On Fri, 27 May 2016 05:45:05 -0700, Lizette Koehler wrote: > >I have only seen BLKSIZE=3120 in the IBM Documentation for TSO XMIT. > >OUTDDNAME(ddname) | OUTFILE(ddname) > specifies the use of a preallocated file as the output data set for the >TRANSMIT command. No data is written to SYSOUT for transmission and the system >limit on the number of records that can be transmitted does not apply. TSO/E >assigns the DCB attributes as LRECL=80, BLKSIZE=3120, and RECFM=FB. Specify the >ddname as either a sequential data set or a member of a partitioned data set. > UNIX files and even POSIX pipes work. I don't know whether they're considered supported. According to introductory matter in "Using Data Sets", data sets can be UNIX files, but I've lost that argument with IBM support more than once.
But it raises the tantalizing possibility of piping from TRANSMIT into ssh and having no requirement for a temporary data set (except for the IEBCOPY case.) > Use OUTDDNAME or OUTFILE in conjunction with the INDDNAME or INFILE operand >of the RECEIVE command. OUTDDNAME and OUTFILE are primarily intended for system >programmer use. >OUTDSNAME(dsname) | OUTDATASET(dsname) > specifies the use of a data set as the output data set for the TRANSMIT >command. No data is written to SYSOUT for transmission and the system limit on >the number of records that can be transmitted does not apply. TSO/E assigns the >DCB attributes as LRECL=80, BLKSIZE=3120, and RECFM=FB. The data set must be a >sequential data set. > >And I agree that 27920 is better. But I have not seen any doc from IBM stating >it was good for XMIT. If you have clarifying documentation, could you share? > Is TRANSMIT one of those accursed utilities that has DCB attributes hard coded and writes them back to the DSCB, possibly corrupting a PDS? -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
