This commit fixes a panic bug caused by two calls to a process’
cmd.Wait() method. If two nodes incurred a fatal error as soon as they
were created, then both the goroutine detected to recovering the stderr
data and the defer statement would case a double channel close due to
the simultaneous calls to Wait().
This commit slightly modifies the existing CT struct in order to
maintain consistency with code-style. As a result of the name change,
all references have also been renamed from `ct` to `t`.
The Error and Errorf methods have been removed in favor of forcing
everything to be reported via `Fatalf`. Additionally a new method
(ProcessErrors) has been introduced to the networkHarness class in
order to encapsulate the underlying channel.
This commit uses the context package’s WithTimeout to ensure tests
which rely on asynchrony behaviors cannot block the execution of the
integration tests for ever. All tests which rely on async gRPC
notifications now perform a 3 way select: on the timeout channel
closure, an error, and the response itself.
Additionally some slight refactoring has taken place within the current
set of integration tests, eliminating some unneeded factory functions.
This commit adds a new RPC command: `channelbalance` which returns the
sum of all available channel capacity across all open channels. The
total balance is currently returned in units of `satoshis`. Additionally
the `networkHarness` has been modified slightly to allow specifying the
additional "extra" command line parameters when creating the initial
seed nodes. Minor refactoring within the integration tests has been
undertaken in order to increase code re-use across tests.
Closes#29.
This commit modifies the OpenChannel helper method within the
networkHarness to open the channel according to the destination node’s
global lighting ID rather than using a hard coded peer ID as before.
With this change the method is now much more general as the prior
temporary hack has been removed.
This commit introduces the concept of a manually initiated “force”
closer within the channel state machine. A force closure is a closure
initiated by a local subsystem which broadcasts the current commitment
state directly on-chain rather than attempting to cooperatively
negotiate a closure with the remote party.
A force closure returns a ForceCloseSummary which includes all the
details required for claiming all rightfully owned outputs within the
broadcast commitment transaction.
Additionally two new publicly exported channels are introduced, one
which is closed due a locally initiated force closure, and the other
which is closed once we detect that the remote party has executed a
unilateral closure by broadcasting their version of the commitment
transaction.
At times when testing one requires access to the logs of a particular
node in order to aide with debugging. Before this commit, one needed to
manually modify the networkHarness code in order to print either the
location of the logs or the logs themselves. With this commit, tests
can now programmatically examine the logs of any node created within
the networkHarness.
It’s worth noting that at times the logs dumped may not be the most up
to date version of the logs files as the logging library employs
intermediate buffering.
This commit fixes some flakiness exhibited in the current basic funding
workflow tests. This test can fail occasionally in resource constrained
environment due to a race condition which arises after Alice learns of
the channel, but Bob is still waiting for Alice’s notification. As a
temporary fix, we now only check Alice’s state for the existence of the
channel.
This commit adds some new networkHarness helper methods which are
mean’t to reduce the verbosity of the previous basic tests, and also to
enable developers to right tests mote easily five a higher level
interface.
