Re: Type in allocator has deeper level than designated class-wide type

[Adam Beneschan <adam@irvine.com>]

On Jun 13, 5:49 pm, Adrian Hoe <aby...@gmail.com> wrote:
> On Jun 14, 1:11 am, Adam Beneschan <a...@irvine.com> wrote:
>
>
>
>
>
> > On Jun 13, 8:31 am, Adrian Hoe <aby...@gmail.com> wrote:
>
> > > Hi all,
>
> > > I have stomped into above error and I am stuck with the error and
> > > runtime exception PROGRAM_ERROR : accessibility check failed.
>
> > > Here's the fragment of my codes:
>
> > > generic
> > >    N : Positive;
> > >    type Item is private;
> > > package High_Cloud is
>
> > >    package List renames Heterogeneous_Doubly_Link_List;
> > >    use List;
>
> > >    type Data_Array               is array ( 1 .. N ) of Item;
>
> > >    type Element_Record           is new List.Node with private;
> > >    type Series_Access            is new List.Access_List;
>
> > > private
>
> > >    type Element_Record is new List.Node with
> > >       record
> > >          Name           : Unbounded_String;
> > >          Color          : Unbounded_String;
> > >          Data           : Data_Array;
> > >       end record;
>
> > > ...
> > > end High_Cloud;
>
> > > In the package body of High_Cloud, I have a procedure:
>
> > >    procedure Add
> > >       ( Cloud          : in out High_Cloud_Record;
> > >         Name            : in     Unbounded_String;
> > >         Color           : in     Unbounded_String;
> > >         Data            : in     Data_Array )
> > >    is
>
> > >       New_Element : List.Access_Node;
>
> > >    begin
>
> > >       New_Element := new Element_Record'
> > > <========== Error/Warning here.
> > >                         ( Constructor.Init ( Name, Color, Data) ) ;
>
> > >       Concatenate ( Cloud.Series, New_Element ) ;
>
> > >    end Add;
>
> > > My intention is to use genericity to create a data which is an array
> > > (1..N) of any_type. This data is then stored in a list. Procedure Add
> > > instantiate (by calling to Constructor.Init) the list's node with
> > > data.
>
> > > I have read the ARM 3.10.2 and 3.10.1 and also John Barnes book yet I
> > > still have no clue how to solve the problem.
>
> > > Can you please help? Thanks.
>
> > I think we need to see the definition of Constructor.Init, at least.
> > (I'm guessing that Heterogeneous_Doubly_Link_List is a library package
> > that defines Node as a tagged record and Access_Node as "access all
> > Node'Class".  But I don't have a guess about what Init is.)
>
> >                                -- Adam
>
> Yes, you are right. In Heterogeneous_Doubly_Link_List:
>
>    type List        is limited private;
>    type Access_List is access List;
>
>    type Node           is tagged private;
>    type Access_Node   is access all Node'Class;
>
> private
>
>    type Node is tagged record
>       Previous : Access_Node;
>       Next     : Access_Node;
>    end record;
>
>    type List is limited record
>       Head    : Access_Node := null;
>       Tail    : Access_Node := null;
>       Current : Access_Node := null;
>       Count   : Unsigned := 0;
>    end record;
>
> And:
>
>    package body Constructor is
>
>       function Init
>          ( New_Name  : Unbounded_String;
>            New_Color : Unbounded_String;
>            New_Data  : Data_Array )
>       return Element_Record is
>       begin
>          return ( List.Constructor.Init
>                      with
>                         Name  => New_Name,
>                         Color => New_Color,
>                         Data  => New_Data
>                 ) ;
>       end Init;
>
>    end Constructor;
>
> Is the problem lying in the Data : Data_Array in Element_Record or in
> the return statement of Constructor.Init?

Where is the above Constructor package defined?  Is it directly within
High_Cloud, or is it defined inside a subprogram in the body of
High_Cloud?

Assuming it's the former, I can't get either our compiler or GNAT
(4.5.2) to generate an error or warning message.  However, I don't
know if any relevant rules about accessibility have changed in Ada
2012.  I'd have to study this one further.  If you're not using our
compiler or GNAT, and it's an Ada 2005 compiler, it could well be an
error in your compiler.

                               -- Adam