On Nov 21, 2009, at 2:06 PM, Gonzalo Aguilar Delgado wrote:
Hi all,
I realized that current UI implementation performs well but is not
completely async.
Let me explain if I I'm right.
We have two execution modes, ui and desktop.
UI
System will render everything as usual but custom YUI code will
be executed
to allow move, resize, place, etc. Windows.
The portlet creation is created in the jetspeed pipeline in
sequential mode like
in the static mode.
Client will hold until everything is created and rendered.
This is current and default render mode.
Desktop
Only basic html is sent to the client.
Ajax Toolkit initialization is sent and everything is handled
with javascript.
Client will start to see contents as soon as they are loaded.
This is obsolete and posibily will not work with new dojo
versiopns.
I suppose that best world is a mix of both were portlets are rendered
empty of content via
a <div> tag and ajax toolkit will take care of move, resize, place,
etc... And load contents
asyncronously...
This way page will get at least portlet configuration if something
goes
wrong with ajax.
What do you think?
This is exactly what we are working on here (see JIRA issues). There
are two modes of the /ui pipeline: server-side rendering and client-
side rendering, the equivalent of the old /portal and /desktop
pipelines - but both sharing the same core engine (unlike in 2.1.x). I
am in the process of having a YUI expert review my architecture and
prototype this week - after that - we will formerly begin developing
the /ui code to have both client side and server side rendering all in
and working by end of year. I will get the roadmap updated so everyone
has a better idea where to contribute
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
