MLS crashes in AWT-Windows thread

[dajoker]

Over the past few months MLS has had a hard life and has crashed quite a few times. I've kept notes each time to try to find a pattern and so far I have been unable to find anything in how I am (or others are) using the program when it crashes. With that said, the hs_err_pid*.log files have a trend that is pretty interesting. Using grep to find the thread at fault turns up the same thread, and almost the same stack pointer, in every case (all but one) somewhere in the AWT thread:

hs_err_pid1132.log:=>0x160d3c00 JavaThread "AWT-Windows" daemon [_thread_in_native, id=1096, stack(0x162e0000,0x16330000)]
hs_err_pid1228.log:=>0x160dac00 JavaThread "AWT-Windows" daemon [_thread_in_native, id=2472, stack(0x162e0000,0x16330000)]
hs_err_pid1484.log:=>0x160e3800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=1412, stack(0x162e0000,0x16330000)]
hs_err_pid2060.log:=>0x160e4800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=1028, stack(0x162e0000,0x16330000)]
hs_err_pid2528.log:=>0x160e6800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=3824, stack(0x162e0000,0x16330000)]
hs_err_pid2564.log:=>0x160e2800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=3232, stack(0x162e0000,0x16330000)]
hs_err_pid2580.log:=>0x160e9000 JavaThread "AWT-Windows" daemon [_thread_in_native, id=5900, stack(0x162e0000,0x16330000)]
hs_err_pid2584.log:=>0x160dec00 JavaThread "AWT-Windows" daemon [_thread_in_native, id=328, stack(0x162e0000,0x16330000)]
hs_err_pid3224.log:=>0x160de800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=1948, stack(0x162e0000,0x16330000)]
hs_err_pid3524.log:=>0x160e7000 JavaThread "AWT-Windows" daemon [_thread_in_native, id=5548, stack(0x162a0000,0x162f0000)]
hs_err_pid3932.log:=>0x160d2800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=884, stack(0x162e0000,0x16330000)]
hs_err_pid640.log:=>0x160e9000 JavaThread "AWT-Windows" daemon [_thread_in_native, id=4580, stack(0x162e0000,0x16330000)]

Also interesting on my Linux laptop where I have a sample unit I get similar crashes by thread, though with obviously different memory addresses:

hs_err_pid39.log:=>0x03310400 JavaThread "AWT-Windows" daemon [_thread_in_native, id=53, stack(0x03710000,0x03810000)]
hs_err_pid40.log:=>0x0330c800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=54, stack(0x03710000,0x03810000)]
hs_err_pid41.log:=>0x0330c800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=55, stack(0x03d70000,0x03e70000)]
hs_err_pid45.log:=>0x0330c800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=64, stack(0x03710000,0x03810000)]
hs_err_pid49.log:=>0x0330c800 JavaThread "AWT-Windows" daemon [_thread_in_native, id=11, stack(0x03930000,0x03a30000)]

In all of these cases the thread indicated is one of approximately thirty total threads running in the JVM. The system is running on windows XP SP3, and Java is version 1.6.0_15 (per the files above so this should be accurate for the MLS JRE). Most of the crashes happen when I'm doing nothing more than moving the mouse (to click 'Print' on a report, or to click 'Ok' on a dialog box, or something similar to that). For comparison see the '/Program Files/LDS Church/MLS' directory and look for files hs_err_pid*.log. I'm currently on MLS 3.4.2 though this has happened since (by these files' dates) May with whichever version of MLS was current at that time

Does anybody know of a way to enable the memory dump for crashes to try to see what is leading up to these?