[877] in java-interest
Re: Java implementations should be tail recursive
daemon@ATHENA.MIT.EDU (Martin Odersky)
Sat Aug 12 11:15:45 1995
To: jcdst10+@pitt.edu
CC: ramsdell@linus.mitre.org, java-interest@java.sun.com, pat@icon-stl.net,
ramsdell@linus.mitre.org
In-reply-to: <Pine.3.89.9508110414.A16815-0100000@unixs1.cis.pitt.edu> (message from James C Deikun on Fri, 11 Aug 1995 04:39:07 -0400 (EDT))
Reply-to: odersky@ira.uka.de
Date: Sat, 12 Aug 1995 13:51:19 +0200
From: Martin Odersky <odersky@ira.uka.de>
James C Deikun writes:
The problem with this, of course, is that without saving the old stack
frame, it's impossible to tell who invoked the new method. This
unfortunately conflicts with many of the security methods that seem to be
used in Java, or with just about any rational sort of security for a
language that can safely and stably allow arbitrary runtime extension of
a system.
Can you be more specific? I can imagine that a stack dump in a
debugger would look stange, but are there other reasons to keep
around the caller's frame in a tail-call?
Cheers
-- Martin Odersky
======================================================================
Prof. Martin Odersky e-mail: odersky@ira.uka.de
Department of Computer Science Tel: +49 721 608 3495
University of Karlsruhe Fax: +49 721 69 40 92
76128 Karlsruhe, Germany http://wwwipd.ira.uka.de/~odersky
C is the ultimate computer virus (Dick Gabriel)
-
Note to Sun employees: this is an EXTERNAL mailing list!
Info: send 'help' to java-interest-request@java.sun.com
