In REXX, "ARG [template]" is just a short form of the instruction "PARSE UPPER ARG [template]". Hence INTERPRET on passed ARG variables produces uppercase results (unless PARSE ARG is coded instead of ARG). There is no REXX equivalent to CLIST's "CONTROL CAPS ASIS". CLIST: https://www.ibm.com/docs/en/zos/2.1.0?topic=reference-control-statement REXX's "VALUE()" is CLIST's "&EVAL()" and AFAIK neither can be nested/recursive - whereas INTERPRET can be recursive, but less so than CLIST's &&s. Cheers, over and out.
On 14/09/2021 21:41, Paul Gilmartin wrote: > On Tue, 14 Sep 2021 13:56:42 +0000, Seymour J Metz wrote: > >> No, the evaluation of the operand is the same as for any other expression >> and the interpretation of the evaluated operand is the same as for any other >> statement in REXX, including the treatment of apostrophes and quotation >> marks: >> >> interpret 'foo = "bar baz"; say "foo="foo' >> > Yes, but the naive programmer is apt to be astonished by such as: > > drop X > A = X > X = 'Roman numeral ten' > interpret 'say' A > interpret 'say A' > interpret say '''A''' > > generally start my EXECs with SIGNAL ON NOVALUE. > So I am astonished sooner than later during testing > and repairs are easier. > > -- gil > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN > . > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
