On Wed, 1 Dec 2021 17:00:55 +0000, Seymour J Metz wrote:
> ... IBM is not even upgrading REXX in TSO to the level of the ANSI
> standard.
>
MVS presents challenges to ANSI Rexx.
What should be the semantic of
address LINKMVS with error ...
Does "error " refer to SYSTERM? Other (specify)?
How should linein() and charin() treat Classic data sets containing '15'x
as a data character? Perhaps just rely on C RTL fopen() and blame
that for any behavior. OMVS paths are natural.
Regina wisely (IMO) directs TRACE and message output to stderr by default.
TSO Rexx designers have declined to do likewise for OMVS Rexx.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
