The current version of Tomcat we use for bigapps testing is Tomcat 4.0.
We are in the process of migrating to Tomcat4.1.24. When dry run the
new test suite, we encounter problem with Tomcat4.1.24.

With Mantis b19, Tomcat 4.1.24 hang after a few iterations.
The problem happened with both -server and -client, on solaris and windows
platforms. ( we have not tried on linux yet )

How to reproduce the problem:
1. telnet to jtg-e420-5.sfbay
2. export JAVA_HOME=/usr/j2se
3. execute the script /net/jtgb4u4c.sfbay/export/sail8/bigapps/tests/runtomcat,v4.1.24.ksh -server
4. log files are under /bt/Tomcat*
   after a few iterations, test hang.

   error message in /bt/Tomcat*/tomcat4.1.24/logs/catalina.out
   "SEVERE: Caught exception executing org.apache.tomcat.util.net.TcpWorkerThread@c06258, terminating thread
java.lang.OutOfMemoryError"

###@###.### 2003-03-26

With the mantis VM, this version of tomcat gets a SEGV because there is a
race between filling the stack traces and another thread counting the depth
of the stack trace. Trace at signal:

  ---- called from signal handler with signal 11 (SIGSEGV) ------
=>[8] java_lang_Throwable::get_stack_trace_depth(0xf2803780, 0xb70238, 0x0,
0xf9814120, 0x14, 0x0), at 0xfe3ba3e0
  [9] JVM_GetStackTraceDepth(0xb702cc, 0xecb80ffc, 0xecb80ef8, 0xf9815e50,
0xf2809a50, 0x0), at 0xfe40e2b4
  [10] Java_java_lang_Throwable_getStackTraceDepth(0xb702cc, 0xecb80ffc, 0x0,
0xf9814120, 0x8, 0x0), at 0xfe79e670
  [11] 0xf980b96c(0xecb80ffc, 0xecb81090, 0x0, 0xf98160d0, 0x0, 0x0), at
0xf980b96b
  [12] 0xf9805774(0xecb81094, 0xecb81098, 0x0, 0xf98160d0, 0x4, 0xecb80f90), at
0xf9805773
  [13] 0xf9805750(0xecb8112c, 0x78, 0x0, 0xf98160d0, 0x4, 0xecb81028), at

With this fixed, the application still runs out of memory fairly quickly.

###@###.### 2003-03-26