Set anti-fee-sniping nLockTime in payjoin-cli

[bc1cindy]

compute the anti-fee-sniping nLockTime in the cli (current tip, ~10% backdated up to 100, 0 on stale/IBD) and set it explicitly on wallet_create_funded_psbt, mirroring Core's DiscourageFeeSniping

payjoin-cli nLockTime then matches Core/Electrum anonymity set.

part of #1676 and #1597

[coveralls]

Coverage Report for CI Build 28193013536

Coverage increased (+0.04%) to 85.555%

Details

• Coverage increased (+0.04%) from the base build.
• Patch coverage: 1 uncovered change across 1 file (50 of 51 lines covered, 98.04%).
• No coverage regressions found.

Uncovered Changes

File	Changed	Covered	%
payjoin-cli/src/app/wallet.rs	51	50	98.04%

Coverage Regressions

No coverage regressions found.

---

Coverage Stats

Relevant Lines:	15175
Covered Lines:	12983
Line Coverage:	85.56%
Coverage Strength:	364.83 hits per line

---

💛 - Coveralls

[benalleng]

The cli uses bitcoind-async-client so the default will follow that instead.
https://github.com/alpenlabs/bitcoind-async-client/blob/5af72a9473105275af53c3c696d784988f0a74c2/src/client/v30.rs#L311

We either need to change this upstream in their library or re-implement this behavior to follow core ourselves.

[bc1cindy]

thanks everyone.

reworked to fund via Core's send RPC (add_to_wallet=false), which applies the native anti-fee-sniping nLockTime, with no re-implementation in the cli

depends on alpenlabs/bitcoind-async-client#126

I'll leave as a draft (CI red) until that releases, then I'll bump the dep.

[benalleng]

I would encourage you to just pin the bitcoind-async-client hash to your repo and commit bc1cindy/bitcoind-async-client@d7ff324

However, I do not think using send is going to work here as it is my understanding that this will create and broadcast the tx not leaving it open for us to modify in the payjoin I see you added the add_to_wallet flag so perhaps this could work though It does seem rather hacky still IMO

[benalleng]

The core issue I opened bitcoin/bitcoin#34987 received mixed feedback as other people seemed to believe that this was the desired behavior as to leave locktime default creation up to us when using wallet_create_funded_psbt and any non-final tx creation process

[bc1cindy]

thanks @benalleng

so, given the Core feedback, I lean to switch this to re-implement, since the caller should own the locktime for non-final tx creation (same as Liana)

compute the anti-fee-sniping locktime (current tip, ~10% backdated ≤100, 0 on stale/IBD) and set it explicitly on wallet_create_funded_psbt

sounds good?

[benalleng]

https://github.com/bitcoin/bitcoin/blob/ea9afb61a1c52979d0e64e4dae9b3a611f7b9a5e/src/wallet/spend.cpp#L993

https://github.com/bitcoin/bitcoin/blob/ea9afb61a1c52979d0e64e4dae9b3a611f7b9a5e/src/wallet/spend.cpp#L979

Adding these to just give myself some better reference for review