We'll need to pack feature vectors for route blinding, so we pull
the encoding/decoding out into separate functions (currently
contained in ChannelType). Though it's more lines of code, we keep
most of the ChannelType assertions so that we strictly enforce
use of the alias.
This commit introduces a filterManager that uses our outbound peers'
feefilter values to determine an acceptable minimum feerate that
ensures successful transaction propagation. Under the hood, a moving
median is used as it is more resistant to shocks than a moving average.
This commit adds a check to ensure that we don't start LND with native
SQL invoice DB if we already have any invoices in our KV channeldb. This
is needed as native SQL invoices is an experimental feature and we do
not yet support migration.
This commit removes a merge artifact introduced by a rebase of a
previous PR.
While we're at it, we streamline the grammar, formatting and general
look-and-feel of the release notes.
When determining the max fee rate of a channel we used to scale
the fee rate depending on our available local balance on this channel.
This lead to a special case that if a channel would be drained we
could especially decrease the fee rate even down to the fee floor.
Now we make sure that our max fee rate will not be lower than the
old fee rate to make sure in case our channel is locally drained
we do not continue to decrease fees too low.
Bitcoind will not report any fee estimation in case it has not
enough data available. We used to just set the min mempool fee
in such cases but this might not represent the current fee situation
of the bitcoin network. We return an error now so that we will use
the fallback fee instead.
In the tower session itest, we make sure to mine blocks on session count
assertion failure. This required because the session count is expected
to change only when 2 things both happen: 1) a closable session is
queued and 2) a new block notification comes through _after_ the session
has been queued. Only then will it be deleted. So we need to do this to
prevent the case where all the block notifications are consumed _before_
the session is queued for deletion. This flake tended to happen in the
windows itest.
Because we pass in the same config into multiple notifiers, we sometimes
get a data race in the unit tests. By creating a copy of the config we
avoid that.
Since we fixed a number of issues in chainntnfs.NewBitcoindBackend that
makes it compatible with bitcoind v26.0, we now want to use that
function in all our unit tests.
With the chainntnfs.NewMiner now being optimized for not creating
nodes with colliding ports, we use it in all unit tests that spin up
temporary miners.
This commit fixes a number of issues with our temporary bitcoind nodes
that we spin up for unit tests:
- We didn't remove the node's temporary data dir after tests finished.
- We used random ports which lead to unwanted port collisions.
- We used ipc:// unix sockets for ZMQ which currently isn't supported
in bitcoind v26.0 due to a regression. Since we can reliably create
new non-colliding ports now, we should use TCP for ZMQ anyway.
With this commit we create a new function that returns system wide
unique ports by using a single file to keep track of previously used
ports. We'll want to use this everywhere whenever we need to listen on a
new, random port during unit or integration tests.
Because we now have a unique source, we don't need to apply the port
offset that was used for the different tranches of parallel running
integration tests before.
Carla Kirk-Cohen