I thought we had an issue for this, but I guess not. We should really use the ChannelMonitorUpdate
s in the FilesystemPersister
instead of always persisting the full ChannelMonitor
.
Because we're moving towards the logic not being encapsulated in the FilesystemPersister
in #1417 we'll have to do a bunch of work there - probably we'll want to move the read_channelMonitors
thing into lightning
using a get
method added to the KVStorePersister
trait, with the ability to roll forward the monitors using updates and then remove the updates after persisting the new monitor.