I've just released Jude version 1.07. This is a relatively minor bug-fix release. Thanks to B.L. for reporting the bugs and helping to isolate them.
Interestingly, one of the bugs reported against the previous version was an ArrayIndexOutOfBoundsException at a spot where such an exception really was not possible. This had me really puzzled--I could not duplicate it. But when I discovered that inserting debugging println() calls made it go away, I realized that this was a JVM problem and not my bug. It turns out that in Java 5 (we tested u17 and u18) on Linux (at least) running with the -server option would cause this spurious exception. Running with -client (which is the default for most installations, I think) would not cause it. The crash never occurred at precisely the same spot in a run, leading me to think it was a GC bug. Unfortunately, I've got no idea how to isolate a bug like this with a simple test case so that I can report it.
