Stuart,

We can probably easily identify cases through an initial pattern in the name on deserialization, and then just stick to simply the class name as the rest when that works and use the bean ID only in special cases where needed.  Sounds easy enough, but identifying the correct cases on serialization might be more involved.  I'll think about it some and might have some time in June to do it, if you don't.

- David

On Fri, May 27, 2011 at 12:53 AM, Stuart Douglas <stuart.w.douglas@gmail.com> wrote:
It is something I have been planing to do something about for a while now, but never quite got around to it. The original reason is that we needed a deterministic way of determining proxy names that will be the same accross JVM's, and using the class name will not always work, so we used bean ID's. It should be possible to simplify this for most cases.

Stuart

On Fri, May 27, 2011 at 5:21 AM, Dan Allen <dan.j.allen@gmail.com> wrote:
On Thu, May 26, 2011 at 02:31, Jaikiran Pai <jpai@redhat.com> wrote:
I am seeing many forum posts where, within the exception stacktrace, I
see really long classnames for proxies generated by Weld.Here's one
example http://community.jboss.org/message/604723#604723. Out of
curiosity, is there any reason why those names are so lengthy instead of
just generating the classnames like java.lang.reflect.Proxy does?

On top of the length annoyance, I think this is one of the reasons Weld doesn't work on the IBM JDK (or I'm mistaken an it's a Solder issue). But from my brief testing, it had something to do with generated class names. Just a heads up.

-Dan

--
Dan Allen
Principal Software Engineer, Red Hat | Author of Seam in Action
Registered Linux User #231597



_______________________________________________
weld-dev mailing list
weld-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev


_______________________________________________
weld-dev mailing list
weld-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev