Normally I would choose performance over @nearly_all_else, but this time I'm deferring the new feature until the POE::XS project is done.
Invoke will make local-session calls about 2-3 times faster in the current code base. The gains may not be as significant when large parts of the back-end classes are written in C. If that's the case, the entire purpose for invoke() becomes moot. If invoke() is still significantly faster, waiting to include it means we've had more time to think of alternative names. "It sucks the least" isn't a positive recommendation for "invoke". -- Rocco Caputo - [EMAIL PROTECTED] - http://poe.perl.org/