Is there a way to fail an automated test upon Netty leak detection? -


I am using Netie 4.0.x on a project where a different project ByteBuf Pass them on buffers and client code layers, which should be responsible for shutting down the buffer.

I have found a leak in some cases and I want to cover the codepath which is an automatic test with those leaks, but Natalie's Resistor detector only reports leaks inside the logs.

Is there a way to fend off automatic JUnat test in case of such leakage? (Such as by some behavior plugin in ResourceLeakDetector)?

Thank you!

PS: Keep in mind that my test will not actually make buffers, the core code (which is one dependency) does.

Indicates the build is unstable when the leak is detected. I'm not sure on the mechanism for this, but possibly to find out the possibility (possibly detecting leaked messages from build logs).

Keep in mind that the bytefuge objects are GCs, when / where the leakage can be detected, it may be far from when / where the leak actually happened. Nati contains a mark of such places where the bytebook is being reached in a meaningful way to help you breathe where the buffer was originally allocated, and possibly a list of those objects, which That can be responsible for issuing buffer.

If you are fine with the above limitations, you can add a throw of exception for personal use.


Comments