beehive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael Merz" <>
Subject RE: endpoint interface
Date Tue, 28 Sep 2004 20:57:24 GMT
Hi Wolfgang,

Thanks for your latest contribution. For some reason, I did get your

I am currently looking at your code. One comment up front: the
WsmAnnotationProcessorFactory really wants to subscribe for
"javax.jws.WebService" only. All source file analysis is originated from
there. If we registered for javax.jws.* as you suggested, we'd be called
several times for the same file from different scopes, which we
currently don't want.

The rest looks good. I'll have another look and check it in when I am
done with it. Thanks again.



-----Original Message-----
From: [] 
Sent: Tuesday, September 28, 2004 9:37 AM
To: Beehive Developers
Cc: Jonathan Colwell; Michael Merz
Subject: Re: endpoint interface

Hi Jonathan and Michael.

I've attached 3 files ( a patch file (wsm_diff.txt),
and )

> If no endpointInterface is specified, things behave as they were and
> rest of the annotations will be loaded from the implemetation as

At the end of page 35 of JSR-181 states 
"If the JSR-181 service implementation bean does not reference a service
interface through the @WebService.endpointInterface annotation, the
JSR-181 processor 
is required to generate the interface from the annotations on the
implementation bean."

So my implementation of will generate a
service endpoint interface 
(actual java source file) and read annotations from the service
implementation bean.
Doesn't need to generate an actual java source file (java interface) ?

> if someone puts
> @WebService(endpointInterface = "") we try to
> load the source

It can load annotations from the source specified by the
However, It currently allows only the full path of the source file.

> if a class name is used we try to find a matching
> source file and if that fails, just load the class and use reflection
> create the object model.

It can load the class and use reflection to create the object model but
find a matching source file.
Implementing the functionality of finding a matching source file is
actually easy , but 
at least I gotta know the base directory where the search started from
and how deep 
the search is allowed to dig into. 
Otherwise, It's gonna take a long time to find out the search succeeded
or actually failed.
So I left the part.

I will refactor to take those options ( the
base directory , 
depth of directory the search is allowed to dig into ) later.

Thanks in advance.

why is the attachments from BEA guys allowed, but me ( mail)

View raw message