Gil, Actually I would guess that this particular change is very possible.
I would say that it would require (relatively minor) changes to the Interpreter and Allocation in addition to the Converter. The changes to the Interpreter and Allocation would be needed to allow the request for a regarding "relative" label (e.g. LABEL=(NEXT,...)) to be handled. The good news is that with these changes in place considerations COND/IF do not obtain, the label sequence number is resolved at allocation time rather than at JCL creation time (e.g. late binding). The bad news is that if this facility existed the label sequence number becomes indeterminate (because of COND/IF). Not the end of the world if the dataset is cataloged (since the label sequence number is kept in the catalog entry) but somewhat more problematic if the dataset is uncataloged :-( But then again it doesn't change the requirement to know the label sequence for uncataloged tape datasets when you want to read them, it just makes it a little more work :-) Personally I like this idea. However it falls outside of the scope I am trying to maintain, to wit the Converter (only). If the scope of this effort grows to encompass the Interpreter and Allocation then I would certainly advocate for this idea. I am very concerned about "scope creep" and thus I am "drawing a line in the sand" in a desperate attempt to stay out of the ocean :-) BTW, thanks for bringing this thread back to it's original spirit :-) John McDowell >Another. I think this is out of the "boiling the ocean" category. > >o Automatic generation of position numbers for multi-file tapes. Something >like: > > LABEL=(1+*.stepname.ddname,SL),... > > or even: > > VOL=REF=*.stepname.ddname,LABEL=(NEXT,SL),... > > where the position will be 1 greater than that in the most recent > step mentioning the volume, regardless that the referenced stepname > is not the most recent. > > (Assuming that no intervening steps are flushed by COND or IF. > accommodating flushed steps is "boiling the ocean".) ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
