[seam-dev] Seam 3/Weld extensions planning notes
gavin.king at gmail.com
Sat Nov 21 14:06:06 EST 2009
On Sat, Nov 21, 2009 at 4:17 AM, Pete Muir <pmuir at redhat.com> wrote:
> 2) support for other languages such as Groovy or Scala (a future direction, looking for a lead per language)
> * in general, this more writing tests and examples, than specific integration code
This should be a low priority. We never got much adoption of that stuff in Seam.
> 4) Support for replicating managed entities
> * this is something we supported in Seam 2, but it caused problems. I have previously
> spoken to the JBoss Clustering team about this, and they plan to add this support to JBoss
> AS. IMO we should use this support, and not try to implement it ourselves, so we should
> check with them to find out the status.
Yes, the only way to really make this work properly is to fix it at
the container level.
> 7) Hot-deploy support (Pete)
> * this is something we supported in Seam 2
> * splits into two parts:
> - redeploy of underlying class definitions
> - rebuild of associated bean and class metadata
> * IMO the class redeploy should be dealt with by the container, not Seam/Weld, but this
> may not happen quickly
We can probably support the same thing as Seam2, i.e. managed beans in
a special deployment location loaded on Weld's own classloader.
Probably worth doing, since Seam users seem to like this. However, the
real fix has to happen at the container level and for that we need
buy-in from the AS6 and EJB teams. So far this has been very, very
> 8) JAX-RS integration (Jozef? and Christian?)
This is something we should try and prioritize.
> 9) Resource module for Seam (?)
> * DocumentStore
> * Classloading stuff from Seam
> * I would prefer to not have this if at all possible
Agreed. JSF2 now has some stuff for this, right? I would definitely
like to search for alternative options.
I think we should try and toss all the stuff we have in Seam2 for
loading "mini" filters and servlets. The servlet spec now has
annotations and a componentized web.xml.
> 10) I8ln support (?)
This has medium priority. Probably need to properly review the design
of all that stuff.
> 11) Servlet support (as a CDI PE?) (Nik)
What does this mean? Conversations?
> 12) Security (Shane)
> * Also going to get Shane together for a couple of days with Anil soon so that they can
> work out how we can better reuse the support JBoss Security provides (e.g. using their
> SSO, NTLM, Kerberos etc. support)
The security stuff is one of the *most* important features of Seam,
but needs to go through a review/redesign. We need to:
* emphasize use of user-defined annotations representing permissions (CDI style)
* ditch use of EL in Java code
* deemphasize drools
* use what the servlet 6 spec now has for authentication (you can call a method)
And we should really look for functionality to "trim" from this stuff.
> 15) Pageflow (?)
This should be for backward compatibility with Seam2 only. I want to
design a typesafe Java-based solution.
> 18) Seam application framework (Dan/Jozef)
Leave this to me. I now have some different ideas that I need to try out.
Make sure that this is a very generic solution for CDI. i.e. it's a
CDI bean remoting layer.
gavin.king at gmail.com
More information about the seam-dev