Thanks Eric, but please make this available as alternate implementation only. Because I honestly think the BDA definition is utterly broken in the spec (*)! It makes assumptions about classloading mechanisms which are _most_ times true for _old_ EE containers but is a) NOT defined in the EE spec and b) is NOT true for modern EE containers anymore (which heavily use OSGi under the hood).
I honestly believe that the mechanisms we have currently does suite professional needs much better than this overly strict BDA stuff does. And we _still_ pass the TCK, so there is imo nothing to worry yet! LieGrue, strub (*) marking a JAR (if available) as containing jars is really fine. And with Pete having added CDATA sections to the schema recently, we could now also add our own namespaced exclude and include rules to beans.xml for speeding up the classpath scanning. This is really neat. BUT there are a few heavily broken thinkgs in the BDA part of the spec. A few examples: .) having to write <alternatives> into EACH BDA where it should be active is just idiotic .) same for <interceptors> .) same for <decorators> ----- Original Message ---- > From: Eric Covener <[email protected]> > To: [email protected] > Sent: Tue, August 10, 2010 3:56:29 PM > Subject: Re: [DISCUSS] releasing owb-1.0.0 next week? > > On Tue, Aug 10, 2010 at 3:37 AM, Mark Struberg <[email protected]> wrote: > > Hi folks! > > > > Gurkan and I had a small chat yesterday about the stability of > > OpenWebBeans. >And > > we both agree that it really would deserve a 1.0.0 number because it's >pretty > > stable. At least a lot more stable than many commercial products with a 1.0 > > release number ;) > > +1, I hope to work on BDA awareness in the scanner service / > {Bean,Alternatives,Interceptor}Manager stuff and I don't want to > derail 1.0 with it! > > Would we branch at 1.0, and would 1.0 still be CTR? > > > -- > Eric Covener > [email protected] >
