Path: utzoo!attcan!uunet!husc6!uwvax!oddjob!ncar!noao!arizona!debray
From: debray@arizona.edu (Saumya Debray)
Newsgroups: comp.lang.prolog
Subject: detecting freedom from side effects (was Re: Clause fusion)
Summary: need to consider asserts
Message-ID: <5504@megaron.arizona.edu>
Date: 16 May 88 21:05:30 GMT
References: <983@cresswell.quintus.UUCP>
Organization: U of Arizona CS Dept, Tucson
Lines: 14

In article <983@cresswell.quintus.UUCP>, Richard A. O'Keefe writes:

> Suppose that the compiler has processed the definition of pop_stack/4
> and knows that it is free of side effects.  [In the absence of call/1,
> this is trivially easy to determine.]

It's also necessary to take "assert" into consideration: if someone,
somewhere, can assert a clause for pop_stack/4, all bets are off.  Doing
flow analysis to deal with assert and retract can get quite hairy.
-- 
Saumya Debray		CS Department, University of Arizona, Tucson

     internet:   debray@arizona.edu
     uucp:       {allegra, cmcl2, ihnp4} !arizona!debray