Event loss because of Timestamp filter

[pLeminoq]

I got the following log which causes unit tests to fail:

21:02:45.055 [pool-1-thread-92 @coroutine#141] INFO  o.o.b.r.u.c.UnitRegistryController FilterData in updateDataToPublish
21:02:45.074 [pool-1-thread-98] INFO  o.o.j.s.r.ProtoBufFileSynchronizedRegistry [LocationConfigRegistry]: 100% consistency checks passed of LocationConfigRegistry after 12 applied modifications.
21:02:45.086 [pool-1-thread-98] INFO  o.o.b.r.u.c.UnitRegistryController FilterData in updateDataToPublish
21:02:45.089 [pool-1-thread-98] WARN  o.o.b.r.u.c.UnitRegistryController UnitRegistryData filterDataForUser with transaction id 221
21:02:45.090 [pool-1-thread-92 @coroutine#141] WARN  o.o.b.r.u.c.UnitRegistryController UnitRegistryData filterDataForUser with transaction id 206
21:02:45.092 [pool-1-thread-92] INFO  o.o.b.r.u.r.UnitRegistryRemote Skip event on scope[/test/thuxohl/registry/unit] because event seems to be outdated!
21:02:45.094 [pool-1-thread-7] INFO  o.o.b.r.u.r.UnitRegistryRemote Transition check failed, received 206 but waiting for 221 of UnitRegistryRemote[scope:/test/thuxohl/registry/unit]

You can see that one thread (the coroutine) performs a requestData while another thread performs a notification. They get mixed up in a way that the latest data (from the notification) is discarded by the remote because of an older timestamp.

[pLeminoq]

We just checked and the print Skip event on scope ... should only happen if no one is logged in with the default session manager. We should check if this is the case during startup of the mock registry. If yes logging in could at least mitigate the problem to cases where no one is logged in.