Output from the Separator Routine

A separator routine can write any kind of separation identification. IBM supplies a routine that constructs block characters. (See Using the Block Character Routine.) The separator routine can punch as many separator cards, or print as many separator pages, as necessary.

The output from the separator program must conform to the attributes of the output data set. To find out what these attributes are, examine the open output DCB pointed to by the parameter list. The attributes are:

Record format (fixed, variable, or undefined length)
Record length
Type of carriage control characters (machine, ANSI, or none)

For printer-destined output, a separator routine can begin its separator records on the same page as the previous job output, or skip to any subsequent page. However, the separator routine should skip at least one line before writing any records because in some cases the printer is still positioned on the line last printed.

After completing the output of the separation records, the separator routine should write sufficient blank records to force out the last separation record. This also allows the error exit routine to obtain control if an uncorrectable output error occurs while writing the last record. One blank record, for printer-destined output, and three blank records, for punch-destined output, are sufficient to force out the last record.