Interestingly, I can't reproduce the issue on OS X, either. The servers where I see the exceptions are debian linux servers, running sun's jdk, version 1.5.0_09-b01 on one server, 1.5.0_10 on another, and 1.5.0_12-b04 on another. All three machines are dual processor machines.

The server where I most frequently see the exception is the one running jdk 1.5.0_09-b01. But it's also the box with the most traffic by far, so it's probably that the frequency of exceptions is a function of the use of the box, rather than the version.

I've got a small test app that I'm going to try running on the box to see if I can recreate the bug on linux, since I can't duplicate it on os x.

Robert

On Jan 19, 2008, at 1/1911:14 AM , Howard M. Lewis Ship (JIRA) wrote:


[ https://issues.apache.org/jira/browse/TAPESTRY-2037?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12560714 #action_12560714 ]

Howard M. Lewis Ship commented on TAPESTRY-2037:
------------------------------------------------

Ok, now I have 13 windows running curl against a couple of different pages, and I'm running interactively. This is fun .. .the performance is still quit good for the interactive client. No sign of an exception yet.

Response rate, measured via FireBug, bounces all over the place: 9ms up to 650ms.


NullPointerException caused by many rapid page refreshes
--------------------------------------------------------

               Key: TAPESTRY-2037
               URL: https://issues.apache.org/jira/browse/TAPESTRY-2037
           Project: Tapestry
        Issue Type: Bug
        Components: tapestry-core, tapestry-ioc
  Affects Versions: 5.0.7
       Environment: jdk 1.5
          Reporter: Howard M. Lewis Ship
          Assignee: Howard M. Lewis Ship
          Priority: Critical
           Fix For: 5.0.8


This was reported on the mailing list.
In certain places, hitting the refresh button rapidly can cause a null pointer exception.
It is believed this is related to a JDK 1.5 bug:
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6550283
Tapestry makes a few uses of ThreadLocal that are consistent with this pattern. ThreadLocals are used to connect service proxies to perthread scope services. We will locate all useages of ThreadLocal and, alas, synchronize access to them.
More discussion: http://markmail.org/message/7bwztu66paz2cfqm

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


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to