There was this
https://issues.apache.org/jira/browse/TRINIDAD-1632
On 1/26/2011 10:15 AM, MAX STARETS wrote:
Blake,
Andy and I have discussed this offline a bit, and his preference was we do not
hack the flag
to be off explicitly (he thought it would create even more confusion). He also
suggested that we look
at fixing the issues in Trinidad sooner rather than later. The users are
getting the following exception:
java.lang.NullPointerException
at
org.apache.myfaces.trinidad.component.UIXComponentBase._getRendererImpl(UIXComponentBase.java:1299)
at
org.apache.myfaces.trinidad.component.UIXComponentBase._cacheRenderer(UIXComponentBase.java:1279)
at
org.apache.myfaces.trinidad.component.UIXComponentBase.__rendererDecode(UIXComponentBase.java:1193)
at
org.apache.myfaces.trinidad.component.UIXComponentBase.decode(UIXComponentBase.java:724)
at
org.apache.myfaces.trinidad.component.UIXForm.processDecodes(UIXForm.java:71)
at
org.apache.myfaces.trinidad.component.UIXComponentBase.decodeChildrenImpl(UIXComponentBase.java:1081)
at
org.apache.myfaces.trinidad.component.UIXComponentBase.decodeChildren(UIXComponentBase.java:1067)
at
org.apache.myfaces.trinidad.component.UIXComponentBase.processDecodes(UIXComponentBase.java:839)
It is caused by the absence of the render kit Id on the view root after state
restoration.
*Gabrielle*, I thought you have prototyped a fix for this issue a while back.
Do you have the the code somewhere?
Thanks,
Max
On 1/26/2011 12:23 PM, Blake Sullivan wrote:
I think we need to do both--essentially force partial state saving off and
document that we are going to force it off.
-- Blake Sullivan
On 1/26/11 6:57 AM, Matthias Wessendorf wrote:
On Wed, Jan 26, 2011 at 3:33 PM, MAX STARETS<[email protected]> wrote:
Hello Everyone,
I think we need to do something about the fact that any Facelets application
with Trinidad 2.0
will fail in a very non-obvious way unless partial state saving is disabled
with the following context
parameter:
<context-param>
<param-name>javax.faces.PARTIAL_STATE_SAVING</param-name>
<param-value>false</param-value>
</context-param>
I see two options:
Document it prominently in the release notes
Try to "fake" the parameter in Trinidad code as always being set.
I am not sure if (2) is feasible, but the advantage would be that the users
will not have to manually re-enable partial
state saving when we resolve the compatibility issues.
I am fine with (2), even if it's a bit hacky :-)
Opinions?
Max
