CANCEL (85) and multiple levels of called programs

Hi,

Could someone confirm my understanding of CANCEL (85 version). As I understa
nd
it, if I cancel a program that I have called, it should reset the working
storage to initial values ready for the next time I call it again. And from 
the
85 version onwards, it should also cancel any programs were called by the
program that I called. Right?

I'm on Windows with Micros Focus Net Express 3.1, and have something like th
is:

Prog1 (exe) --CALL--> Prog2 (dll) --CALL-> Prog3 (dll)
Prog1 then cancels Prog2.
Prog1 (exe) --CALL--> Prog2 (dll) --CALL-> Prog3 (dll)
Prog2 working storage has been reset.
Prog3 working storage has *not* been reset.

This would seem to contradict my understanding. So am I wrong, or is MF doin
g
something wrong?

TIA,
Richard

"Richard Antony Burton" <richardaburton-NOSPAM-@hotmail.com> wrote:

>Hi,
>
>Could someone confirm my understanding of CANCEL (85 version). As I underst
and
>it, if I cancel a program that I have called, it should reset the working
>storage to initial values ready for the next time I call it again. And from
 the
>85 version onwards, it should also cancel any programs were called by the
>program that I called. Right?
>
>I'm on Windows with Micros Focus Net Express 3.1, and have something like t
his:
>
>Prog1 (exe) --CALL--> Prog2 (dll) --CALL-> Prog3 (dll)
>Prog1 then cancels Prog2.
>Prog1 (exe) --CALL--> Prog2 (dll) --CALL-> Prog3 (dll)
>Prog2 working storage has been reset.
>Prog3 working storage has *not* been reset.
>
>This would seem to contradict my understanding. So am I wrong, or is MF doi
ng
>something wrong?

MF is doing it right. Prog3 must be cancelled explicitly by prog2, its logic
al
parent, or by prog1, which would be poor structure IMO.

The Standard says a cancel will automatically cancel programs INCLUDED in pr
og2,
not those called by prog2. Included means nested compilations i.e. in the so
urce
file of prog2.

If you want initial values always, an alternative to cancel is INITIAL in
program-id.

"Robert Wagner" <robert.deletethis@wagner.net> wrote in message
news:40d0bdd5.22483499@news.optonline.net...

> MF is doing it right. Prog3 must be cancelled explicitly by prog2, its
logical
> parent, or by prog1, which would be poor structure IMO.

Thanks for clearing that up.

Richard.