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=-1.9 required=5.0 tests=BAYES_00 autolearn=ham
	autolearn_force=no version=3.4.4
X-Google-Thread: 103376,5aa1f17e1cbe0afb
X-Google-Attributes: gid103376,domainid0,public,usenet
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news1.google.com!news2.google.com!news1.google.com!news.glorb.com!peer1.news.newnet.co.uk!194.159.246.34.MISMATCH!peer-uk.news.demon.net!kibo.news.demon.net!news.demon.co.uk!demon!diphi.demon.co.uk!jpt
From: JP Thornley <jpt@diphi.demon.co.uk>
Newsgroups: comp.lang.ada
Subject: Re: SPARK and static unit checking?
Date: Sun, 13 Apr 2008 20:39:34 +0100
Message-ID: <O1nEKjD2FmAIJwBH@diphi.demon.co.uk>
References: 
 <af9b42f7-b24c-4ca5-8a49-1ea9c15e1f37@y21g2000hsf.googlegroups.com>
NNTP-Posting-Host: diphi.demon.co.uk
Mime-Version: 1.0
Content-Type: text/plain;charset=us-ascii;format=flowed
X-Trace: news.demon.co.uk 1208115699 11420 80.177.171.182 (13 Apr 2008
 19:41:39 GMT)
X-Complaints-To: abuse@demon.net
NNTP-Posting-Date: Sun, 13 Apr 2008 19:41:39 +0000 (UTC)
User-Agent: Turnpike/6.07-S (<tazlimhlassK3RFOF7bcA7ehMP>)
Xref: g2news1.google.com comp.lang.ada:20914
Date: 2008-04-13T20:39:34+01:00
List-Id: <comp.lang.ada>

In article 
<af9b42f7-b24c-4ca5-8a49-1ea9c15e1f37@y21g2000hsf.googlegroups.com>, 
Jacob Sparre Andersen <jspa@nykredit.dk> writes
>I have found a rather annoying (but in some ways very reasonable) rule
>in SPARK:
>
>   You are not allowed to (re)declare operators for a type.
>
>This prevents me from using my standard trick for static unit type
>checking:
>
>   type Length is private;
>   function "+" (L, R : Length) return Length; -- not SPARK
>   ...
>   Meter : constant Length;
>
>   type Area is private;
>   function "*" (L, R : Length) return Area; -- not SPARK
>   ...
>
>I have an idea for a (clumsy) solution:  Generate two packages with
>the same types (name-wise).  The one as above, the other simply as:
>
>   type Length is new Float;
>   type Area is new Float;
>
>Then I can use the liberal version with the SPARK tools, and swap the
>restrictive in for the proper compilation.

Ummmm, sorry no - derived types aren't allowed either (except for 
extending tagged record types).

>
>Any proposals for an elegant solution?

The nearest you can get to derived types is simply:

    type Length is digits 6;

but this will require type conversions that are not required by the 
"proper" version (and which will be required by the derived types as 
well).

I guess the only way is an unconstrained subtype:

    subtype Length is Float;

(hardly elegant, but ...)

Phil Thornley

-- 
JP Thornley