Hi, Sun has some serious problems in the 1.1.X series with finalization. Basically the finalizer thread can hold locks and threads waiting to be finalized can hold locks and can be waiting for finalization. Obviously, this can easily result in deadlocks (the finalizer needs to wait to obtain a lock while the thread holding that lock is waiting to obtain a finalization queue lock). Sun knows that this is a serious problem, but to fix it given the way finalization is done under 1.1.X is not easy and so they decided to only include a fix in 1.2 (which we don't have access to yet). To prevent the deadlocks, in jdk116_v3 and higher, we have not allowed threads to wait in the finalization queue. Instead they just return false. This greatly slows down finalization but does not stop it (and prevents deadlocks!). If the finalizer queue is empty and finalize() is called, the finalization will be done. If you truly want your application to work well on *all* java platforms, please don't use the finalize() call to do almost anything! (even loading a class can be a problem since there is a class loading lock). At least until 1.2 is out and everybody is using it on all platforms. Our changes might be causing your problems (but they are fixing others, ie. the deadlocks that prevented many programs from launching at all). If you have a SMALL sample program that illustrates this problem, we would be happy to try and track down the specific problem and try to get all things working. Also, you might want to try checking the return value from finalize() to see if it needs to be called again. If you can generate a small sample java program, please submit it using the Blackdown bug database. Thanks, Kevin Hendricks Blackdown JDK porting team member ---------------------------------------------------------- Kevin B. Hendricks Associate Professor, Operations & Information Technology School of Business, College of William & Mary Williamsburg, VA 23187, [EMAIL PROTECTED] http://business.tyler.wm.edu
