On Wed, 12 Sep 2018 17:47:00 +0200, R.S. wrote: >>> -------------------------------------------------- >>> ------------------------------------------------- >>> ** SYSUT1.T0849100 PS U 0 666 >>> ... >> It's shortsighted design that many utilities don't suport deleting "strange >> files". > >IMHO it's even worse that any program/utility is able to create such >creatures. > I agree. The rules should be enforced uniformly by a single central component (DFSMS?) and apply alike to DYNALLOC, JCL, TSO,ISPF, ... APIs have no business enforcing such rules. They should pass the argument, unmodified to the basic service and report any errors returned. If DFSMS allows mixed case in (uncatalogued) data set names, TSO should provide a means to manipulate data sets so named.
JCL adds complexity because of the desirability of validating names in the reader, not the interpreter. One might invoke Postel's Principle to argue either side of this matter. I have only contempt for the practice of Security by Naïveté: that I might be able to create an object with such an obscure name that you don't know how to manipulate it. >Regarding above datasets - those are quite normal. SYSUT1 can be legal >HLQ, second qualifier is also legal one. DCB parameters are legal as >well, which is irrelevant for delete operation. > I was wondering about that. They look like <ddname>.<jobname>. Temp DSNs are similar, but have more qualifiers. Did RACF intervene? -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
