Re: Passing unwanted parameters

Hi, again, Rory:

Correction -- for ILE bound *PGMs, the MIN is always set to 0 and the MAX is in fact always set to 255.
(Thanks to Simon Coulter for reminding me; I did a DSPPGM on an ILE bound *PGM to verify this.)

But why did IBM do this? :-o

<RANT>
Perhaps yet another example of "dumbing down" OS/400 or i5/OS to make it more like Unix, that "lowest common denominator" of all operating systems. :-(

What possible benefit can come from "silently ignoring" the error of passing more parameters than the called program expects? :-/

(Recall that in most Unix-like systems, many kinds of errors are just silently ignored; this seems to be a basic design philosophy of Unix.)
</RANT>

Ugh.

Mark S. Waterbury

> Rory Hewitt wrote:

All,

Probably a silly question, but I haven't managed to find an *absolute*answer:

If I call a program passing too many parameters, what will happen? I'm
assuming the called program won't throw a fit, but will OS/400 take note? I
know some IBM programs throw an error, but I assume (again!) that this is
program-specific code which checks for too many parameters.

If I define a prototype for the called program with the extra parameter
being CONST, could there be any memory overwriting problems? I don't think
so, but I'd like to know.

Although this is the RPG list, if anyone can confirm the behavior for other
languages as well, that would be nice :-)

Thanks,

Rory