Last comment, this not including Ryan:
On 9/12/09 10:25 PM, Dan Allen wrote:
---------- Forwarded message ----------
From: *Ryan de Laplante* <ryan(a)ijws.com <mailto:ryan@ijws.com>>
I read about the new RichFaces 4.0 being built for JSF 2.0 here:
http://www.jroller.com/a4j/entry/richfaces_4_0_0_alpha1
They are still using AJAX4JSF instead of JSF 2.0's built-in AJAX library
because:
JSF 2 Ajax implementation still very limited and not contains our
autoupdatable zones, limitations for autoupdate, statuses, table
partial updates, server side processing options and so on. Queue in
JSF not configurable at all..
Some of these points could be solved at application level by using
some event's handling for example. But for some features - you will
even not have a point to attach the functionality to. If the
features that we provides with our ajax enchancements was
implemented in JSF - we would not started reimplementing the weel
for sure ;)
Table partial updates work AFAIK, thanks in large part to a bugfix
submission by ICEFaces (thanks guys). If they fail, it's a bug, and
needs to be reported - I can promise it'll be quickly fixed if at all
possible.
There's a queue in JSF 2 client, but it's a single FIFO stack. There's
good reasons not to like this (performance being the main one), but
other queue proposals (including the one by Alexander) had the major
disadvantage of leading to a non-deterministic call order, as well as
saturation of client connection resources - both of which would have
resulted in bugs that were almost impossible to catch by the end user,
as well as complicating the job of writing a comet overlay tremendously.
A real queuing system, as well as request aggregation of any kind, will
simply have to wait until we sort out how we want to handle Comet in JSF
in a compatible way - which is, I hope, on the agenda for 2.1.
Statuses were already commented on by Werner and I in a different part
of the thread.
As for auto-updating, perhaps that was part of the discussion before I
joined - but I wasn't a part of those discussions. Since I don't know
what they'd entail exactly, I can't comment on them.
Jim