bah... changing the event store query function to return the events serially messes with the model of how two tier (layer2) event stores work

for now, ok, going to change it to make the assumption that all events in the L2 are known about, just there is only their index and event ID stored... i think that's ok, but it is an extra overhead to consider in the garbage collector... it kinda implies that the low/high water marks need to be lower to allow space for the size of the L2 indexes

but i guess it's not that huge, i mean... ok, tags can take up a lot of space... but compared to the event as well, and the tags are just truncated 8 bytes out of 32

ah, engineering complex things is complex