Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!ames!ucbcad!ucbvax!KL.SRI.COM!STEINBERGER
From: STEINBERGER@KL.SRI.COM (Richard Steinberger)
Newsgroups: comp.os.vms
Subject: Re-entrant code
Message-ID: <12296042060.20.STEINBERGER@KL.SRI.COM>
Date: Mon, 20-Apr-87 11:16:21 EST
Article-I.D.: KL.12296042060.20.STEINBERGER
Posted: Mon Apr 20 11:16:21 1987
Date-Received: Tue, 21-Apr-87 01:12:48 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 23


As I once again learned, you have to disable interrupts (e.g. with a
SYS$SETAST call) before calling non-re-entrant routines if AST routines
also call them.  This is (obviously) to prevent the data structures in the
non-re-entrant code from getting trashed.

So what's the concern:  is it possible to write re-entrant code on a VAX
in higher level languages like FORTRAN?  If so, are there guidelines written
anyplace?  Thanks to all who reply.

-Ric Steinberger
steinberger@kl.sri.com


BTW:  The symptom that something was wrong in my coding was as follows:
The program would fail with some sort of fatal error or other.  I would
use the debugger to trace control flow and find that the received arguments
in a subroutine that doesn't alter them were different that what they were
in the calling program.  Then I considered that the receiving subroutine
was also called by an AST. . . .grrrr!


-------