On Thu, 9 Jun 2011, Felix Schumacher wrote:
Am Donnerstag, den 09.06.2011, 04:53 +0100 schrieb Jocelyn Ireson-Paine:
Hi,
I'm getting sporadic null-pointer exceptions from 'response.encodeUrl'.
This is with Tomcat 7.0.14, the latest stable version to which I upgraded
an hour ago, and Java 1.6.0_26, which again I upgraded to at the same
time, under Windows XP. I also got the error under Tomcat 5.5.33, which is
why I tried upgrading. I uninstalled old Tomcats and Javas before
upgrading.
Here's a trace of the error:
java.lang.NullPointerException
at
org.apache.catalina.connector.Response.toAbsolute(Response.java:1594)
at
org.apache.catalina.connector.Response.encodeURL(Response.java:1198)
at
org.apache.catalina.connector.ResponseFacade.encodeUrl(ResponseFacade.java:422)
at sharesim.ValueGame.encodeUrl(ValueGame.java:468)
The fourth line of the trace is my code.
To get some diagnostics, I encapsulated 'response.encodeUrl' in the
following method, and used 'println' to print to Tomcat's log. This method
You can use the servlet api's log method to log inside a web
application, or even better use a logging-framework.
is the one mentioned on the fourth line of the above trace.
private final static String encodeUrl( HttpServletResponse response
, String url
)
{
System.out.println( "encodeUrl" );
System.out.println( "response=" + response );
System.out.println( "url=" + url );
String result = null;
if ( response == null )
result = "not on Web";
else
result = response.encodeUrl( url );
System.out.println( "result=" + result );
return result;
}
As mentioned above, the errors are sporadic. My redirection code gets
called on my server when I submit one of my forms. But sometimes, the
submit works; sometimes it doesn't. Here are two successive traces from
the above method:
encodeUrl
response=org.apache.catalina.connector.ResponseFacade@1a5ec6c
url=/ResearcherValueGame1.jsp
result=/ResearcherValueGame1.jsp
encodeUrl
response=org.apache.catalina.connector.ResponseFacade@1a5ec6c
url=/ResearcherValueGame1.jsp
The first one worked; the second crashed rather than returning the encoded
URL. The URL to be encoded is the same in both cases, and 'response' is
evidently the same instance. I don't know how to dump 'response' in order
to show relevant fields (whichever they are), but am happy to try it if
someone can suggest how.
You haven't shown us the code, which calls your encodeUrl. Most of the
times someone reports sporadic errors while working with responses, he
is saving references to the responses in a class variable.
Could this be your problem?
Hi Felix, and thanks for the comment. What is it about saving references
to the responses in a class variable that can cause problems? Is it to do
with threading: for example, do different threads all see the same class
variable, thus potentially interfering with one another? This is something
I need to know.
More specifically, I'd like to understand why you want to know this in
addition to the diagnostics I included with the mail. These showed two
calls to 'response.encodeUrl'. In both, the URLs were the same. The
response instances were presumably also the same, since they had the same
instance ID, ResponseFacade@1a5ec6c. In particular, neither response
instance was null. So although one call crashed, it wasn't because of a
nulled response, caused perhaps by losing the contents of a class
variable. I wondered whether you suspect that that might be my problem.
Although I suppose the response instances could have been interfered with
by code that updates some of their fields. As it happens, my code
doesn't do that (unless I remember wrongly), but you obviously wouldn't
know that.
I'm not criticising your comment, by the way; just trying to understand
why you asked. My code is quite long, but I can summarise the call chain
and response assignments if we don't solve the problem before.
Felix
Jocelyn
I've found a few mentions on Google of such errors, but nothing definite
about it happening in Tomcat 7. One discussion which I thought I'd seen,
but can't find again, seemed to suggest that the problem occurred on
version 5, possibly because of a missing 'synchronized'. I've not seen
anything in recent postings to this list.
I wondered whether the problem might be caused by the browser creating
malformed cookies, or some such. (I'm using session-handling.) However,
this seems unlikely, as the error occurs regardless of whether I use
Firefox, Internet Explorer, or Google Chrome.
Any ideas?
Jocelyn Ireson-Paine
http://www.j-paine.org
Jocelyn's Cartoons:
http://www.j-paine.org/blog/jocelyns_cartoons/
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
