On 05/12/2014 18:42, Mark Thomas wrote: > On 04/12/2014 14:21, Rémy Maucherat wrote: >> 2014-12-04 15:01 GMT+01:00 Mark Thomas <ma...@apache.org>: >> >>>> Ah actually, it is a mix then of the two, it should implement what is >>>> marked as WSC-4.8.1 in the spec instead. >>> >>> I thought that was what was implemented (prior to r1643008). What did I >>> get wrong? >>> >> The test is fairly explicit that the annotated methods from superclasses >> should be used. And that's what the spec is really saying actually, but >> it's nearly impossible to understand IMO. > > Ah. Now I get it. > >> It mostly defines some exceptions >> to that inheritance: >> - if the subclass does not have its @*Endpoint annotation (fine) [spec >> quote: For example, a Java class >> that inherits from a Java class annotated with class level WebSocket >> annotations does not itself become an >> annotated endpoint, unless it itself is annotated with a class level >> WebSocket annotation.] >> - if an annotated method is overridden by a non annotated method (harder) >> >> It starts with a confusing statement [The websocket annotation behaviors >> defined by this specification are not passed down the Java class >> inheritance hierarchy. They apply only to the Java class on which they are >> marked.]. So the method level annotations would still be there in the >> superclass and used unless the subclass overrides them. > > This is the point I missed in my original meaning. I agree with your > interpretation. > >> So I tried to do that now, but my commit (r1643065) is lost somewhere. > > That should have come through by now so I'll go and have a look.
Found it. Looks good. I made one minor tweak but no functional change. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
