Wonderful news...

I will rework the javascript ajax apis this week on thursday.
Generally there is not too much work on the api side to do
a helper function has been added.


Werner


Simon Lessard schrieb:
Hi all,

This is a simple post to inform you that the final JSF 2.0 draft was released yesterday (http://jcp.org/aboutJava/communityprocess/pfd/jsr314/index.html) and that the 2.0 branch is in sync with it (at least to my knowledge). Not everything is integrated of course, but all API classes and methods should be there (except enums for component attributes, I hope we can "pluginize" those). If you check the code you can find different kinds of TODO:

// TODO: IMPLEMENT HERE
Means the logic should be coded inline where the comment is. Can be found in both API and IMPL.

// TODO: IMPLEMENT IMPL
Means the logic should be implemented in myfaces-impl while the api side most likely throws an UnsupportedOperationException. Can be found in API only.

// TODO: PLUGINIZE
Means the class might be generable by maven plugin, but adding the metadta/code for it to work as yet to be done. This is mainly for the new enums for component attributes, the TODO is in the UIComponent class.

// TODO: VALIDATE
Means the code should be reviewed against the spec or the use case should be thought about to determine if the algorithm used is correct. Can be found in both API and IMPL.

// TODO: DEFINE
Means the spec is silent on the algorithm or the constant values. Found in API, currently only in ExceptionQueuedEventContext I think. I'll go to the EG with it.

// TODO: REPORT
Means that I must raise something I consider invalid as an issue to the EG. Found in both API and IMPL.

// TODO: REFACTOR
Means the code should be moved to a more logical spot. Found in Facelets mainly where I had to move some code around so that it would compile with the new API.

// TODO: PROFILE
Means I think an alternative algorithm could be faster, but think more thourough thinking should be put to it. Mostly found in Facelets and its extensive use of Arrays.binarySearch instead of faster but more memory consuming HashMap.


Help on any part is welcome, but what's left of error handling (validating where exceptions should be shallowed or rethrown and some missing ErrorHandler code) and component tree visitiing was reserved for Google SoC I think so let not deal with those for now.


Thanks,

~ Simon

Reply via email to