Mark Jacobs wrote:
This might not be the right forum for this question, but...
Doing some very limited in initial research I've found three
documented methods of performing Dynamic Allocation in COBOL (
Enterprise COBOL 4.2), BPXWDYN, CEEENV or setenv.
Q1) Are there any others?We already use a home grown assembler program
for dynamic allocation, but our direction is to move as much to the OS
as possible.
Q2) Is there any reason to pick one over the others?
I use BPXWDYN where I can. Unfortunately, with tape datasets all the
needed parameters a sometimes not there. So I have a alternative
process for those uses. COBOL can call the C runtime library
functions. C has dynalloc() and dynfree() to dynamically allocate and
free datasets. The linkage is a bit different. And C doesn't set the
high bit in it's parameter lists, so one has to call the function with a
extra parameter to keep the high bit in the address from being set like
this:
CALL 'DYNALLOC' USING BY REFERENCE LS-DYNT
BY VALUE ZERO
RETURNING WS-RETURN-CODE.
A gotcha is being C, character strings need to be NULL terminated (PIC
Z). And another is some of the functions in the C runtime have to be
statically linked at compile time. I use a small COBOL wrapper that
links the C routine statically which I then can call dynamically. The
runtime functions also allow text units to be passed which make them
flexible. A nice thing is since they are part of the IBM C runtime, you
don't have to maintain a routine or distribute one. It's already there
with Language Environment.
Alan
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
