On 28/02/2008, at 1:31 AM, Mike Schrag wrote:
>> Unless I'm remembering incorrectly, WOBuilder only showed you the
>> keys that were in the model, not every possible key available to
>> you in the class. This is a huge step forward!
> I'm not sure what heuristic they use to determine this ...
>
> * At the top level, I would think you'd have to show all of the
> possible bindings on the component (sans "system" ones, which we
> remove now, too). It looks like WOB puts session and application up
> at the top of the list, which sort of makes sense
>
> * I think WOB also split out actions from regular bindings, which is
> actually a little bit tricky. I'm not sure how they determine which
> ones are actions compared to just regular "get" methods. No way
> they looked at the inheritance hierarchy to only show methods that
> extend from WOActionResults
> (and even though we can do that easily, it's really expensive to
> actually do that). I suppose it could look for methods that don't
> take params and either return void, or return non-void but don't
> have a correspondingly named setXxx method, to disambiguate it from
> a regular getXxx method (or xxx() method)? If someone wants to run
> some experiments on WOB, that might be interesting.
Done. WOB appears to look for zero arg methods with a return type of
WOComponent (or a subclass). Although I think this might be better
done as zero arg methods with a return types that implements
WOActionResults because that is what the runtime is really expecting.
-- Seeya...QQuinton Dolan - qdola..mail.com Gold Coast, QLD, Australia (GMT+10) Ph: +61 419 729 806
This archive was generated by hypermail 2.0.0 : Wed Feb 27 2008 - 19:22:58 EST