Re: Who is wrong, me or GNAT? :-)

[Ed Falis <efalis@mediaone.net>]

Mark Lundquist wrote:

> OK, here's the code:
> 
> --------------------------------------------
> -- foo.adb
> --------------------------------------------
> 
>  procedure Foo is
> 
>   package P is
>    type T is abstract tagged null record;
>    function Create return T is abstract;
>   end P;
> 
> 
>   package P1 is
>    type T2 is new P.T with private;
>   private
>    type T1 is new P.T with null record;
>    function Create return T1;
> 
>    type T2 is new T1 with null record;
>    --
>    -- Gnat takes error on this ^^^ line.
>    --
>    -- But T2 should inherit 'Create' from
>    -- (concrete) type T1.
> 
>   end P1;
> 
>   package body P1 is
>    function Create return T1 is
>    begin
>     return (null record);
>    end Create;
>   end P1;
> 
>  begin
>   null;
>  end Foo;
> ----------------------------------------
> 
> GNAT 3.13(p) says:
> foo.adb:16:30: type must be declared abstract or "Create" overridden
> 
> Is this a bug in GNAT?  If not, what am I missing?
> 
> I was trying to come up with an idiom for inherited constructors for
> descendants of an abstract tagged type, and this seemed like it ought to do
> the trick (although it suffers from the fact that the intermediate derived
> type, e.g. T1 in this case, would have to override all the abstract
> primitives).
> 
> Any help appreciated...
> thx
> -- mark

RM 3.9.3 (4-6)

Functions with a controlling result must be explicitly overridden when 
the parent type is derived (or the derived type has to be declared 
abstract), as they're considered abstract in the derived type.