Although it it not that unreliable, it causes false negatives once in
a while. This is because the standard Java SE API does not provide a
way to force a *full* garbage collection run, we need to resort to
unreliable hacks to trigger one.
The test itself is still useful to diagnose or refute alleged memory
leaks.
This commit also move the test from JUnit 4 to Junit 5.
remove PacketMultiplexListener, RoomListenerMultiplexor and
ConnectionDetachedPacketCollector(Test), which was a bunch of
(in same cases redundant) code that formed a complex construct that
presumably tried to make MultiUserChat instances easily garbage
collect-able.
Now, MultiUserChat should be eligible for gc if the userHashLeft() is
invoked before the reference to the instance is dropped, which should be
the case in the most scenarios. Otherwise the connection may references
the MultiUserChat instance over Packet(Listener|Interceptor)s preventing
the gc.
instead of using a PacketListener, which means that the user has to
downcast the Packet to Message, we now use a Listener which callback
parameter is already Message/Presence.
It is necessary to introduce MessageListener and PresenceListener, which
are interfaces that have a callback for Message/Presence instead of
Packet. The 'old' MessageListener is renamed to ChatMessageListener.
Use Generics in ConnectionDetachedPacketCollector.
instead of using the old baseName=smack appendix=project.name approach,
we are now going convention over configuration and renaming the
subprojects directories to the proper name.
Having a prefix is actually very helpful, because the resulting
libraries will be named like the subproject. And a core-4.0.0-rc1.jar is
not as explicit about what it actually *is* as a
smack-core-4.0.0-rc1.jar.
SMACK-265
