On Thu, 24 Aug 2017 08:09:18 -0500, Elardus Engelbrecht wrote: >Paul Gilmartin wrote: > >>>> ... RECFM=FB. ... >>Why is that still a thing in the 21st Century? > >Backward compatibility all the way back to Julius Ceasar and Nero ... ;-D > >What would you like to have in place of FB? > It could be done, with extreme backward compatibility. Consider that for a UNIX file if the application OPENs it for INPUT with RECFM=FB, the access method (it was briefly called "XSAM") pads each record to LRECL with blanks and removes the NL character; for RECFM=VB, it converts each NL to a synthesized RDW. For OUTPUT, the operations are reversed (but trailing spaces are not removed). (Except for Binder which pigheadly overrides JCL DD attributes and does its own stupid thing.)
Decades ago, before OMVS was envisioned, I was maintaining/enhancing the runtime library for a FOSS Pascal system and wished the access method would do something similar for Classic data sets. It would have saved me considerable support code and a Pascal application could have seen everything as RECFM=V. Such support for Classic data sets still merits an RFE, IMHO. Even better than mere backward compatibility. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
