On Sun, May 24, 2009 at 1:21 PM, Dan Allen <dan.j.allen@gmail.com> wrote:
On Fri, May 22, 2009 at 5:50 PM, Clint Popetz <cpopetz@gmail.com> wrote:

Is there any reason seam-bridge-api and seam-webbeans-bridge shouldn't move to
webbeans-extensions?  I have this jira issue:


to allow the webbeans wicket support to work with any jsr-299 impl,
and part of making that work is providing a way for the wicket
extension to grab the current manager instance.  It seems like this
is useful independent of seam.

I think we are engaging in a struggle over ownership of features and defining boundaries between Seam and WB.

True, it seems like an arbitrary boundary.

In my definition, Seam is the universal middleware stack at Red Hat which is most definitely independent of JSR-299 implementation. Therefore, it's my position that anything that "extends" 299 towards "rich web" should be a module of Seam. Extensions which don't go in that direction, such as the Java SE support, are a better fit as a webbeans extension. Furthermore, I don't think you should feel like you "have to depend on Seam" because it isn't going to be like Seam 2 where it is all or nothing. Modules are starkely independent and can be seen as a lose collection of extensions.

Perhaps others will disagree with my definition, or think it too liberal, but if we don't define where Seam begins, then you could easily rationalize all modules of Seam be extensions of Web Beans...I will admit though that the bridge is a gray area in that it could go either way.

This issue has actually gone away for my own use-case, because the things wicket-webbeans needs are now part of the new SPI.