Event binding improvements

[cervengoc]

This PR contains several minor improvements related to the event and submit bindings.

• Refactored the submit binding to reuse the event binding. This way some concepts can be easily shared, like preventing default behavior or stopping bubbling.
• Added a check in the bound event handler to throw an exception in case of the bound value is not a function. This adds some level of additional usability and makes debugging strange issues easier.
• Added support for stopping event propagation from the handler by returning an explicit false value. This behavior is similar to how jQuery handles return values from event handlers and it adds support for run-time event propagation handling (in contrast to the existing "design-time" *Bubble feature).

The latter point introduces somewhat a breaking change, but I don't think that many code is affected as returning false explicitly to prevent default behavior and stop propagation is an old convention in my opinion.

Please let me know if we should refine anything here.

[cervengoc]

For some reasons phantomjs tests seem to fail, I'll have a deeper look at it in the days.

[cervengoc]

@mbest I've double-checked several times the code and cannot see why it doesn't work in phantomjs tests. Could you please have a look at it?

Also, for some reason I can't run the phantomjs tests locally after a clean npm install, it just doesn't seem to do anything.

[codymullins]

you could just do typeof handlerFunction !== function, the null and undefined check seems redundant in my opinion, shouldn't those be covered by the !== function check?

[cervengoc]

No, in this case this condition is stricter and that was intentional. I wanted to allow null or undefined values as those can be kind of valid values in some cases.

[codymullins]

Ah, good point. I was under the assumption we didn't want null or undefined here. Thanks for the explanation.

[cervengoc]

@mbest Would you please share your concerns or thoughts on this one if you have any?