Clojurians Log v2

I am thinking about changing the index layout of Datahike for better performance to get closer to FoundationDB, basically by putting eid, attribute ids and maybe also value prefixes into flat arrays that can be compared against each other much faster. @grischoun looked into this a few years ago. I am not sure yet how to exactly do it, but it would reduce general index scan overhead considerably. This would be a breaking change though (ideally I don't want to maintain two indices going forward), so I would like to hear people's opinions, in particular @alekcz360 and @hoertlehner. How painful would it be for you?