On Dec 30, 2008, at 1:35 PM, Gavin King wrote: > The spec requires that injection points be validated at init time for > any Bean (for XxxxxDependencyException). > > Nick from the RI team has noticed that for this to be implementable, > we need to add metadata about the injection points of a bean to the > Bean interface. It shouldn't be necessary, since "SimpleBean implements Bean" should be responsible for its own validation. It doesn't make sense for the Manager implementation to validate the Bean, since that breaks the encapsulation model of both Manager and Bean. It's cleaner for Manager to assume the Bean is valid.

On a related note, it would be cleaner if the Bean.create() method was changed to Bean.create(Context) in order to handle circular references. The logic of Bean.create(Context) would be 1. check if Bean is in Context (for circular ref), if so return the bean. 2. instantiate the Bean (do not inject/initialize) and put it in Context (so we have a ref) 3. inject/initialize the bean as usual (this may trigger the circular reference, which would be resolved by #1 on a second call).

Actually I think there is a better way to solve the problem, but I need to talk it over with Pete.

Just my comment: For adding the new API for this complicates the specification. I think, specification provides the necessary API related with Bean creation. As I said, these concerns are left to the implementations. Gurkan ________________________________ From: Gavin King <gavin(a)hibernate.org&gt; To: Scott Ferguson <ferg(a)caucho.com&gt; Cc: Michael Keith <michael.keith(a)oracle.com>; WebBeans <webbeans-dev(a)lists.jboss.org>; Java Community Process JSR #299 Expert List <JSR-299-EG(a)jcp.org&gt; Sent: Friday, January 2, 2009 7:19:20 AM Subject: Re: [webbeans-dev] Re: Bean.getInjectionPoints() On Fri, Jan 2, 2009 at 3:42 PM, Gavin King <gavin(a)hibernate.org&gt; wrote: > Actually I think there is a better way to solve the problem, but I > need to talk it over with Pete. The alternate solution is something like: interface CreationalContext { void add(Bean<X> bean, X instance); //registers an "incomplete" instance } interface Contextual<T> { public abstract T create(CreationalContext ctx); ... } interface Context { public <T> T get(Contextual<T> bean); //get any existing instance public <T> T get(Contextual<T> bean, CreationalContext ctx); //get or create an instance public Class<? extends Annotation> getScopeType(); boolean isActive(); } interface Manager { Object getInstanceToInject(CreationalContext ctx, InjectionPoint ij, Type type, Annotation... bindings); ... } And getInstanceToInject() would be smart enough to return any object that had already been registered with the CreationalContext. Note that this solution addresses the problem of circular references, and also fixes a hole in how the Manager knows about the InjectionPoint. -- Gavin King gavin.king(a)gmail.com http://in.relation.to/Bloggers/Gavin http://hibernate.org http://seamframework.org _______________________________________________ webbeans-dev mailing list webbeans-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/webbeans-dev _______________________________________________ webbeans-dev mailing list webbeans-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/webbeans-dev

But what you're really saying is that it's better to handle circular references from Bean.create() than from Manager.getInstance().

> That only works for Bean implementations provided in the > specification. It does not work for Bean implementations provided by > 3rd-party extensions. We need the dependency validation stuff to work > in the second case also. The validation would be a requirement on the extension, which should be reasonable since a new Bean type would be a relatively rare and sophisticated thing. If a 3rd party is capable of introspecting the injection points, it's fully capable of calling the Manager methods for validation.

Mostly, though, it's an encapsulation issue. The proposal exposes and constrains information that really should be internal to the Bean implementation just for the purpose of a relatively straightforward validation.

True, but I did like the old DependentScoped concept of the previous draft, and that kind of scope isn't available from the Manager. (I'm not exactly sure why the dependent concept changed between drafts. I assumed the new way was easier to define.)

> But what you're really saying is that it's better to handle circular > references from Bean.create() than from Manager.getInstance(). OK, > that's really interesting, let me chat to Pete Muir about that and get > back to you. > > ...The responsibility for associating the new object with the Context > should be in Bean.create() instead of in Context.get(). Right. It's easier to manage the circular problem there, or at least was easier for me.

Hi; Actually the spec requires to return Proxy objects from the Manager.getInstance(). So the dependency injection does not occur until Bean method is called, i.o.w Context.get() is not called before the any method calls on the Bean. So, circular reference may not be happened with Proxy usage. Gurkan ________________________________ From: Gavin King <gavin(a)hibernate.org&gt; To: Scott Ferguson <ferg(a)caucho.com&gt; Cc: Michael Keith <michael.keith(a)oracle.com>; WebBeans <webbeans-dev(a)lists.jboss.org>; Java Community Process JSR #299 Expert List <JSR-299-EG(a)jcp.org&gt; Sent: Friday, January 2, 2009 6:29:46 AM Subject: Re: [webbeans-dev] Re: Bean.getInjectionPoints() On Fri, Jan 2, 2009 at 2:37 PM, Scott Ferguson <ferg(a)caucho.com&gt; wrote: > > On Dec 30, 2008, at 1:35 PM, Gavin King wrote: > >> The spec requires that injection points be validated at init time for >> any Bean (for XxxxxDependencyException). >> >> Nick from the RI team has noticed that for this to be implementable, >> we need to add metadata about the injection points of a bean to the >> Bean interface. > > It shouldn't be necessary, since "SimpleBean implements Bean" should be > responsible for its own validation. It doesn't make sense for the Manager > implementation to validate the Bean, since that breaks the encapsulation > model of both Manager and Bean. It's cleaner for Manager to assume the > Bean > is valid. That only works for Bean implementations provided in the specification. It does not work for Bean implementations provided by 3rd-party extensions. We need the dependency validation stuff to work in the second case also. > On a related note, it would be cleaner if the Bean.create() method was > changed to Bean.create(Context) in order to handle circular references. > > The logic of Bean.create(Context) would be > 1. check if Bean is in Context (for circular ref), if so return the bean. > 2. instantiate the Bean (do not inject/initialize) and put it in Context > (so we have a ref) > 3. inject/initialize the bean as usual (this may trigger the circular > reference, which would be resolved by #1 on a second call). Bean already has a reference to the Manager and can call Manager.getContext(myScope). So the extra parameter would not be necessary. But what you're really saying is that it's better to handle circular references from Bean.create() than from Manager.getInstance(). OK, that's really interesting, let me chat to Pete Muir about that and get back to you. -- Gavin King gavin.king(a)gmail.com http://in.relation.to/Bloggers/Gavin http://hibernate.org http://seamframework.org _______________________________________________ webbeans-dev mailing list webbeans-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/webbeans-dev _______________________________________________ webbeans-dev mailing list webbeans-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/webbeans-dev