We should run all or some of our tests at the PARANOID leak detection level, and ensure the build fails if a leak is detected. Note it might be necessary to use releaseLater for buffers created in the tests themselves to guard against false positives.
http://netty.io/wiki/reference-counted-objects.html
We should run all or some of our tests at the PARANOID leak detection level, and ensure the build fails if a leak is detected. Note it might be necessary to use
releaseLaterfor buffers created in the tests themselves to guard against false positives.http://netty.io/wiki/reference-counted-objects.html