Re: [REVIEW] Issue #66 - Possible optimization regarding execution of PhaseListeners

From: Ed Burns <Ed.Burns_at_Sun.COM>
Date: Thu, 11 Nov 2004 17:21:19 -0800

Thanks for doing this. I like the performance enhancement, but you need
to change one thing to conform to the new spec, which has been vetted by
EG discussion so I don't want to change it.

>>>>> On Mon, 08 Nov 2004 20:48:04 -0600, Jacob Hookom <> said:

JH> *Note* New implementation does scope exception catching on a
JH> per-PhaseListener where the previous implementation was scoped on
JH> before/after phase. I would like to propose not catching at all since
JH> the API does not specify any exceptions to be thrown on beforePhase or
JH> afterPhase. Also, all the tests ran successfully with the one change to
JH> TestLifeCycleImpl in leu of changes to exception scoping.

JH> Issue #66

JH> Performance Enhancements to LifeCycleImpl

JH> SECTION: Changes

JH> M src/com/sun/faces/lifecycle/
JH> - Used ArrayList cloning to prevent lengthy synchronization blocks.
JH> - Exceptions from PhaseListeners will not prevent other pertinent
JH> listeners from being fired.

JH> M test/com/sun/faces/lifecycle/
JH> - In leu of the changes to exception handling, all registered listeners
JH> will be called, even if 'b' throws an exception.

This test must remain unchanged since it accurately tests that the
implementation conforms to the spec. For example.

If there are three phaseListeners installed, A, B, C. If B throws an
exception on beforePhase, only A must be called on afterPhase.

Can you produce a version of your change that has the good performance
properties, but continues to allow TestLifeCycleImpl to execute



|  | {home: 407 294 2468, office: 408 884 9519 OR x31640}
| homepage:         |
| aim: edburns0sunw | iim:
To unsubscribe, e-mail:
For additional commands, e-mail: