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=unavailable autolearn_force=no version=3.4.4
Path: 
 eternal-september.org!reader01.eternal-september.org!reader02.eternal-september.org!news.eternal-september.org!mx02.eternal-september.org!.POSTED!not-for-mail
From: Simon Wright <simon@pushface.org>
Newsgroups: comp.lang.ada
Subject: Re: Generic parent accessing his child?
Date: Sat, 06 Dec 2014 07:49:41 +0000
Organization: A noiseless patient Spider
Message-ID: <lyoarh2p62.fsf@pushface.org>
References: <cef4t5Ff8n5U1@mid.individual.net>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: mx02.eternal-september.org;
 posting-host="d7244c06a78575e4dbdf34623acd89dd";
	logging-data="4362"; mail-complaints-to="abuse@eternal-september.org";
	posting-account="U2FsdGVkX1/C/Jp86JG7ciEhbG3SDzxlZmcm62FzLbg="
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.3 (darwin)
Cancel-Lock: sha1:PW1WBtYr2GBKpUKj4Fi1Iykv+SA=
	sha1:i2dwEkXYlbG5micMG02dPDGpK9g=
Xref: news.eternal-september.org comp.lang.ada:23895
Date: 2014-12-06T07:49:41+00:00
List-Id: <comp.lang.ada>

hreba <hreba@terra.com.br> writes:

> Without generics I had no problem letting a parent access his child, a
> "with" was enough. With generics I have the following problem:
>
> -------------------------------------------------------
>
> generic		-- a package used as generics parameter
> package T is
>    type	Ttype is private;
>    procedure Init (t: in out Ttype);
> private
>    type Ttype is new Integer;
> end T;
>
> package body T is
>    procedure Init (t: in out Ttype) is
>    begin
>       t:= 0;
>    end Init;
> end T;
>
> --------------------------------------------------------
>
> with T;		-- the parent
> generic
>    with package U is new T(<>);
> package P is
>    type AD is abstract tagged null record;
>    type A is access all AD'Class;
>    function NewA return A;
> end P;
>
> with P.Q;
> package body P is
>    function NewA return A is
>    begin
>       return A(P.Q.NewB);	--  <-- ERROR
>    end NewA;
> end P;
> -------------------------------------------------------
>
> generic		-- the child
> package P.Q is
>    type BD is new AD with
>       record
>          i:	U.Ttype;
>       end record;
>    type B is access all BD;
>    function NewB return B;
> end P.Q;
>
> package body P.Q is
>    function NewB return B is
>       b0: B;
>    begin
>       b0:= new BD;
>       U.Init(b0.i);
>       return b0;
>    end NewB;
> end P.Q;
>
> -----------------------------------------------------
>
> The error message refers to the line marked above and reads:
>
> Invalid prefix in selected component "Q".
>
> What would be the correct way?

Compiled with -gnatwa (enables many additional warnings) gives

    22. with P.Q;
              |
        >>> warning: unit "P.Q" is never instantiated

    23. package body P is
    24.    function NewA return A is
    25.    begin
    26.       return A(P.Q.NewB);
                        |
        >>> invalid prefix in selected component "P.Q"

    27.    end NewA;
    28. end P;

because P.Q is itself generic.

You can get the code to compile by

   with P.Q;
   package body P is
      package My_Q is new P.Q;
      function NewA return A is
      begin
         return A(My_Q.NewB);
      end NewA;
   end P;

if that does what you want.