On Mon, 16 Mar 2020 18:24:11 +0000, Seymour J Metz wrote: >> Dammit! Regina gets it right.
>No - the syntax that you are using pertains to the shell and will work with >any command that expects input from stdin. It's called a here document. > >>532 $ regina /dev/fd/3 3<<endrexx > here-document, yes; stdin, no. (I often find "called" condescending.) I used it as an extreme instance of a pathname as a Rexx script. >>it allows (sometimes requires) a function's definition before its use, >WTF? REXX function are normally written after their use, for very good >reason. The one exception is that when one function calls another it doesn't >matter which comes first. > I stand corrected; lapsus calami. On Mon, 16 Mar 2020 18:34:13 +0000, Seymour J Metz wrote: >I made the same mistake you did; parse external reads from the terminal, >not frohm the external data queue. The concept of end-of-file is inapplicable >for terminal input, although both TSO and *ix have kludges for files >mapped to the terminal. > I would have said, rather, "mapped to stdin"; hardly a kludge. Likewise IRXJCL. Though Rexx presupposes that PARSE EXTERNAL is reading from a terminal. To its credit, it does not use /dev/tty. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
