[Kstars-devel] branches/work/kdeedu_kstars_htm/kstars/kstars/skycomponents

James Bowlin bowlin at mindspring.com
Thu Aug 9 16:31:22 CEST 2007


On Thu August 9 2007 7:52 am, Thomas Kabelmann wrote:
> The first proposal is pure theory and doesn't scale well with a
> growing number of methods. As I didn't looked at the code which
> components use this method, I just suggested to define this method
> for all components. If this method is just used by SkyMapComposite
> than your proposal is the better way as we don't pollute the main
> interface. If in the future other components, not derived from
> SkyMapComposite, but not all components need this functionality, we
> should define a new interface which the  corresponding components
> have to implement. Drawback is then that we have check the type at
> runtime, but better than casting of good luck. So currently I prefer
> your solution too.

I agree with this and will do it this way.  All of those casts will
go away.  I will change the signature when it seems safe the location
in the c/c hierarchy won't change and if we need the same functionality
elsewhere I will add a virtual function (most likely in SkyComponent).


-- 
Peace, James

There are many ways to be free. One of them is to transcend reality by
imagination, as I try to do. 
-- Anais Nin


More information about the Kstars-devel mailing list