I am attempting to set up an automated procedure for running a PREPARE command
and checking the output. The server runs under MVS, and I am issuing the
command using the MVS administrative client in a background TSO session. I use
the -outfile option to direct the output to a file, and the wait=yes option to
get the command to run in the foreground. The output consists mostly of ANR
messages, many of which are broken into two lines. This happens even though
the record length of the file is amply sufficient to fit the entire message
into a single record. I have not found any simple and unambiguous test that
would allow a parsing program to recognize split messages and reassemble them.

Is there a way to prevent the messages from being split across lines in the
first place? Failing that, is there a simple and reliable way of identifying
continuation lines?

As a general matter, I have found that ADSM makes automated operations
incredibly difficult. Most IBM products that run under MVS generate return
codes which provide an easy way of determining success or failure. Why is ADSM
such a glaring exception to this general rule?