Re: Findbugs and locks?

Jeff Higgins wrote:
if (n < lockArray.length) {
ReentrantReadWriteLock rrwl = lockArray[n];
WriteLock lock = rrwl.writeLock();
try {
lock.lock();
// do some disk I/O
} finally {
lock.unlock();
}
}
}

Is there a chance that lock.lock() would fail?

If so, would lock.unlock() cause any harm?

The usual pattern for resource acquisition and release with try-finally is

resource.acquire();
// break flow on acquire() failure
try
{
...
}
finally
{
resource.release();
}

That way the release() only occurs if the acquire() succeeds. Moving the acquire() into the try{} means that the release() will occur even if acquire() fails. This is how Knute coded it, and I don't see why his way caused a warning.

I would be more comfortable if Knute's way didn't cause any warnings, because I don't understand it either. However, I'm unfamiliar with Findbugs (although many recommend it very highly).

--
Lew
.