On 3/9/06, Werner Punz <[EMAIL PROTECTED]> wrote: > Manfred Geiler schrieb: > > On 3/9/06, Werner Punz <[EMAIL PROTECTED]> wrote: > >> Mario Ivankovits schrieb: > >> > >>> And - IMHO the killer, the shared is not interoperable, means it cant > >>> put stuff in the request and use it again if called from another package > >>> as they are different classes - different package names. > >>> See my other thread regarding ExtensionsFilter and Dummyform. > >>> > >> Actually this is a serious showstopper, due to the fact that often some > >> kind of request singleton is absolutely needed. > > > > Why? Do we really need a singleton that is a shared between impl and > > tomahawk. How can we be sure that we do not have a runtime dependency > > between tomahawk and impl then? > > > Actually you have to think out of the box as well... > for instance a singleton between Tomahawk and ADF in the future would > break that way. Even if it is defined in the commons and no other > interference is there!
What do you mean by "a singleton between Tomahawk and ADF"? If both, Tomahawk and ADF, really need one common singleton during runtime, this functionality is the perfect candidate for the future commons project. And it's the only allowed place where something like that could live in the future! We do *not* want to have runtime dependency between Tomahawk and ADF, right? On the other hand a dependency like that is perfectly ok: * tomahawk depends on commons-jsf.jar * adffaces depends on commons-jsf.jar BTW, please do not muddle up shared and commons. There is no commons module right now. What we have is a *tool* project called "shared" that is based on the former commons classes and is meant to produce the two libs shared-impl and shared-tomahawk. We will have a *real* commons lib in the future (you know: with stable api, api/impl separation, etc.), but right now there is none. Manfred
