Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!pasteur!ucbvax!NAIF.JPL.NASA.GOV!PJS
From: PJS@NAIF.JPL.NASA.GOV (Peter Scott)
Newsgroups: comp.os.vms
Subject: CALLs and SUBROUTINEs
Message-ID: <880422114821.0000009E131@naif.JPL.NASA.GOV>
Date: 22 Apr 88 19:48:21 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 32

Okay, so I started to get into subroutines, everything was going
great until I ran up against a problem that I boiled down to this:
if I run the following command procedure:
------------------------------------------------------------------------
$	CALL a
$	EXIT
$ !
$ a:	SUBROUTINE
$	d = 1
$ loop:
$  	CALL c
$	d = d + 1
$	GOTO loop
$	EXIT
$	ENDSUBROUTINE
$ !
$ b:	SUBROUTINE
$	EXIT
$	ENDSUBROUTINE
$ !
$ c:	SUBROUTINE
$	EXIT
$	ENDSUBROUTINE
------------------------------------------------------------------------
then after the line "CALL c", both d and loop are undefined (you can
take loop out if you want, I was just illustrating that labels are
affected too).  However, if I remove subroutine B, everything goes
swimmingly.  (In the original file, B was called later on.)

What is going on???  How can I lose the scope of my pseudo-procedure
just by calling subroutine c?  Should I write an SPR?

Peter Scott (pjs%grouch@jpl-mil.jpl.nasa.gov)