Tapestry 4.0 (final) has been released. This is a great relief to me, because I really think I could explode, or at least spontaneously combust, if this didn't happen.
I'm exceptionally proud of this release; to some degree, we've rewritten Tapestry from the inside out. The new structure, based on HiveMind, makes Tapestry exceptionally extensible, and performance appears to be better than Tapestry 3.0 (so much less reflection going on). More importantly, the coding model has improved quite a bit. Component parameters now just work ... no fumbling with "direction", listener methods are really flexible and easy to use, the validation framework is much, much simpler, and the annotation support (still optional in 4.0) is very succinct and code-focused.
We're already doing some planning for Tapestry 4.1 which better not take two years ... I'm hoping more like 6 - 9 months. I have several tasks on my plate:
- A fast, useable, productized integration test suite (that can be used to test applications)
- Convert to a Maven 2 build
- Rework form support to eliminate the rewind phase
- Remove the need for pages and components to extend from Tapestry base classes
Meanwhile, time to sit back and see what people make of this new release.