Re: Using Red-Black Trees

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

On Thu, 18 Nov 2010 07:27:20 +0100, J-P. Rosen wrote:

> Le 18/11/2010 00:39, Yannick Duch�ne (Hibou57) a �crit :
>> You can have only one at a time. If you select an implementation, you
>> select it system-wide, and you cannot select one for that usage and
>> another for another usage in the same system.
>> 
> If that's really what you want, there is an easy solution.
> Provide two packages with identical (or compatible enough for your
> purpose) specifications:
> 
> package Binary_Tree is....
> package Linked_List is....
> 
> Then declare at library level:
> with Binary_Tree;
> package My_Structure renames Binary_Tree;

Yes of course it is a simple, but non-Ada solution. BTW, I am using it for
years, with a small addition that the implementation is selected by the
gpr-file.

But it is not Ada it is pure C, because nothing is checked. An Ada way
would be abstract packages, i.e. a package that describes the public part
of some other package. The program could be developed in terms of the
abstract package(s) required to be substituted by an implementation
package. Upon substitution the actual package specification would be
checked. This is close to a formal generic packages but "typed." In effect
you have a package type and packages are values of.

Something of this sort could be done in existing Ada, but in a quite ugly
way:

generic
   type T is private;   -- The formal parameters play role
   with procedure F;  -- of a "specification" of Abstract_P type
package Abstract_P is
   -- Nothing here
end Abstract_P;

generic
   with package Concrete_P is new Abstract_P;
package My_Stuff is
   ...

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