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: a07f3367d7,e276c1ed16429c03
X-Google-Attributes: gida07f3367d7,public,usenet
X-Google-NewGroupId: yes
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news2.google.com!news3.google.com!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From: "Vinzent Hoefler" <nntp-2010-10@t-domaingrabbing.de>
Newsgroups: comp.lang.ada
Subject: Re: Ada is getting more popular!
Date: Thu, 14 Oct 2010 01:05:34 +0200
Message-ID: <op.vkjefkhi0k3wt7@jellix.jlfencey.com>
References: <i93mac$73v$1@speranza.aioe.org>
 <bbe0c03c-6afd-40de-bf9e-04aa04b53504@n7g2000vbo.googlegroups.com>
 <i94sak$47q$1@tornado.tornevall.net>
 <020ea727-e8e9-4d0d-8c5f-b8ad907fc3c9@u10g2000yqk.googlegroups.com>
 <op.vkjcy2vv0k3wt7@jellix.jlfencey.com>
 <8edaf552-0104-4150-bb55-5557e8e24be3@i5g2000yqe.googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
X-Trace: individual.net OatKDpf93bqopRItP8bp1wJ6xTrOPKNbfuy0bR9s+pof4yMokK
Cancel-Lock: sha1:85amjvlySHPrBxxw0IyOQQmsHiU=
User-Agent: Opera Mail/10.62 (Win32)
Xref: g2news2.google.com comp.lang.ada:15496
Date: 2010-10-14T01:05:34+02:00
List-Id: <comp.lang.ada>

On Thu, 14 Oct 2010 00:49:40 +0200, ramon_garcia <ramon.garcia.f@gmail.com> wrote:

> On Oct 14, 12:34 am, "Vinzent Hoefler" <nntp-2010...@t-
> domaingrabbing.de> wrote:
>> Huh? Since when do "out" parameters imply a copy?
>
> If the size is determinated at runtime, the caller cannot know the
> size in advance, and cannot allocate it. In fact, GNAT has a
> "secondary stack" for returning variable-size objects.

For functions returning such types, yes. Not for procedures.

>> I don't understand that. Either you are creating the array
>> dynamically, then getting an access to it should not be a problem
>> or the caller already knows the bounds.
>
> The array was created statically, but I do not want code in the caller
> to depend on that.

type Buffer        is array (Positive range <>) of ...;
type Buffer_Access is access all Buffer;

A : Buffer        := (1 .. 4 => ...);
P : Buffer_Access := A'Access;

should work, while

A : Buffer (1 .. 4);
P : Buffer_Access := A'Access;

does not.

> Or a container record was created dynamically, and the size of the
> array is part of the discriminant.
>
> type container(n_elemen: natural) is record
>      a: buffer(n_elemen);
> end record;

In that case you lose, because you cannot create A without explicit
bounds. And with explicit bounds, the x.A'Access won't do.

> Imagine that this type is defined in the private part of the package.
> A method is provided to obtain an access to the buffer, and from that
> access the bounds of the array can be obtained at run time. This is
> not posible in general.

Why don't you provide access to the record then?


Vinzent.

-- 
There is no signature.