[
https://issues.apache.org/jira/browse/PLUTO-523?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12667826#action_12667826
]
Ate Douma commented on PLUTO-523:
---------------------------------
- adding new PortletContainer enum Method type defining the specific supported
container methods (.e.g. render/resource/action/admin/load)
- recording current invoked Method enum value in ContainerInvocation so it can
be evaluated/used within container processing logic
- fixing the PortletRequestDispatcher.forward method to actually call (servlet)
rd.forward instead of "faking" it through an rd.include
- adjusting DefaultPortletInvokerService to use forwards for
PortletContainer.Method.RESOURCE
- adding get/set properties methods to PortletURLProvider as formally required
by the spec (no real implementation needed: this is intended for "vendor"
specific properties handling)
- replacing the RequestPropertyProvider (which internal implementation was
really, really messy and broken in many ways) by more generic PropertyManager
(very similar to old Pluto 1.1.x PropertyManager actually).
- PropertyManager.setProperty (response headers) will write them to the
servlet.response when PortletContainer.Method.RESOURCE, setProperty(Cookie)
will now always be written out
> Further abstractions of the Pluto SPI to support embedding in and extending
> by other portals
> ---------------------------------------------------------------------------------------------
>
> Key: PLUTO-523
> URL: https://issues.apache.org/jira/browse/PLUTO-523
> Project: Pluto
> Issue Type: Task
> Affects Versions: 2.0.0
> Reporter: Ate Douma
> Assignee: Ate Douma
> Fix For: 2.0.0
>
>
> While the mayor Pluto 2.0 SPI refactoring from PLUTO-481 is done, we're
> encountering new, although smaller, issues with getting Jetspeed running with
> Pluto 2.0.
> As an example, the Pluto PortletRequestImpl.getContextPath() uses the
> PortletApplication name to derive the contextPath.
> This works for "normal" portlet applications but Jetspeed also has a "local"
> portlet application feature, which means the context path of such an
> application actually is the (Jetspeed) portal application.
> To support overriding this getContextPath() resolution (without wrapping each
> and every PortletRequestImpl), I'm moving this to a new
> InternalPortletContext.getContextPath (similar with Servlet 2.5
> ServletContext.getContextPath).
> As Jetspeed already overrides/wraps the InternalContextPath, we now can
> easily adjust the returned contextPath based on the type of portlet
> application (local vs. plain webapp)
> This issue, I'll keep open as overall item to record similar small SPI
> refactoring changes.
> Another coming up (not done yet) is the PortletURLProvider interface which
> doesn't yet encapsulates resourceURL Cacheability or ResourceID state.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
