"MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx> wrote on 08/05/2016
12:31:48 PM:
This has to be a command line thing. I wonder if I need to create the
whole
command in a variable and then use QCMDEXEC to execute it. Any ideas?
Normally, there isn't an issue between command use from a command
line vs. command use from a CL program. As long as the command can be
prompted without errors in either environment then the editor will format
it correctly in the CL program source for compilation and subsequent
execution. The issue comes when calling a CL program without using a
command interface.
The reason is that the command-line processor (in either
environment) has to make some assumptions about literals used as
parameters in the CALL parameter list. Those assumptions are (1) that
string literals will be blank-padded to a minimum length of 32 characters
or a maximum length equal to the actual length of the string literal, (2)
that numeric literals are always formatted as *DEC 15 5, and (3)
floating-point literals must be formatted using floating-point syntax. The
processor for a command interface doesn't have that issue because the
command contains all the data type definition information required to
correctly format the interface to the command processing program.
So, that said, I presume you can prompt the "failing" command in
your CL source and the editor reformats the results as needed and without
errors? If so, then it is the command processing program associated with
the command interface, itself, that is failing to handle the parameters
appropriately in an *IPGM or *BPGM environment.
Sincerely,
Dave Clark
midrange.com
