Joel Ewing has made a valid point about programs potentially having LRECL expectations. COBOL is good for that.
Tim Brown is silent on what he actually wants to do this for. Until then it's difficult to suggest something concrete. Ditch the blocksize has been said, making the LRECL smaller has no obvious benefit has been said. Just to add that the LRECL can always be "overridden" on the DD for a subsequent reference. On Wednesday, 2 March 2016 08:24:22 UTC, Ted MacNEIL wrote: > Why not just create a VBA file with a very long LRECL and not worry about it > at all? Longer LRECLs don't introduce any more ov‎eReader than short ones. > > -teD > Original Message > From: Kjell Holmborg > Sent: Wednesday, March 2, 2016 02:54 > To: [email protected] > Reply To: IBM Mainframe Discussion List > Subject: Re: rexx and tso alllocate > > One suggestion might be that your rexx program writes records to a stem > variable and you could keep track of the longest record and then just before > writing the contents of the stem variables to the dataset you do a TSO > Allocate with the longest record as a variable to the ALLOCATE command. > > /Kjell > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
