Re: Ada is getting more popular!

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Wed, 13 Oct 2010 16:51:19 +0200, Georg Bauhaus wrote:

> On 13.10.10 14:07, ramon_garcia wrote:
>> type my_container(counter: integer, discriminated_attribute: integer)
>> is record
>>    data: array(0..counter) of contained(discriminated_attribute);
>> end record;
>> 
>> This straightforward way is not posible because one cannot use an
>> anonymous array type inside a record (by the way, I find this feature
>> unjustified).
> 
> I think it's justified insofar as you can't have both separate
> compilation and dope-free representation of My_Container.

Note that the following is perfectly legal:

   type My_Container (Counter : Natural) is record
       Data : String (1..Counter);
   end record;

The issue has nothing to do with array dopes. The problem is that arrays
may not have discriminants. Therefore you cannot constrain an array
otherwise than in its bounds. If arrays had discriminants, OP could achieve
his goal:

   type Not_A_Ragged_Array (Length : Natural) is
       array (Positive range <>) of String (1..Length);
   type My_Container (Counter : Natural, Length : Natural) is record
       Data : Not_A_Ragged_Array (Length, 1..Counter);
   end record;

[ I wonder that OP complaints about features (constrained types) of which
other languages lack slightest traces. ]

> procedure P(Simple : in out My_Container) is
> 
>    -- Problem: what's the address of Simple.Data(N)
>    -- in case there are no dope vectors?

Maybe Simple.Data(N) does not have a unique address at all. Why should you
care how does the compiler pack arrays? 

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de