Seam 3 space on JBoss Community
by Dan Allen
Two forces are converging that cause us to rethink where we are hosting Seam
3 project documentation and discussions. I'll enunciate those forces and
propose a solution that Lincoln and I have devised.
Seam 3 is a very different beast than Seam 2. In some ways, I liken it to
the relationship between Maven 2 and Maven 1. Both the Maven projects are
build tools that are founded on the same principle, a software project
management and comprehension tool that leverages an extensive ecosystem of
plugins that tie into a unified build lifecycle. But Maven 2 is built on a
different foundation than Maven 1, making them, in the end, very different
solutions.
Seam 3 relinquishes the context and dependency management and event
notification responsibility to CDI, something that evolved out of the Seam 2
core. Seam 3 is also focused on extensibility punctuated by container
portability. While you can expect similar convenient APIs and integrations
in Seam 3, and a bridge module that allows you to use or emulate Seam 2
components, pragmatically speaking, it's new project that is going to be
organized and governed very differently.
Having established that Seam 3 differs from Seam 2 is several fundamental
ways, mixing Seam 2 and Seam 3 information is massively confusing for
developers. I'm referring specifically to the forums and the wiki. The
driving force here is to establish a new "space" for Seam 3. Let's set that
point on the table and move to the next force.
Seam and Weld are projects brought to you by JBoss and its valued community.
Yet, for some time now, these projects have isolated themselves from the
JBoss community by being hosted outside (seamframework.org) of the JBoss
community umbrella (http://community.jboss.org). The creator of the
seamframework.org software is no longer with the team and the rest of us
must assume the responsibility of keeping it up to date. The problem is, any
work we spend on seamframework.org takes away from our ability to improve
Seam and Weld, review patches, analyze issues or write documentation. It's
an unnecessary drain. What makes it even more ridiculous is that we have a
entire team, the JBoss Community infrastructure team, that works full time
on providing a collaboration site to host projects, discussions and
documentation. A resource we are not currently leveraging.
Although it's a bigger discussion, we are reconsidering the use of the JBoss
Community for technical reasons. Lately we've identified some pain points
with seamframework.org that will immediately be solved by using Jive SBS
(the Java-based software behind community.jboss.org).
- Inability to monitor community discussion via "subscribe to this
thread"
- Inability to monitor comments on wiki pages
- Inability to post deep, complex links - including links to our own
threads on swfk.org
- A rigid navigation structure that cannot easily be altered
In short, we are concerned that we are lacking proper tools to effectively
manage and grow the Seam and Weld projects and community.
As simple as it sounds, there are still some considerations to take into
account before deciding to return Seam and Weld home to the JBoss Community.
That migration will require more planning. However, what Lincoln and I
suggest we can do as a first step is to take advantage of the need to create
a new space for Seam 3 and initiate that space in the JBoss Community. In
addition to the technical benefits, this would give us:
- a clean break for Seam 3, avoiding muddling with Seam 2 discussions and
information
- a proving ground to decide how we like the JBoss Community
collaboration platform, considering a full migration of
seamframework.orgin the future
- a chance for Seam community members across the board to establish their
JBoss Community accounts
- a better experience, in our mind
We propose:
- Creating a Seam 3 space at http://community.jboss.org with a user
discussion forum and wiki
- Creating a development subspace with a wiki (no discussion forum)
In the short term, we'll still use the seam-dev mailinglist for development
discussions (though we can consider enabling the discussion forums in the
development subspace if the e-mail bridge can be activated).
Would the majority of you be agreeable to this approach? It really will
provide a lot more clarity for Seam 3 and will show that we are sincere
about reuniting with the JBoss Community. Understand that this move will not
take away from the focus on portability in any way.
-Dan
p.s. Take a look at http://hibernate.org
--
Dan Allen
Senior Software Engineer, Red Hat | Author of Seam in Action
Registered Linux User #231597
http://mojavelinux.com
http://mojavelinux.com/seaminaction
http://www.google.com/profiles/dan.j.allen
--
Dan Allen
Senior Software Engineer, Red Hat | Author of Seam in Action
Registered Linux User #231597
http://mojavelinux.com
http://mojavelinux.com/seaminaction
http://www.google.com/profiles/dan.j.allen
14 years, 5 months
- 19
- 60
Seam 2 race conditions
by Ondřej Skutka
Hi!
JBSEAM-2419 (IAE when concurrently called a factory) had been solved by
introducing a lock on the getInstanceFromFactory method. It turned out
to cause deadlocks under some circumstances (JBSEAM-4669). So I've moved
the lock deeper into the method [1]. Could someone review this change
and approve it?
It would be also nice if there was a lock per every factory. It might
have some performance benefits. Or disadvantages. Profiling should give
the answer.
Thanks, Ondra
[1] https://jira.jboss.org/secure/attachment/12335273/JBSEAM-2419.patch
14 years, 6 months
- 1
- 0