On Thursday, April 1, 2004, at 12:34 PM, <[EMAIL PROTECTED]> wrote:
I just noticed that a portlet render method is called twice as result of an
ActionRequest.
That shouldn't happen.
Im in the middle of refactoring and componentizing the navigation and URL code, so its very likely things will break.
Im putting in unit tests around all the components
Very much appreciate your help here, but please bear with me for a few days while I transition the code to a more manageable state
and decouple the navigational state from Pluto's implementation.
Furthermore, the RenderRequest contains all the ActionRequest parameters as
well (which is a violation of the specs).
Yes I know about that.
Log it as a bug if you'd like.
I just opened it up so that post parameters made it back and so portlet developers could write portlets that use form processing now
Debugging isn't so easy as the render events happen in seperate o.a.j.aggregator.impl.Worker threads.
Have you tried debugging with Eclipse 3.0? It handles multithreading nicely.
Or try changing the aggregator component's algorithm to single threaded to make debugging easier
I know Pluto will send a redirect after a processAction so I'm a bit puzzledThis sounds very interesting.
where this is coming from.
This is partically troublesome as I'm currently trying to implement Struts
as a Jetspeed portlet in which I perform Struts action processing from
processAction and Struts view rendering from doView. Because I need to
transport render information over from the processAction to the doView
handling I put these temporarily in the session. The doView handling removes
it from the session on first usage (just like Struts can temporarily store
its messages in the session).
Now with the doView being called twice this setup clearly is falling apart.
Will this be a general solution that you plan on contributing to J2?
I think a Jetspeed/Struts portlet application framework is a valuable contribution.
Hopefully someone has an idea?Yes, Im pretty much dedicated to working on cleaning up this part of the code over the next week.
No big deal, it will get fixed soon. Im with a client today and tomorrow though
ThanksI've already created an issue for this: http://nagoya.apache.org/jira/browse/JS2-5
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
