Simple COBOL program aborts when running under gcos simulator
Simulator Information
This was done on GHM so I don't have access to the simulator directly.
OS information not available
What Happened?
When running a gcos batch job that compiles a COBOL program that just outputs "Hello World", the execution aborts after the output. I believe that the simulator is at fault because looking at the dump output of the job, the execution somehow ended up in a data area and attempted to execute a zero opcode.
While the linker reports three missing routines, it inserts a MME GEBORT in place of those and this failure was not caused by a MME GEBORT call.
I can't completely eliminate an issue in Multics as the problem but it really seems most likely that this is a simulator issue.
Expected result
The program should have successfully completed.
...
Steps to reproduce
Note that this can only be reproduced on GHM right now because it requires a patched gcos module that is not available in a Multics Release yet. To get this patch, you must run the following command after logging in:
in >sysbuild>mcrs>MCR10130>bound_gcos_ -a -fc
The GCOS batch job can be created by pasting the following into a Multics session:
delete cob02.jcl
qedx
a
$ SNUMB COB02
$ IDENT ANDERSON
$ OPTION COBOL
$ COBOL CBL68,LSTOU
$ FILE C*
$ DATA S*
IDENTIFICATION DIVISION.
PROGRAM-ID. COB02.
ENVIRONMENT DIVISION.
CONFIGURATION SECTION.
OBJECT-COMPUTER. 6000 WITH EIS.
DATA DIVISION.
PROCEDURE DIVISION.
START.
DISPLAY "HELLO, WORLD."
END
$ EXECUTE DUMP
$ ENDJOB
\f
w cob02.jcl
q
The you can run the batch job with the command:
gcos cob02.jcl
The job output can be viewed with (note that it will also queue a print job with this output):
print cob02.jcl.sysprint.list
...