Re-ordered the cleanup in handleSweeps so completed batches are deleted only
after any re-add attempt, preventing a finished batch from being dropped before
it can absorb the re-add.
Added a fresh status/parent-batch lookup inside handleSweeps so the logic now
reacts to swaps that complete between AddSweep and the event-loop handler; fully
confirmed sweeps are routed straight to monitorSpendAndNotify instead of
spawning a duplicate batch.
Added regression test testSweepBatcherHandleSweepRace, which replays the
AddSweep/confirmation race and verifies that only the original batch remains and
the sweep keeps its parent assignment.
When doing git-clone to a subdir, do not keep tags, since there can be local
tags affecting `git describe` and buildvcs info. Instead pull the tags from
upstream.
Collect artifacts in "tmp-..." instead of final place. If build fails in the
middle, we don't want to leave the artifacts directory with partial result.
Also remove the build directory (with Git clone) in the end.
Create a temporary directory inside loop/ directory and make a clone of Git repo
to that dir. This ensures that the version built is exactly what is committed
without any unstaged files and not dirty.
To test "make docker-release" it is convenient to pass a regular commit to it
(an output of `git describe`). In this commit release.sh is changed to detect
such "tags" and skip tag signature verification for it.
Also if release.sh is used without an argument, a unique directory name is
generated from the current time.
The Problem:
The test TestPresigned/dust_main_output was causing a deadlock. When the test
called batcher.AddSweep for the second time (with a valid sweep), the batcher
would create a new batch and attempt to register for a spend notification.
However, the test was not waiting for or consuming this notification
registration, leading to a blocked goroutine and a timeout.
The Fix:
I have added a line to the test to consume the spend notification registration
from the mock lnd's channel, similar to how other tests in the same file are
structured. This prevents the deadlock.
The problem was introduced in commit 54652dc641.
If a quote request contains an amount and flag
SelectDeposits set to true the quoting coin-
selects the required deposits to meet the
swap amount in order to quote for the number
of deposits.
In this commit we add a new function SelectDeposits
to the loop-in manager. It coin-selects deposits that
meet an arbitrary swap amount provided by the client.
We have to ensure that the server creates the correct
change outputs for the htlc- and sweepless sweep
transactions.
The QuoteRequest adds a field auto_select_deposits to signal
that the specified quote amount should be coin-selected from
the client's deposits to determine the number of deposits to
quote for.
The StaticAddressLoopInRequest gets a new amount field that
either indicates that the swap amount should be allocated
from the passed deposits or coin-selected from all available
deposits.
The whole project was analyzed by AI and the context was saved to the file.
This file is picked up by agents automatically and it should facilitate AI
understanding of the project.
