RE: MCH3602 - Pointer not valid

[Scott Klement <midrange-l@xxxxxxxxxxxxxxxx>]

On Mon, 5 Jun 2006, Elvis Budimlic wrote:
> OPM compiled programs maintain a count of parameters and if you pass extra
> in you'll get exceptions.  ILE compiled programs will ignore extra parms.

This is true of CL programs, but not of RPG programs. OPM RPG programs 
ignore extra parms, just like the ILE ones do.


On Mon, 5 Jun 2006, Michael Naughton wrote:
> The thing is, the first program (the once called from the Domino
> application) runs fine -- it accepts the parm, removes the null terminator
> and pads it with spaces, and then calls a CL program passing the cleaned
> version as a parm. The CL program then bombs the first time it tries to
> use that parm.

Is it possible that the program that removes the null terminator and pads 
it with spaces is using teraspace memory to do that?   If so, that'd 
explain why you get the MCH3602, and why it works with ILE programs (which 
understand how to work with teraspace) but doesn't work with OPM programs 
(which don't understand teraspace).

I don't see why a V5R2 -> V5R4 upgrade would cause this problem, unless 
you replaced the "first program" (the one that removes the null and pads 
with blanks) when you upgraded it.

The only other possibility that comes to mind is that one of the programs 
(possibly the one removing the null and adding spaces) is overwriting 
memory that doesn't belong to it.  It's possible that in V5R2, you got 
"lucky" and overwrote memory that wasn't in use, but that a change in V5R4 
caused the layout in memory to change so that you're now overwriting part 
of a pointer variable.