[jira] Commented: (WICKET-1010) Contract of Session.attach() and Session.detach()

["Martin Funk (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/WICKET-1010?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12530650 ] 

Martin Funk commented on WICKET-1010:
-------------------------------------

If I interpret this right:
http://www.nabble.com/Contract-of-Session.attach%28%29-and-Session.detach%28%29-tf4517840.html#a12887573

the patch should read like this:

$ svn diff ./jdk-1.4/wicket/src/main/
Index: jdk-1.4/wicket/src/main/java/org/apache/wicket/Session.java
===================================================================
--- jdk-1.4/wicket/src/main/java/org/apache/wicket/Session.java (revision 579923)
+++ jdk-1.4/wicket/src/main/java/org/apache/wicket/Session.java (working copy)
@@ -305,7 +305,22 @@
         */
        public static void unset()
        {
-               current.set(null);
+               try
+               {
+                       Session session = (Session)current.get();
+                       if (session != null)
+                       {
+                           session.detach();
+                       }
+               }
+               catch (RuntimeException re)
+               {
+                       log.error("there was an error unsetting the current session", re);
+               }
+               finally
+               {
+                       current.set(null);
+               }
        }

        /** A number to generate names for auto create pagemaps */

> Contract of Session.attach() and Session.detach()
> -------------------------------------------------
>
>                 Key: WICKET-1010
>                 URL: https://issues.apache.org/jira/browse/WICKET-1010
>             Project: Wicket
>          Issue Type: Bug
>            Reporter: Martin Funk
>
> What is the contract of Session.attach() and Session.detach() ?
> Especially, is it intended that after a call to attach() that there will
> be at least one call to detach() before the request goes back to the client?
> If that's the case, then there might be a bug in Session and I propose
> the following patch on org.apache.wicket.Session 
> Index: .
> ===================================================================
> --- .    (revision 579354)
> +++ .    (working copy)
> @@ -305,6 +305,11 @@
>      */
>     public static void unset()
>     {
> +        Session session = (Session)current.get();
> +        if (session != null)
> +    {
> +        session.detach();
> +    }
>         current.set(null);
>     }
> In my current project well fell over this looking at:
> WicketFilter.getLastModified(final HttpServletRequest servletRequest)
> where cachable resources lead over Session.findOrCreate to Session.set(Session) to Session.attach()
> but the Session.unset() doesn't lead to a Session.detach()
> Martin 

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