Failed to find coin for address re-use in regtest

[ffranr]

This issue tracks a bug reported here: jamaljsr/polar#775 (comment)

Some aspects that I've found interesting whilst investigating so far:

• This bug was encountered in Polar.
• The relevant LND wallet in Polar manages a single UTXO where as in our integration tests each wallet manages ~100 UTXOs.
• This issue/bug may only be relevant in a regtest environment where mining can happen significantly faster than on mainnet.

Developers who have encountered this issue are requesting a new feature which will allow them to customise the coin lock time in a regtest environment. Or to observe minting rather than a set lock time. See this request: jamaljsr/polar#775 (comment)

This is the coin lock/lease 10 min constant:
https://github.com/ffranr/taproot-assets/blob/main/tapfreighter/wallet.go#L35-L37

A similar issue was reported here: https://lightningcommunity.slack.com/archives/C03B3556HQ8/p1694700896132179

[ffranr]

Attempt at reproducing bug: #515

[ffranr]

I've spoken again with @jamaljsr . He mentioned that there was a delay message between send events. And that the coin locking error message was produced whilst attempting to kick off send during that delay.

I'm pretty certain that the delay was caused by the proof courier's backoff procedure erroneously resuming (with a delay) from the stale state of a previous send. This is what I've observed via debugging.

Not too long ago we merged this commit into main: 328e909 It stops the itests from complaining when the courier's backoff procedure resumes from a bad state.

I've created this PR to fix the courier's backoff procedure initial state: #529

I think that the coin locking error message is valid and doesn't need to be changed. If the coin was locked in a send event (which was ongoing because of the erroneous courier delay) then it was correct to raise the error message it did.