Preparing Mozart 1.4.0

Boriss Mejias boris.mejias at uclouvain.be
Sun Jun 29 22:52:31 CEST 2008


Raphael Collet wrote:
> Boriss Mejias wrote:
>>
>> Another clarification here. A while ago, when we decided that mozdss 
>> should go as a branch in the official Mozart repository, we discussed 
>> with Erik and Raph about the support to C++. We decided that it would 
>> be unnecessary work, so we dropped it. From that point, DSS evolved 
>> mainly thinking in terms of Mozart, and the support for C++ became 
>> obsolete.
>>
>> One of the big changes done in the implementation was the transition 
>> from "Mediator + Abstract Entity" to only an Abstract interface on 
>> DSS, with its implementation in the Glue layer of Mozart. This 
>> implementation is more efficient, but it request more work on the Glue 
>> layer.
> 
> Well, in fact it *simplified* the Glue layer. 

Sure. It simplifies the architecture of AE + Mediator by only having one 
entity now. What I mean is that before, part of the job was done by the AE in 
the DSS, and the other part by the Mediator in the Glue. Now, the AE does not 
exist in the DSS any more (it's only an interface), therefore, more work is 
done at the Glue layer.

cheers
Boriss

> The solution relies on 
> multiple inheritance in C++, with "virtual"s put at a few well-chosen 
> places.  I basically merged two layers of objects into one.  The merge 
> actually makes sense, but the class hierarchy required a bit of C++ 
> mastery to be made right.
> 
>> DSS remains as a language independent middleware for distribution in 
>> its design, but it would require non-negligible work to implement a 
>> Glue to be used with C++.
> 
> Cheers,
> raph
> _________________________________________________________________________________ 
> 
> mozart-hackers mailing list                           
> mozart-hackers at mozart-oz.org      
> http://www.mozart-oz.org/mailman/listinfo/mozart-hackers


More information about the mozart-hackers mailing list