Fellow Listers,

I have a CM utility that is used to 'scan' file sets for user specified
strings.  I need to enhance the utility to handle bytestream files.  The
utility uses 'MR;NOBUF' access to the files and uses the record size, for
fixed record files, or the variable block structure, for variable record
files, to 'step' through the large block of data while searching for the
specified string.

The utility aborts when it 'runs' into a bytestream file since its
'foptions' settings 'identify' it as a variable record file but it lacks
the 'required' block structure.  In addition it seems that the 'NOBUF'
'request' is being ignored.

Does anyone know the particular bit settings that would identify a
'bytestream' file to a CM utility?  Is there a way of enabling the 'NOBUF'
option and if so, how would one 'extract' the qualifying records within
each block?

Regards
Paul Christidis