Swing Reliability test run hangs at 12th hour of execution, and heap overflow exception appeared, while pass criteria for that test is running for 72 hours without hanging and any exceptions. This looks like memory lick. The failure screenshot is attached. Test run log directory can be accessed at /net/vice.russia.sun.com/export/home0/results/1.5.0_10/Build01/swingbasher/winXPhome-x86/
OS : Win XP home edition with SP2
Hardware : x86
C:\>jdk1.5.0_09\bin\java -version
java version "1.5.0_09-ea-b01"
Java(TM) 2 Runtime Environment, Standart Edition (build 1.5.0_09-ea-b01)
Java HotSpot(TM) Client VM (build 1.5.0_09-ea-b01, mixed mode, sharing)
VM_OPTS : -server
How to reproduce :
Test suite location : /net/vice.russia.sun.com/export/home0/dtf/dtf_ws/suites/1.5.0_URv09/swing/
Full test procedure is attached (Swing_TOI.sxw).
Brief test procedure :
1. copy the test suite to test machine, install jdk under test
2.
$export SB_TIMELIMIT=4320 (timeout in minutes)
$mkdir c:/SwingBasher/
$mkdir c:/SwingBasher/classes
$mkdir c:/SwingBasher/results
$cd ${TEST_SUITE_LOCATION}/src/Swing_AllComponents/Automated/SwingBasher/tiger/
$sh Run.sh <JAVA_HOME> <SUITE_DIR> <OUTPUT_DIR> <RES_DIR> <JTABLE_PRINT_LOG> <VM_OPTS>
JAVA_HOME : Specify the JAVA HOME directory path
SUITE_DIR : Specify the SUITE directory path.
OUTPUT_DIR : Specify the Output directory path.
RES_DIR : Specify the Result directory path.
JTABLE_PRINT_LOG : Specify the store/remove( It tells the harness whether to store/remove the files generated as a result to JTABLE print testacy).
VM_OPTS : Specify the VM options.
Ex:
$sh Run.sh c:\jdk1.5.0_09 c:\swing\src\Swing_AllComponents\Automated\SwingBasher\tiger c:\SwingBasher\classes c:\SwingBasher\results store -server
Screen saver should be switched off since the test requires display.
Please let me know if you need some kind of assistance in reproducing the issue. I'm also able to provide a test machine.
Thanks for your attention,
Ilya. ###@###.###
The fix is verified for update 9 build 2. Currently I'm checking update 10 build 2. Will update the CR as soon as I would have results of this run.