I haven't z/OS 2.1 to experiment with. Some of this is empirical under
1.13; I can't find documentation.
I observe the following behavior under 1.13, JES2:
o If no SYSIN record exceeds 80 bytes, the data set is opened with
DCBRECFM=F, DCBLRECL=80.
o If all SYSIN records have identical length >= 80, the data set is
opened with DCBRECFM=F, DCBLRECL=(that length).
o If any two records have different length, and one of them has
length > 80, the data set is opened with DCBRECFM=V,
DCBLRECL=(4 + greatest length).
Is the behavior the same under 2.1?
http://pic.dhe.ibm.com/infocenter/zos/v2r1/topic/com.ibm.zos.v2r1.ieab600/xddsymbols.htm
http://pic.dhe.ibm.com/infocenter/zos/v2r1/topic/com.ibm.zos.v2r1.ieab600/jclsymstr.htm
Is the determination made before or after symbol substitution? For
example, if all records have length = 80 before substitution, and
substitution causes some (but not all) records to have length > 80,
what gets merged into the DCB?
http://pic.dhe.ibm.com/infocenter/zos/v2r1/topic/com.ibm.zos.v2r1.ieab600/iea3b6_Relationship_to_other_parameters9.htm
I observe that with 2.1 DCB=LRECL (but not DCB=RECFM; why!?)
is newly compatible with SYSIN. How does this interact with symbol
substitution's possibly elongating records?
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
