Re: gnubol: subsets

[RKRayhawk@aol.com]

In a piece of one of my previous posts that was quite busy I mention a notion 
such as ...

<<  we will want to see
 a procedure reference as a statement that types a particular
procedure (you know like a SORT INPUT/OUTPUT procedure); which
is not declared in the section or paragraph itself.  >>

This kind of typing is not urgent, but designing it might be. Later 
optimization features might be much more reluctant to optimize aspects of 
procedures refered to with SORT or declaratives, because of the intimacy with 
external control points, such as access methods. This is really just a 
sweeper type of function this kind of typing. A section that is not 
referenced at all, and can be determined to not be enterable by fallthru is 
ripe for optimization right out of here.  Other procedures, _might_ have 
entrances and exits simplified under certain optimization insistances from 
the user. But we may have some procedures where we do not want that because 
of how they are referenced.  

So I am arguing in favor of an attribute of some kind in the procedure SYMT 
that accumulates toggles that record how the procedure is referenced. All of 
which is a big tangent. It is just that that is easier to do if we see it 
comming as a requirement, assuming eventual agreement to it as a requrement.

Best Wishes
Bob Rayhawk


--
This message was sent through the gnu-cobol mailing list.  To remove yourself
from this mailing list, send a message to majordomo@lusars.net with the
words "unsubscribe gnu-cobol" in the message body.  For more information on
the GNU COBOL project, send mail to gnu-cobol-owner@lusars.net.