aug 15, 2013, at 16:49 , Van Snyder <Van.Snyder at jpl.nasa.gov> wrote:
> >
> >
> > 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.
>
>
> Van might be referring to a memory pool idea I had for f08.
>
> The program would name pools, and a pointer or target could be
> associated with a pool.  Inter-pool aliasing could not exist.
>
> It was seen as not enough of a win to be worth the effort.
>
> --
> Cheers!
>
> Dan Nagle
>
>
>
>
> _______________________________________________
> J3 mailing list
> J3 at mailman.j3-fortran.org
> http://mailman.j3-fortran.org/mailman/listinfo/j3
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.j3-fortran.org/pipermail/j3/attachments/20130815/3672b756/attachment.html