Message11197

One side-note: Looking at the jstack-output it looks like your local file-paths appear for .py-files (see #2535). I thought this had been fixed as of https://hg.python.org/jython/rev/45e1a6b52935. Do you have an idea why this fails in your case?


Regarding the deadlock, I suspect it is kind of a hidden deadlock, see https://www.javacodegeeks.com/2013/02/java-concurrency-the-hidden-thread-deadlocks.html.

Please run

jstack -l

next time. The -l option includes more info about "ownable synchronizers".

Also, it is maybe related to https://github.com/jythontools/jython/commit/b142d46c2968b32a1823e438cea41ccbc6ac2978
(pobably all our recently observed deadlocks are, they all involve fromClass, createType and such). The main problem with analyzing this is that the message

waiting for ownable synchronizer 0x00000000eab642c8, (a java.util.concurrent.locks.ReentrantLock$NonfairSync),
  which is held by "MainThread"

Doesn't tell where and how MainThread got this lock. (Maybe it's a read-lock on what PyType.getClassToType() returns.) The link above explains that read-locks don't show up in jstac output. So this will be involved.