Review representation of `mask` in `Event`

[hannobraun]

Review representation of `mask` in `Event`

[hannobraun]

I just had this thought regarding this issue:
The current test suite covers the basics, assuming that anything not explicitely covered will work fine, since we don't do anything fance that's specific to the type of event. Before implementing this issue, we should have a full test suite in place that covers all types of events, to make sure all our assumptions are correct.

Extending the test suite in such a way would be a good idea anyway, but it would be especially important before we hardcode any assumptions about how the event mask is structured.

[hannobraun]

Another tought:
I think it would be better to model event queue overflows as an error, not just another kind of event.

[hannobraun]

Additional thoughts:

• The cookie could be removed. Instead it could be added as a field to the rename event variants of the EventKind enum. Then an Eq implementation or a dedicated method could be used to determine whether the rename events match.
• IN_IGNORED events could be made more explicit. Right now they're just another kind of event.