RE: [Cobolforgcc-devel] OCCURS (was Redefines

[William M. K. <wm...@ix...>]

I think that (eventually) you will want to emulate something like MERANT's
ODOSLIDE directive - that allows for *either* fill-up *or* slide-in
behavior.  IBM works like the latter - so MUCH of the code that you get
probably will "rely" on it.

> -----Original Message-----
> From: cob...@li...
> [mailto:cob...@li...]On Behalf Of Tim
> Josling
> Sent: Tuesday, January 02, 2001 2:35 AM
> To: cob...@li...
> Subject: Re: [Cobolforgcc-devel] OCCURS (was Redefines
>
>
> Bill,
>
> First up I am implementing the standard with needed enhancements
> to allow calling C routines. I had my fingers well and truly
> burned implementing an earlier extension. The earlier extension
> was the use of fully nested COPY with REPLACING - a can of worms.
> Once you go beyond the standard it raises a lot fo indefined
> issues.
>
> In the OCCURS example, the issues are what happens to the holes
> in the occurs. Do you fill them up or leave them vacant? So I
> will leave this alone for now.
>
> Later on, we need to have a look at what the existing compilers
> have done. What do you think is the best approach,
>
> Tim Josling
>
> "William M. Klein" wrote:
> >
> > Tim,
> >  For OCCURS are you going to stick (at least initially) to the "strict"
> > ANSI/ISO rules - which require that an ODO (Occurs Depending
> On) be the LAST
> > item in a record - and that they not be nested (or have other
> OCCURS nested
> > in them)?
> >
> > IBM, MERANT, and most "similar" compilers have extensions to
> "remove these
> > restrictions" - but what they do about the run-time results can differ
> > DRAMATICALLY (compare MERANT's ODOSLIDE directive).
> >
> > > -----Original Message-----
> > > From: cob...@li...
> > > [mailto:cob...@li...]On Behalf Of Tim
> > > Josling
> > > Sent: Monday, January 01, 2001 2:12 PM
> > > To: cobolforgcc-devel
> > > Subject: [Cobolforgcc-devel] Redefines
> > >
> > >
> > > I have uploaded the code and tests for redefines. Apart from
> > > trivia like filler, the main hard thing left in the data division
> > > is occurs.
> > >
> > > The debugging information is not all coming out. This needs a bit
> > > more work to find out which magic fields I need to fill in.
> > >
> > > Ted Seward is working on sort merge. The basic routines are
> > > mainly done, with only the basic merge to go - actually the most
> > > interesting algorithm.
> > >
> > >
> > >
> > > Tim Josling
> > >
> > > _______________________________________________
> > > Cobolforgcc-devel mailing list
> > > Cob...@li...
> > > http://lists.sourceforge.net/mailman/listinfo/cobolforgcc-devel
> >
> > _______________________________________________
> > Cobolforgcc-devel mailing list
> > Cob...@li...
> > http://lists.sourceforge.net/mailman/listinfo/cobolforgcc-devel
>
> _______________________________________________
> Cobolforgcc-devel mailing list
> Cob...@li...
> http://lists.sourceforge.net/mailman/listinfo/cobolforgcc-devel