Does SET ADDRESS OF X TO Y work?
Estimated Reading Time: 2 Minutes
Question:
I have some code that uses M$ALLOC and then assigns the address of an array (in the linkage area) to the space just allocated. Code fragment below. There is a note in the migration section that “SET X TO ADDRESS OF Y” requires compiling with the -cp option to enable full POINTER support. But I am using only "SET ADDRESS OF X TO Y" and am not calling C functions.
000419 CALL "M$ALLOC" USING IMG-ROW-SIZE, 000420 IMG-ROW-PTR, 000421 SET ADDRESS OF IMG-ROW TO IMG-ROW-PTR,Where IMG-ROW is an array in the linkage section.
Answer:
The answer is yes. SET ADDRESS OF X TO Y is supported. Use the -ca compiler option.
Note: If you plan to exchange pointers with C functions then enable full POINTER support by compiling with the -cp option.
If you are using a USAGE POINTER item internally, within COBOL (i.e. not to pass to C functions), SET X TO HANDLE OF Y works as desired and gives the same behavior as SET X TO ADDRESS OF Y in ACUCOBOL.
In addition, "SET ADDRESS OF X TO Y" works as expected.
Because of Java constraints, with isCOBOL, M$ALLOC returns a handle, not an actual memory address. When you set the address of a linkage item to the handle value, the program behaves as desired because internally Java uses handles to identify objects, not pointers.
If you compile with -ca, then the program you described will compile and run as it did with ACUCOBOL. If you don't compile with -ca then you simply need to change USAGE POINTER to USAGE HANDLE wherever it occurs in the data division.
Here is an example program:
Message 1
Message 2
Message 3
Message 4
Message 5
Message 6
If you are using a USAGE POINTER item internally, within COBOL (i.e. not to pass to C functions), SET X TO HANDLE OF Y works as desired and gives the same behavior as SET X TO ADDRESS OF Y in ACUCOBOL.
In addition, "SET ADDRESS OF X TO Y" works as expected.
Because of Java constraints, with isCOBOL, M$ALLOC returns a handle, not an actual memory address. When you set the address of a linkage item to the handle value, the program behaves as desired because internally Java uses handles to identify objects, not pointers.
If you compile with -ca, then the program you described will compile and run as it did with ACUCOBOL. If you don't compile with -ca then you simply need to change USAGE POINTER to USAGE HANDLE wherever it occurs in the data division.
Here is an example program:
id division.
program-id. malloctest.
data division.
working-storage section.
77 img-row-ws pic x(50).
77 img-row-ptr usage handle.
linkage section.
01 img-row.
03 filler pic x occurs 50.
procedure division.
main-logic.
set address of img-row to address of img-row-ws.
move "Message 1" to img-row.
display img-row.
move "Message 2" to img-row.
display img-row-ws.
move "Message 3" to img-row-ws.
display img-row.
move "Message 5" to img-row-ws.
CALL "M$ALLOC" USING length of img-row,
IMG-ROW-PTR,
SET ADDRESS OF IMG-ROW TO IMG-ROW-PTR,
move "Message 4" to img-row.
display img-row.
display img-row-ws.
move "Message 6" to img-row-ws.
move "Error" to img-row.
display img-row-ws.
The output of program is: Message 1
Message 2
Message 3
Message 4
Message 5
Message 6