Unexpected change of behavior of ForkJoin common pool clearing thread locals

As part of https://bugs.openjdk.java.net/browse/JDK-8246585 the ForkJoin common pool has changed it behavior wrt ThreadLocals

The change introduced makes the common pool clear the ThreadLocals since the common pool is now created using InnocuousForkJoinWorkerThread.

Is this change intended?

We found this when using logstash-logback-encoder together with the common pool - https://github.com/logfellow/logstash-logback-encoder/issues/722