On Fri, 20 Oct 2017 10:46:27 +0800, David Crayford wrote: >>> >> pax works well with archives in Classic data sets. (I've used DSORG=PS; >> never tried PO.) But why not pass the DSN(MEMBER) in PARM rather than >> a DD statement. If 100-character limit is intolarable, alternatives include >> STDPARM, PARMDD, and STDENV. > >Data set names in parms are ugly! > How?
>> And beware. No standard utilities (e.g. pax) are specified to work with >> //DD: >> and only a handful with //DSN. Otherwise, it's an accident if they work, and >> unsupported if they break. > >What's the difference between //DSN and //DD:? The way I see it it's >just an argument to fopen() and either will work. Maybe IBM should give >an example with DD: to keep the pedants quiet but I wont lose any sleep >over it. > It is nowhere documented that the six (exactly) standard utilities that support data set names use fopen(). That's a good guess, but IBM implicitly retains the liberty to change that implementation in the future. You are relying on things that happen to work but IBM makes no commitment to support. If it breaks, you're entitled to keep both pieces. Otherwise, RFE? Oops! I RTFM and stand semi-corrected. The 2.3 UNIX Command Ref., in the section, "Appendix K. Specifying MVS data set names in the shell environment" mentions, with error bars, a Restriction on using "the //DD:DDNAME format". Ironic, because the (rather informal) syntax in that section never mentions that format's being supported at all. I believe I'll submit an RCF. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
