> But back to the extending singleton problem. I don't see a clear solution
> for extending singletons (have two instances of a singleton is really no
> solution!). So I think we should keep the current solution.

I'm not sure about this. The issue here is to *derive* from a singleton
on the class level, which seems like a valid idea to me. If you want a
refined child class, it shouldn't matter that the parent is a singleton
class. It might be sensible to demand that the child class is also a
singleton. But what you are saying is that singleton classes are
effectively final.

The issue you are actually raising is that of run time  *instances* of
such a derived class. That's a different problem. Maybe that could be
dealt with in the qooxdoo run time, e.g. by checking for another
instance polymorphically (i.e. along the class hierarchy) when the
constructor of a singleton class is invoked. That way only a single
instance of that *hierarchy* can be instantiated. To do that the object
registry could keep a list of instantiated singeltons (maybe it does
that already).

T.

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
qooxdoo-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel

Reply via email to