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=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID
	autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,9677a3433b89e25b
X-Google-Attributes: gid103376,public
From: Niklas Holsti <nholsti@icon.fi>
Subject: Re: attributes in generic procedure
Date: 1999/11/29
Message-ID: <3841CE51.862ACED3@icon.fi>#1/1
X-Deja-AN: 554229881
Content-Transfer-Encoding: 7bit
References: <818esd$t6l$1@nnrp1.deja.com> <383DC78E.63A556E5@callnetuk.com>
Content-Type: text/plain; charset=us-ascii
X-Trace: read2.inet.fi 943835121 194.252.1.51 (Mon, 29 Nov 1999 02:25:21 EET)
Organization: Space Systems Finland Ltd
MIME-Version: 1.0
NNTP-Posting-Date: Mon, 29 Nov 1999 02:25:21 EET
Newsgroups: comp.lang.ada
Date: 1999-11-29T00:00:00+00:00
List-Id: <comp.lang.ada>

Nick Roberts wrote:
> 
> daelen@my-deja.com wrote:
> >
> > I would like to create a generic procedure as shown below which will only be
> > used for scalar types. Does anybody have a hint how to accomplish this?
> >
> > generic
> >    type A_Type is private;
> > ...
> > end Update;
> 
> The answer is simple:
> 
>    generic
>       type A_Type is (<>);
>    ...

According to RM 12.5.2, this is a formal discrete type definition,
not a formal scalar type definition. This means that it allows only
enumeration and integer types, but not real types (3.2(3)).

In fact, there is no formal_scalar_definition in the RM. I, too, would
have liked to have one when I recently wrote a generic function to
take the maximum value of a vector of some generic element type.
The Max function needs only the attributes 'first and 'max, which
exist for all scalar types, but different generic functions are
needed for a discrete element type and a real element type, although
the bodies of the two functions are identical. One can of course work
around this by making 'first and 'max be additional generic formal
parameters, instead of using type attributes, but it is a little messy.

I don't know if it's worthwhile to consider adding a formal scalar
class to Ada. It might be difficult for compilers that share generic
code across instantiations 

Niklas Holsti
Working at but not speaking for Space Systems Finland Ltd.