Reply To: What happened to PrettyFaces?

Splash Forums PrettyFaces Users What happened to PrettyFaces? Reply To: What happened to PrettyFaces?



Thanks! Now I’m fully armed and ready to start the migration and update my utility libs, which are based upon old PrettyFaces project.

Nice to hear from you, Lincoln. I was busy too, lived between Moscow, Germany and US, fund raising and stuff like that.
Forget about rewrite-config-prettyfaces)) Thanks to Christian, I have solved all of the migration issues and now I prefer to stick to the Rewrite project rather than to compatibility library which is less likely to be supported as goog as your new “child”))

I’d like to raise the URLMappings topic again. Christian has offered the use of the ConfigurationProvider for JSF beans having the same business logic for different viewId’s (there is a discussion earlier in this thread). I feel uncomfortable with this idea, not because I will need to do more migration job, but rather because the logic should be splitted between the two absolutely different contexts – I wont have an access to the FacesContext, @EJB injection should be only singleton ones, no @ManagedProperty inside the ConfigurationProvider, etc. Now I understand that I will definitely create dozens of empty @Join-annotated sub-classes to get things working as they did with URLMappings. No problem, I can live with that, but I just do not understand why.

I had a quick look at the sources and it seems to me that there is nothing that could prevent you from supporting something like this:


That would possibly cause the @Rule name to be placed inside @Join (or created automatically) and some ambiguity checks in the Navigate and/or AnnotationJoinResourcePathResolver classes to be implemented, but that’d definitely give JSF part of the project more flexibility and saved it from the functional degrade which I think the one-to-one (class-to-page) approach definitely leads to.

Am I missing something here? What do both of you think about it? As an old and experienced PF-user I’d be glad to contribute if you think this idea is worth implementing.

Btw, there is a bug (I believe it is) in the initialization of the HttpConfigurationProvider – two entities of the class are created if there is an entry in META-INF/services/org.ocpsoft.rewrite.config.ConfigurationProvider – one from the services entry, another from some class-scanner I believe. Don’t think this is the expected behavior.