Re: C/C++ programmer giving Ada95 a chance -- writing an emulator.

[tmoran@bix.com]

>And the conversion to Op_Root'Class to force dispatching
>seems reasonable to me (you can find this on p546 of Cohen).
  On p546 Cohen says:

> Given (the declaration File_Name_Copy : Heirarchical_File_Name;)
> if we were to replace the statement
>
> This_Delimiter := Delimiter(File_Name'Class(File_Name));
>   (he presumably meant to say:
> This_Delimiter := Delimiter(Heirarchical_File_Name_Type'Class(File_Name));
>   )
>
>    with
>
> File_Name_Copy := File_Name;
> This_Delimiter := Delimiter(Heirarchical_File_Name_Type'Class(File_Name_Copy));
>
> the call on Delimiter would no longer redispatch based on the
> tag of File_Name.  The assignment to File_Name_Copy throws away
> the tag of File_Name, and the call on Delimiter always dispatches
> to the Heirarchical_File_Name_Type version of the function.

I believe that your assignment

>    Op_Table : array (Opcode) of Op_Root
>     := (Op_Root (O0),
>         Op_Root (O1),
>        etc
likewise throws away the tags of the things copied into spots in
Op_Table, and the call on
>       Func (Op_Root'Class (Op_Table (IR)));
always dispatches to the Op_Root version of the function.
  I think we have here both an error in Gnat and an error in
Cohen's book!

> It is also obvious my code wouldn't work if the derived types
> added components.
Exactly.