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,FREEMAIL_FROM
	autolearn=ham autolearn_force=no version=3.4.4
X-Google-Thread: 103376,86c76377d1077f30
X-Google-Attributes: gid103376,public
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news2.google.com!news4.google.com!border1.nntp.dca.giganews.com!nntp.giganews.com!newsfeed00.sul.t-online.de!t-online.de!tiscali!newsfeed1.ip.tiscali.net!proxad.net!cleanfeed1-b.proxad.net!nnrp11-2.free.fr!not-for-mail
Message-ID: <44c9a4a3$0$691$626a54ce@news.free.fr>
From: Yves Bailly <kafka.fr@laposte.net>
Subject: Re: Casting from interface type
Newsgroups: comp.lang.ada
Date: Fri, 28 Jul 2006 07:46:08 +0200
References: <44c94166$0$11790$626a54ce@news.free.fr>
 <1154050141.462907.39690@b28g2000cwb.googlegroups.com>
User-Agent: KNode/0.10.2
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7Bit
Organization: Guest of ProXad - France
NNTP-Posting-Date: 28 Jul 2006 07:46:11 MEST
NNTP-Posting-Host: 81.56.171.53
X-Trace: 1154065571 nnrp11-2.free.fr 691 81.56.171.53:38197
X-Complaints-To: abuse@proxad.net
Xref: g2news2.google.com comp.lang.ada:5979
Date: 2006-07-28T07:46:11+02:00
List-Id: <comp.lang.ada>

Adam Beneschan wrote:

>> Using the latest GNAT 2006, I receive the following errors:
>> proc.adb:14:19: invalid tagged conversion, not compatible with
>> type "I'Class" defined at line 3
> 
> Type T is not descended from I, so there cannot be any object in
> I'Class that has type T.

This makes sense to me?

> However, the following should be legal, 
> although GNAT probably won't accept it (see below):
> 
>     var_t := T (T'Class (p));
> 
>> proc.adb:15:21: invalid tagged conversion, not compatible with
>> type "I'Class" defined at line 3
> 
> I think this is legal and GNAT is wrong.  My printed copy of
> 4.6(23.1/2) says that type conversions between *any* two class-wide
> types are legal if one of them has an interface type as its specific
> type, even if the other one doesn't have an interface ancestor, because
> some type in the class *could* have an interface ancestor (as DT does
> in your example).  I tried to check to make sure this section hasn't
> changed since I printed it, but unfortunately www.adaic.com is down
> right now.

Just checked, it hasn't changed since when I downloaded it. This is
this precise clause which made me wondering if I misunderstood
something. Well, maybe I'll fill a bug report...

Thanks for your help.

Regards,

-- 
(o< | Yves Bailly  : http://kafka-fr.net   | -o)
//\ | Linux Dijon  : http://www.coagul.org | //\
\_/ |                                      | \_/`