I started using commons-pool 2 yesterday to implement connection pooling for Elasticsearch (mainly for health checking, though who knows where this internal project ends up going) and I noticed that Exception is used to track users of objects from the pool. In Log4j, we use a myriad of strategies to obtain caller class info that are much faster than generating a stack trace such as the internal sun Reflection class, a SecurityManager, and we're also looking at the Java 9 StackWalker class which provides a standard way of doing it. Is there any particular reason why DefaultPooledObject uses Exception instead of faster strategies?
-- Matt Sicker <boa...@gmail.com>
