When we cancel a confirmation request, we should remove the request from
the height map regardless of the current height. Otherwise we end up in
the situation when the height is reached, the notification is attempted
sent which results in a crash.
If a concurrent call to cancel is made while the notifier is shutting
down, this would cause a panic (close of closed channel) since the
events are not removed from the notification sets.
This addresses a panic when a notification is canceled after its been
detected as included in a block and before its confirmation notification
is dispatched.
This commit introduces an optimization for the chain scanning
used in historical confirmations notifications.
It now uses batch request instead of requesting filters one by one.
Earlier GetCFilter could return a nil-error along with a nil-filter in
cases where we failed to fetch the filter. We would just ignore the
block causing us to miss potential crucial information.
Now that GetCFilter will return an error in cases where the fetch
failed, the filter will never be nil (even for blocks with only coinbase
transactions), and we'll instead return an error.
We also increase the number of retries for the filter from the default 2
to 5.
We would establish the connection to the chain backend and start getting
block notifications before we had started the concurrent queues, which
would lead to the OnBlockConnected call being blocked, and a deadlock
(since GetBestBlock would never return).
Instead we make sure to start the queues before establishing the
connection, consuming the notifications right away.
In this commit, we lower the mempool spend check timeout to be twice as
long as the trickle interval of the miner node, which will greatly
improve the execution time of this specific test. We're able to do this
now since we can specify custom trickle intervals for our test
harnesses.
This prevents a deadlock while tearing down the TxNotifier if it's
currently blocked delivering a notification. By closing the quit chan
first, we ensure blocked sends/reads can exit and allow the TxNotifier
to proceed tearing down.
A height hint not being set would cause lnd to scan for the
confirmation/spend of a txid/outpoint/address from genesis.
The number of confirmations not being set within a confirmation request
would cause the internal TxNotifier to deadlock when dispatching
updates.
Since GetUtxo is a potentially long running call, we would see
RegisterSpendNtfn block in some cases, in particular on starting the
chain watcher, causing lnd to hang on startup.
This commit makes the call to GetUtxo in a go routine, letting
RegisterSpendNtfn return immediately when the notification events are
created.
