when thinking of it. I also don't like removing the methods.
Because then the learning effect is greater. Because just looking at the
IModel interface
you notice detach methods and so on. If you have just annotations you have
no idea..

johan


On 11/2/06, Eelco Hillenius <[EMAIL PROTECTED]> wrote:

Sounds good. I don't think it makes the contract fuzzy if we support
both an annotation and detach() btw. I would see the annotation as
optional. In fact, I wouldn't be against annotations being optional
anywhere (saves some more rewriting for WIA too).

Eelco


On 11/2/06, Igor Vaynberg <[EMAIL PROTECTED]> wrote:
> should we extend the support of these annotations to behaviors and
models?
>
> behaviors can benefit from all 4, while models can benefit from
> attach/detach ones
>
> im not sure where imodel.detach() would fit then. i would love to keep
it
> since it makes the contract clear, but it is a bit weird if we support
both.
>
> -igor
>
>
> On 10/30/06, Johan Compagner <[EMAIL PROTECTED]> wrote:
> >
> > On 10/30/06, Igor Vaynberg <[EMAIL PROTECTED]> wrote:
> > >
> > > On 10/29/06, Johan Compagner <[EMAIL PROTECTED]> wrote:
> > > >
> > > > yes private or package scope methods should always be called
> > >
> > >
> > > not quiet, package scoped methods can still have overrides, see
> > > ReflectionUtils.overrides(Method,Method) for some nasty code :)
> >
> >
> >
> > ahh you are right.
> > I don't use package scope things a lot but
> >
> > Package y
> > class A { void test(){}}
> > class B extends A { int test(){return 1;}}
> >
> > fails. (java 4 compile covariant not supported)
> >
> > but move B to another package:
> > Package z
> > class B extends A { int test(){return 1;}}
> >
> > works..
> > And then ofcourse the A.test() is nowhere to be found..
> >
> > so yes if in the same package this works:
> > void test()
> >     {
> >         super.test();
> >         return;
> >     }
> >
> > Then it is really as it is protected (instead of what you also could
thing
> > private)
> >
> >
> >
> >
> > >
> > > the initial implementation of this is in. it only works on Component
so
> > > far,
> > > but there is soom room for improvement i think.
> > >
> > > for example both requestcycle and session have detach() but no
attach(),
> > > so
> > > we can extend annotation support to those as well.
> >
> >
> > RequestCycle doesnt have a detach() (only a private one that is our
one
> > cleanup method)
> > RC has onBeginRequest and onEndRequest.
> >
> > Session does have detach. But WebSession also have attach.. Why that
is i
> > don't know....
> >
> >
> > johanm
> >
> >
>
>

Reply via email to