[jira] Issue Comment Edited: (TRINIDAD-1923) CLONE -Issue in org.apache.myfaces.trinidad.model.ProccessUtils.getMaxVisitedRowKey

["Pavitra Subramaniam (JIRA)" <de...@myfaces.apache.org>]

    [ https://issues.apache.org/jira/browse/TRINIDAD-1923?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12914297#action_12914297 ] 

Pavitra Subramaniam edited comment on TRINIDAD-1923 at 9/23/10 8:46 PM:
------------------------------------------------------------------------

Uploaded a patch file with the fix for this issue on trunk. The patch was worked on together by Josu (see base issue 1844) and Pavitra. 

      was (Author: pasubra):
    Uploaded a patch file with the fix for this issue on trunk. The patch was worked on together by Jose (see base issue 1844) and Pavitra. 
  
> CLONE -Issue in org.apache.myfaces.trinidad.model.ProccessUtils.getMaxVisitedRowKey
> -----------------------------------------------------------------------------------
>
>                 Key: TRINIDAD-1923
>                 URL: https://issues.apache.org/jira/browse/TRINIDAD-1923
>             Project: MyFaces Trinidad
>          Issue Type: Bug
>    Affects Versions: 2.0.0.3-core
>         Environment: Mac OSX 10.5 with JVM 1.6.0_20.
> Running using ADF with JDeveloper 11.1.1.3.0
>            Reporter: Pavitra Subramaniam
>            Priority: Minor
>         Attachments: ProcessUtils.java
>
>
> (Created this issue for fixing on trinidad trunk)
> I have been running the following example: http://adf-samples.googlecode.com/files/SampleTrainModelWithCustomAction.zip
> I just changed the view.train.TrainIdMenuModel class constructor to:
>     public TrainIdMenuModel() {
>         super();
>         super.setMaxPathKey("MyMaxPathKey");
>     }
> This is done to get a max visited node behavior.  However when I click on the button 'Next' to navigate on the tree the nodes are not enabled as expected and the user has to click on 'Back' and then 'Next' twice to get the nodes enabled correctly.
> The example can be found in this blog entry: http://jobinesh.blogspot.com/2010/04/custom-model.html  which is based on the ADF rich client demo (http://www.oracle.com/technology/products/adf/adffaces/11/doc/demo/adf_faces_rc_demo.html).  The rich client demo can also be used to reproduce the issue.
> I accept that using a case test that relies in a whole framework like ADF is not ideal, but looking at the code of org.apache.myfaces.trinidad.model.ProcessUtils I think that the problem lies there.  In the javadoc it is stated: "If set but the focus rowKey is after maxVisitedRowKey, set maxVisitedRowKey to the focus rowKey."
> However if we look at the code, we can see that this is not respected since the maxPath variable is 'cached' at the request map.
> If we assume that a MenuModel is going to be updated during the processing of a request  BUT that the ProccessUtils.getMaxVisitedRowKey is invoked BEFORE the model is actually updated, the maxPath value will be calculated (and cached) using the non updated model.  In particular the maxPath will be calculated incorrectly if the focusRowKey of the model is NOT updated before the first call to ProccessUtils.getMaxVisitedRowKey occurs during the request.  After this, if the focusRowKey is updated in the model and ProccessUtils.getMaxVisitedRowKey is called (always being in the context of the same request), the contract will not be respected and the cached maxPath value will be returned (which can correspond to a node BEFORE the current focus node, which is wrong).

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.