RE: Problem with prototyping main procedure

[Joel Cochran <jrc@xxxxxxxxxx>]

On Fri, 2004-05-21 at 13:27, CWilt@xxxxxxxxxxxx wrote:
> Joel,
> 
> I know you said ignore this but, I don't think having the prototype in a
> copy member clarifies this answer.

Actually, for me it did... I've been lazy with external program calls
and always defined the prototype in the calling member (the exception
being API calls which I do prototype and store in a /copy).  To be fair,
though, I don't do this a lot.  If I find a program that I think will
get called a lot then that tends to look like a candidate for a *SRVPGM
instead.  If it was a *PGM that would frequently be called from other
*PGMs, then this makes perfect sense.

> Do you not use the extpgm keyword on the prototype for your mainline in
> conjunction with the PI for the mainline?

Nope, never have.

> Even when my prototype is in the same source member, I use the extpgm
> keyword.  It was my understanding from Barbara that the extpgm keyword was
> required in order to use the PI in place of an *ENTRY PLIST.
> 
> Does your work without it?

Yes, as long as the main name is the same as the *PGM.

Here is a sample from a program called NUMPARCELS, in source member
NUMPARCELS:

0001.00 d NUMPARCELS      pr
0002.00 d  localCode                     3
0003.00
0004.00 d NUMPARCELS      pi
0005.00 d  localCode                     3

No extpgm parameter required.

Joel
http://www.rpgnext.com