From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.5-pre1 (2020-06-20) on
	ip-172-31-74-118.ec2.internal
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=3.0 tests=BAYES_00 autolearn=ham
	autolearn_force=no version=3.4.5-pre1
Date: 10 Apr 93 19:52:36 GMT
From: alex@MIMSY.CS.UMD.EDU  (Alex Blakemore)
Subject: Re: and then
Message-ID: <66288@mimsy.umd.edu>
List-Id: <comp.lang.ada>

In article <1993Apr9.180808.19494@scf.loral.com> bashford@srs.loral.com (Dave B
ashford) writes:

> I can't believe my eyes - "desired side-effect" in Ada ?
> The only reason I can think of not to optimize to short circuit forms is to c
atch
> undesired side-effects.

typically I agree
but what's important is to have the same behavior on different systems so 
if a function may raise an excepion - it should be called just in case.

> Are there "desired side-effects" ?
 
sometimes:

  one class of beneficial side affect doesnt effect semantics but may
affect performance - .e.g. data structures that reorganize on queries
to optimize requests for recently visited info - akin to caching

  somtimes a side effect is intended every time a function is evaluated:
such as in a pseudo-random number generator or a lexical scanner perhaps
but its by far the exception and should be documented.

> Should compilers optimize when, because of poor programming, the
> behaviour of the program might change ?

this caused a lot of debate in the Ada9X forum in the case when 
certain range checks could be safely eliminated (optimized away) 
without changing the semantics _except_ in the erroneous case
where a variable is referenced before being initialized.

If I understand the new 9X proposal, they have a reasonable solution.
-- 
---------------------------------------------------
Alex Blakemore alex@cs.umd.edu   NeXT mail accepted