On Wednesday 23 September 2009 15:13:47 Duncan Mac-Vicar Prett wrote: > On Tuesday 22 September 2009 14:41:43 Martin Kudlvasr wrote: > > This work was done to show, that there are other (and many think better) > > approaches to the problem than dynamic creation of classes in > > service_resource.rb. So I especially ask Duncan and Klaas, what other > > problems service_resource solves and YastResource does not. > > > > Please let me know what you think. > > I think this is a great step forward.
Let's not forget kudos to Martin Vidner and Karel. > > The argument of being able to add code to the client proxy that otherwise > would go in the model is enough to convince for the benefits. > > Only the name of the class does not convince me, may be the name > YaST:ServiceResource::Base can be reused (there is already such class > there) The YastResource name was chosen quickly as we knew it was going to be only temporary. I personally have no preferred name. If we are going to plan a transition, we should keep 2 names for some time. Both service_resource.rb and YastResource can be used by plugins at the same time. This way we can achieve a fluent transition. > > Also the initialization that needs to be done looks weird, but I can't > think on something to avoid it. > Well, I combined the 2 steps in "init_interface_and_find" method which combines the 2 lines and does exactly what "load_proxy" does - init the class and call find. As Josef has pointed out the initialization part can be placed in a before filter, which would have to be different for each plugin/model. -- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
