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=-2.3 required=5.0 tests=BAYES_00,INVALID_MSGID,
	MAILING_LIST_MULTI,WEIRD_PORT autolearn=no autolearn_force=no
	version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,306c0e782f999b0e
X-Google-Attributes: gid103376,public
From: Simon Wright <simon@pogner.demon.co.uk>
Subject: Re: Booch components
Date: 2000/10/18
Message-ID: <x7vg0lthm8b.fsf@smaug.moho>#1/1
X-Deja-AN: 683900139
Content-Transfer-Encoding: 8bit
References: <MABBJLHONCNCBAPGDDMMKEFPCAAA.lial@khrts.khv.ru>
To: comp.lang.ada@ada.eu.org
Content-Type: text/plain; charset=iso-8859-1
X-Complaints-To: usenet@enst.fr
X-Trace: menuisier.enst.fr 972068535 2150 137.194.161.2 (20 Oct 2000 19:02:15
 GMT)
Organization: ENST, France
List-Id: comp.lang.ada mail<->news gateway <comp.lang.ada.ada.eu.org>
X-Mailman-Version: 2.0beta5
X-BeenThere: comp.lang.ada@ada.eu.org
Mime-Version: 1.0
Reply-To: comp.lang.ada@ada.eu.org
NNTP-Posting-Date: 20 Oct 2000 19:02:15 GMT
Newsgroups: comp.lang.ada
Date: 2000-10-20T19:02:15+00:00

"Alexey V. Litvinov" <lial@khrts.khv.ru> writes:

> When I tried to use Booch components in my program,
> I encountered some trouble with next code (It's abstraction of
> database table as file of simular entities.=20

Could you possibly set your newsreader not to generate these '=20'
characters? I think it may have something to do with "quoted
printable"

> With little diffirent mechanism of deleting/appending of
> records from Dbase-like. Because I'll have little couple=20
> of deleted records with my application and I don't wish=20
> make packing of table, I decide use unbounded queue for list=20
> of deleted records and use values from it when I need to=20
> append new records).

Unless you really need the polymorphic aspects of Lists, I *strongly*
recomment that you use Collections.

> with Ada.Direct_IO;
> 
> generic
>    type RecDataType is private;
>    FileName : String;
> 
> package DB_Table is
> 
>    type RecCountType      is private;
>    type RecPosvCountType  is private;
> --.
> --.
> --. -- routines to read/write/append/etc...
> --.
> 
> private
> 
>    package Table_IO is new Ada.Direct_IO(RecDataType);
> 
>    type RecCountType      is new Table_IO.Count;
>    type RecPosvCountType  is new Table_IO.Positive_Count;
> =20
>    type DEItem is record
>       Pos: Table_IO.Positive_Count;
>    end record;
> 
>    package DE_IO is new Ada.Direct_IO(DEItem);
> 
>    TableFile : Table_IO.File_Type;
>    DEFile   : DE_IO.File_Type;
> 
> end DB_Table;=20
> 
> with Ada.Direct_IO;
> with BC;
> with BC.Containers;
> with BC.Containers.Queues;
> with BC.Containers.Queues.Unbounded;
> with BC.Support.Unmanaged_Storage;
> 
> package body DB_Table is
> =20
>    package Unmanaged_Storage renames BC.Support.Unmanaged_Storage;
>    Storage: Unmanaged_Storage.Pool;
> 
>    package Containers is new BC.Containers(Item =3D> DEItem);

The problem is that the BCs are based on tagged types; and you can
only derive new tagged types in a package *spec*. Move all these
instantiations into the spec and you should be OK.

>    package Queues     is new Containers.Queues;
>    package UBQueue    is new Queues.Unbounded
>                                             (Unmanaged_Storage.Pool, =
> Storage);
> 
> --.
> --.
> --. -- routines to read/write/append/etc...
> --.
> 
> end DB_Table;
> 
> After compliling with GNAT 3.13p:
> gcc -c -gnatf db_table.adb
> db_table.adb:13:04: instantiation error at bc-containers.ads:132
> db_table.adb:13:04: parent type must not be outside generic body
> gnatmake: "db_table.adb" compilation error

Yes, this is a strange error message, I'll make a problem report to ACT.

> But this trouble arising only if my package is generic, when I made=20
> just simular simple package compling was saccessful.
> 
> Good Old Question: knew somebody what I'm doing wrong,=20
> or why things are going wrong?
> 
> PS: at bc.containers.ads:132 we have
> private
> 
>   -- We need access to Items; but we must make sure that no actual
>   -- allocations occur using this type.
> 
>   type Item_Ptr is access all Item;
>   for Item_Ptr'Storage_Size use 0;
> 
> <this>
>   type Container is abstract new Ada.Finalization.Controlled with null =
> record;
> </this>

? I don't see what you think is wrong here?



I would have posted this reply in c.l.a, but for the moment my
newsfeed is under revision :-) and I can only read.