There's a small bug in the WatchDog GTest. Simply what it does is that it's
sending TWO Content-Length headers in the same request, and Apache correctly
translates those two headers into one single headers like:
Content-Length: <firstval>, <secondval>
This small patch fixes the problem, but Tomcat 4.0 SHOULD throw back an
error to the client too (IMO).
Pier
Index: src/tools/org/apache/tomcat/task/GTest.java
===================================================================
RCS file:
/home/cvs/jakarta-watchdog-4.0/src/tools/org/apache/tomcat/task/GTest.java,v
retrieving revision 1.1
diff -u -3 -r1.1 GTest.java
--- src/tools/org/apache/tomcat/task/GTest.java 2000/11/03 17:07:20 1.1
+++ src/tools/org/apache/tomcat/task/GTest.java 2001/08/05 18:25:14
@@ -419,9 +419,6 @@
try {
pw.println(request);
- if( content != null) {
- pw.println("Content-Length: " + content.length());
- }
// Now sending any specified request headers
if ( requestHeaders != null )
{
