Hi Max,
If you haven't already committed the patch can you do it by eod?
Thanks!
Pavitra
On 10/20/2010 10:55 AM, MAX STARETS wrote:
No problem, Pavitra.
I am planning to commit the patch today unless there are objections
from anybody on this list.
Max
On 10/20/2010 12:47 PM, Pavitra Subramaniam wrote:
Thanks for the review Max!
Can you commit the changes if there are no other concerns?
Thanks
Pavitra
-------- Original Message --------
Subject: [jira] Updated: (TRINIDAD-1942)
ViewDeclarationLanguageFactoryImpl implementation should cache
physical URI for the current viewId on the ViewMap
Date: Wed, 20 Oct 2010 12:28:26 -0400 (EDT)
From: Pavitra Subramaniam (JIRA) <[email protected]>
To: [email protected]
[https://issues.apache.org/jira/browse/TRINIDAD-1942?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Pavitra Subramaniam updated TRINIDAD-1942:
------------------------------------------
Attachment: jira-1942-II.patch
This patch (jira-1942-II.patch) contains the code review fix suggested by Max.
> ViewDeclarationLanguageFactoryImpl implementation should cache physical URI
for the current viewId on the ViewMap
>
-----------------------------------------------------------------------------------------------------------------
>
> Key: TRINIDAD-1942
> URL:https://issues.apache.org/jira/browse/TRINIDAD-1942
> Project: MyFaces Trinidad
> Issue Type: Bug
> Components: Components
> Affects Versions: 2.0.0-alpha-2
> Reporter: Pavitra Subramaniam
> Attachments: jira-1942-II.patch, jira-1942.patch
>
>
> In an included page (or fragment) if a composite component gets used, the
CompositeComponentTagHandler retrieves the logical viewId
(context.getViewRoot().getViewId()) first and then gets its physical URI in order
to determine the VDL to use. This is usually not a problem in vanilla JSF
applications or even vanilla Trinidad applications, but in frameworks (such as
Oracle) where the viewId of an included fragment comes into play (iow, the
fragment is in context and not the main page), the PageResolver implementation can
no longer return the correct physical URI for the logical viewId. It merely
returns the logical viewId.
> The problem is further exasperated if the logical viewId of the page does
not contain an extension. As there is no extension available, the VDL returned is
the JspViewHandlingStrategy always. This is disastrous when the page is being
rendered using FaceletsVDL.
> One way to resolve this issue is to cache the physical URI of the page on
the ViewRoot's viewMap and return the cached value whenever the
VDLFactory.getVDL() is called for a viewId that matches the current viewRoot's
viewId.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
