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: 17 Dec 92 22:32:25 GMT
From: aio!dnsurber@eos.arc.nasa.gov  (Douglas N. Surber)
Subject: Re: Novice Question on Record Representation
Message-ID: <dnsurber.724631545@node_26400>
List-Id: <comp.lang.ada>

In <1992Dec17.182403.6353@telesoft.com> garym@telesoft.com (Gary Morris @pulsar
) writes:

>Unfortunately, you can't rely on assertion like this.  The constant a1 is
>not used anywhere and an optimizing compiler could remove a1 and the
>expression used for its initial value (TSize = T'Size).  So the range check
>never occurs and no exception or compile time warning is generated.  This
>optimization is allowed by 11.6(7), when the only effect of a predefined
>operation is to raise an exception. 

But a1 could be used in the package body, therefore it can't be optimized
out during compilation of the package spec.  As I recall there is no
requirement to issue a compile time warning in cases like this, but I
am not aware of any compiler that doesn't.  Does anyone else know of any?
I use the subtype Assertion a lot in my code and it occasionally catches
things, but I guess I really shouldn't rely on it.


--
Douglas Surber             "Would you rather debug at
Lockheed                    compile time or run time?"
Houston, TX                      --Michael B. Feldman