Re: Need Some RPG Technical Questions

Hi Jerry,

There has long been an option (NOMAIN) in the Header specs to strip out
all of the cycle clutter. Especially useful for service programs; I've
never tried to use that option for a "regular" program. Hmmm...wonder if
that would work if one did one's own OPEN's etc.?

The reason it won't work for a "regular" program is that there's no main procedure. When you run the CALL command to call a program, it tries to run the UEP (User Entry Procedure) which in RPG is the program's main procedure. (By "main procedure" I mean the "normal mainline part" of your program -- i don't mean a procedure named 'main')

So NOMAIN won't work for a regular program.

in V6R1, IBM is providing a new H-spec keyword called MAIN. With this keyword, your module will be pretty much identical to a NOMAIN module -- it'll have no RPG cycle, *INLR won't do anything, etc... The only difference between MAIN and NOMAIN is that you'll have a main procedure -- which is to say, a procedure that gets called when you invoke it as a *PGM object. Also, you'll be able to designate WHICH procedure this is.

Personally, I find myself thinking of it as a NOMAIN module where I can designate a subprocedure that's to be invoked when called as a program.

For example:

H MAIN(ScottProc)

D ScottProc PR ExtPgm('SCOTTPGM')
D parm1 10a
D parm2 5a

P ScottProc B
D ScottProc PI
D parm1 10a
D parm2 5a
//
// This is the procedure called when the program starts
//
P E

Hopefully I have the syntax correct -- I don't actually have access to V6R1 at this point, but this is the way I understand it to work.

If you code neither a MAIN or NOMAIN H-spec keyword, you'll get the original behavior, including the RPG cycle, et al. So don't worry there, you're not losing anything... And we'll still be able to have the same holy wars about the cycle.