[svn] 3779: Continue fix for BLZ-262: null pointer exception on redeploy of web app/servlet.

Question

Revision: 3779
Author: [email protected]
Date: 2008-10-21 10:23:55 -0700 (Tue, 21 Oct 2008)
Log Message:
Continue fix for BLZ-262: null pointer exception on redeploy of web app/servlet.
Remove finally block in servlet init() as the call to destroy() in the catch block will clear ThreadLocals
if there is an error. Move it to the end of try block instead since the init() thread doesn't need any
thread locals set.
Log the startup error before calling destroy() in init() catch clause.
QA: Yes
Doc: No
Ticket Links:
http://bugs.adobe.com/jira/browse/BLZ-262
Modified Paths:
blazeds/branches/3.0.x/modules/core/src/java/flex/messaging/MessageBrokerServlet.java

Answer 1

Hello!
I had a similar problem under WebLogic 6.1 SP4 after deploying a war-file with
a bug.
After deploying the corrected file, it seemed that the old version was deployed
again.
Perhaps this behaviour will help you to load the new application/ejb:
- undeploy
- remove from targets
- move to targets
- deploy
Claudia
"Marco" <[email protected]> wrote:
>
hi all,
i have found a werid problem in deploying applications in Solaris 8
for weblogic
7.0.
i have deployed one EJB, and the first time i access it (thru teh web
app embedded
in the .ear) it gave me a null pointer exception because i was failing
to lookup
one environment variable.
so, i fix the code, rebuild the ear and redeployed, and the same error
happened
again.
i am sure i did not do any mistakees because i have put some system.err
statements
in the bean code to show the workflow, and when i have deployed it on
Window 2k
the statemens appeared.
but in solaris they didnot appear, and i got the same exception that
i had before..
it looks like wls is caching teh ejb, because whenever i redeploy it
(adding more
and more system.err statements) it always execute the faulty code of
when deployed
the very first time.
anyone can help?
regards
marco