> A few weeks ago, I wrote:
>
> The Kissme in CVS and in the Kissme 0.0.31 release are currently
> incompatible with the latest Classpath in CVS ...
>
> Compatibility should be restored in a day or so ...
Last night I fixed the last major bug and checked in a large number of
changes to Kissme. I believe that compatibility between the CVS
versions of Kissme and Classpath should now be restored. If you have
problems, please email the kissme-general list.
In the process of doing the Thread / VMThread refactoring, I uncovered
and fixed a number of pre-existing bugs in Kissme's locking code. For
example:
* The exception handling code was not releasing locks when unwinding
the stack past a call to a synchronized method.
* The code that inflated thin locks to fat locks was broken.
* The code for locking and unlocking fat locks didn't always inc/dec
the recursion count correctly.
* The code for releasing/regaining a lock in a 'wait' didn't work
if the lock was a fat lock.
Fixing these bugs may go some way to curing the unexplained lockups
I was seeing previously. But I've also noticed some other inefficiencies
and potential memory leaks that need fixing. And I need to write some
'hitme' tests that will exercise the locking more thoroughly.
I'm sorry this took so long, but it was a much bigger job than I thought.
-- Steve
_______________________________________________
Classpath mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/classpath
