I did try it, and I agree with Gil, and would add that the whole EXPORT facility is rather bizarre. First, the SYSUT1 SYSIN received 'AFTER' for the &S symbol. I suppose that means it is substituted sometime after the JCL is interpreted. Maybe it was easiest to implement it this way (along with weird extra JCL inserts).
sas On Fri, Sep 16, 2016 at 8:44 PM, Paul Gilmartin < [email protected]> wrote: > On Fri, 16 Sep 2016 23:37:57 +0000, Frank Swarbrick wrote: > > > >Maybe this isn't a real enough situation to matter, but it seems odd to > me. > > ... > Also, I was astonished by the behavior of (try it!) > > //STEP EXEC PGM=IEBGENER > //SYSIN DD * > //SYSPRINT DD SYSOUT=(,) > //* > // EXPORT SYMLIST=S > //* > // SET S='Before' > //SYSUT1 DD *,SYMBOLS=JCLONLY > The value of S was set &S DD SYSUT1!? > //* > // SET S='DURING' > //SYSUT2 DD SYSOUT=(,&S) What's the writer? > // SET S='After' > //* > Experts tell me it's intuitive; I feel otherwise; they could have done > better. > The behavior is irritating when I try setting a symbol before calling a > PROC > and get unexpected results. > > -- gil > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN > -- sas ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
