It's long enough for DLM=THISISAUNIQUEDELIM -----Original Message----- From: IBM Mainframe Discussion List <[email protected]> On Behalf Of Radoslaw Skorupka Sent: Thursday, May 8, 2025 5:30 AM To: [email protected] Subject: Re: Check for ENQ at GDG level from rexx (non-ispf)
W dniu 07.05.2025 o 05:35, Paul Gilmartin pisze: > On Tue, 6 May 2025 22:02:03 +0000, Lennie Bradshaw wrote: >> ... >> Or you could supply the incoming data set instream (in xmit format for >> example). This latter method requires you to be sure you can terminate your >> instream data set correctly. So, the first 2 characters of each record MUST >> NOT match the termination string you specify in your JCL DLM value. I have >> used that method in the past, but scanned the incoming stream for several >> pairs of characters, then choose one that was not present. >> ... > That's interesting. I'm pretty sure I've at tines supplied instream SYSYN > with no > explicit delimiter. Ir was simply the last data set in the last job step. > But was > there nonetheless an implied delimiter; a digraph that I fortuitously avoided? > I.e. is there no construct notionally "DLM=none"? Update: since few years the delimiter in can be 2 to 18 characters long. (it applies to JES2, not JES3). 18 characters. IMHO it's enough to find unique string. It's more than enough to find a string which is not a part of any programming language, command language or any human language. Of course there's still possibility to use DD * for binary or generated random data, but in real world I would call it very unlikely use of DD *. -- Radoslaw Skorupka Lodz, Poland ---------------------------------------------------------------------- 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
