From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=0.7 required=5.0 tests=BAYES_00,INVALID_DATE,
	MSGID_SHORT,REPLYTO_WITHOUT_TO_CC autolearn=no autolearn_force=no
	version=3.4.4
Path: utzoo!attcan!uunet!actnyc!djs
From: djs@actnyc.UUCP (Dave Seward)
Newsgroups: comp.lang.ada
Subject: Re: Alsys compiler bugs???
Message-ID: <960@actnyc.UUCP>
Date: 10 Jun 88 23:24:01 GMT
References: <8806081155.AA02381@wooglin.scc.com>
Reply-To: djs@actnyc.UUCP (Dave Seward)
Organization: InterACT Corporation
List-Id: <comp.lang.ada>

In article <8806081155.AA02381@wooglin.scc.com> chuck@WOOGLIN.SCC.COM (Charles Williams) writes:
>15    type Fix is delta 0.01 range -1.0 .. 1.0;
>16    My_Fixed : Fix := Fix'Last;
>                        ^______^
>                        1      1
>1 :   *EXP Value outside the range of the target type. NUMERIC_ERROR will be
>            raised if executed.
>
>This error is very confusing to me as the syntax appears perfectly legal and
>the operation also looks fine.
>...and who's wrong? and more importantly, why?

Fixed point has several tricky points like this. In the simplest terms, the
delta implies a "small" value, usu. a power of 2, in this case 2.0**(-7).
The representable values of the type are integral multiples of the small value.
The specified bounds, ie -1.0 and +1.0, don't have to be representable, as
long as they are within small of being representable. In this case, the largest
representable number is (1.0 - 2.0**(-7)). The logic for this is discussed
in 3.5.9(6). This is a vast simplification, but might get you started toward
understanding it.

BTW, I'm posting as opposed to Emailing as I think a fair number of people
are confused by the same issues, and if so a discussion will ensue.

Dave Seward
uunet!actnyc!djs