On Fri, 16 Dec 2022 15:03:54 -0600, Kirk Wolf wrote: >The OP didn't specify IBM FTP, so I'll mention that could use Co:Z SFTP, which >has a subcomment "dsput" which will allocate a dataset on the target system >matching the source system and then upload the dataset in binary, preserving >record formats. > ... >Note: this works for RECFM=VBS with LRECL <= 32760, which is the range allowed >by dynamic allocation. If you don't specify a DCB LRECL on your IFASMFDP >job, you will get RECFM=VBS with the magic LRECL=32767 (LRECL=X). > Why not respect the attributes in the DSCB, which are merged into the DCB by the OPEN macro? Or are you relying on a technique similar to IEBCOPY which represents a PDS block in the PDSU as a logical record in a RECFM=VBS data set.
> ... This is IMO a bug, since the maximum record length for SMF is 32756. > So put DCB=(RECFM=VBS,LRECL=32760) and stay away from that nonsense. > I believe that LRECL=X is represented as '8000'x (32768). I learned this on one of these fora; I nave not found it in any Ref. (But I haven't scrutinized all the Data Areas manuals.) In any case, BPXWDYN accepts 32768 and creates a LRECL=X data set. (The BPXWDYN designer may have been influenced by Doug Gwyn's maxim: <https://quotefancy.com/quote/1711236/Douglas-Gwyn-Unix-was-not-designed-to-stop-you-from-doing-stupid-things-because-that>.) And the envelope should contain a representation of the attributes so they would be restored automatically on the return trip. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN