With CLJS advanced optimizations on, I get 18 warnings about unreachable core from Konserve, 4 from Datahike itself, and 1 from distributed-scope. Probably not immediate cause for concern, but potential optimization opportunities? Just FYI
Sorry about that, I think they have crept in with the last releases. I have landed memory model improvements for 1.0 (you can have concurrent writers now by default including browsers against S3), but GC is still in flux, once this is landed I will do more cleanups.