Why are we switching from lazily to eagerly initia...
# development
a
Why are we switching from lazily to eagerly initialising?
w
because we could (since we knew when the fork was completed), and because it was less code
a
Some of the initialising is expensive… e.g. the LMDB stuff ends up doing a bunch of filesystem operations, where we probably don’t actually use the LMDB from most of the forks…
I’m not sure about the grpc stack; I’m pretty sure it starts threadpools, I’m not sure whether it performs any RPCs, but it may…
w
we always will use the lmdb stack currently
hm, that's not true for simpler goals actually.
the relative cost is low though.
a
ok 🙂