I’ve just lost 10 mins of work. I HATE JIRA.
There is a set of merging rules that can be summarized as ‘what happens in a deploy, stays in a deploy*', or the cases with the labels:
* (depending on definition of ‘happens in a deploy’ this can be false, as shown by the PERSISTENT_* cases. TL;DR a COMM entirely in the deploy can contain a persistent event that could potentially conflict outside of the deploy)
Since the other merge rules don’t refer to the notion of deploy (they’re just concerned with the type of incoming events, and what waits in the tuplespace), there are chances they cover the deploy-boundary-dependent rules mentioned above.
If that were the case, we could remove any prod code referring to the notion of volatile/local event, and either remove the corresponding tests or leave them as additional safety net.