王正航 wrote:
> This bug was found by a Chinese java community long time age:
> http://mislay.javaeye.com/blog/348338.
We don't poll the 235 000 000 web sites each month to see if someone has
post a message about a MINA problem. We do that only once a year ;).
Seriously, JIRA is the place to fill a bug.
> I think it's best to fix like this:
>
> private boolean await0(long timeoutMillis, boolean interruptable) throws
> InterruptedException {
> long startTime = System.currentTimeMillis();
> //long endTime = System.currentTimeMillis() + timeoutMillis;
>
> synchronized (lock) {
> if (ready) {
> return ready;
> } else if (timeoutMillis <= 0) {
> return ready;
> }
>
> waiters++;
> try {
> for (;;) {
> try {
> long timeOut = Math.min(timeoutMillis,
> DEAD_LOCK_CHECK_INTERVAL);
> lock.wait(timeOut);
> } catch (InterruptedException e) {
> if (interruptable) {
> throw e;
> }
> }
>
> if (ready) {
> return true;
> } else {
> //if (endTime < System.currentTimeMillis()) {
> if (System.currentTimeMillis() - startTime >
> timeoutMillis) {
> return ready;
> }
> }
> }
> } finally {
> waiters--;
> if (!ready) {
> checkDeadLock();
> }
> }
> }
> }
>
> The difference between two time values will never be overflowed!
>
The code has already been patched. See
http://svn.apache.org/viewvc?rev=781711&view=rev
--
--
cordialement, regards,
Emmanuel Lécharny
www.iktek.com
directory.apache.org
