Chain swaps #51

dangeross · 2024-06-14T12:48:48Z

This PR adds Chain Swaps. It should include:

Full Chain swap Boltz API.
Claiming cooperatively / non-cooperatively, including posting the claim tx details to exchange partial signatures.
Refunding cooperatively / non-cooperatively.

Tested:

LBTC-BTC claim cooperative flow
LBTC-BTC refund cooperative/non-cooperative flow
BTC-LBTC claim cooperative flow
BTC-LBTC refund cooperative/non-cooperative flow

i5hi · 2024-06-22T14:32:49Z

Massive!

Thanks @dangeross !

Will review over the weekend.

i5hi · 2024-06-24T04:12:01Z

@dangeross couldnt' find tests for this. Attempting to put one together now.

dangeross · 2024-06-24T04:59:50Z

couldnt' find tests for this. Attempting to put one together now.

Thanks @i5hi, let me know if I can help at all.

It would be cool if @michael1011 could confirm the boltz states used in the chain swaps.

i5hi · 2024-06-24T05:49:31Z

@dangeross When creating a ChainSwap from BTC to LBTC we would need to use the BtcSwapScript/TxV2 structs right?

i5hi · 2024-06-24T06:07:48Z

Additionally, its a bit tricky for the user of the library to know who part of the ChainSwapResponse to use as ChainSwapDetails in this function:

 pub fn chain_from_swap_resp(
        side: Side,
        chain_swap_details: ChainSwapDetails,
        our_pubkey: PublicKey,
        )

There is a claim and lockup - ChainSwapDetails - so maybe better for this function to take the ChainSwapResponse object and let the method handle which of the two ChainSwapDetails to use?

dangeross · 2024-06-24T06:43:34Z

When creating a ChainSwap from BTC to LBTC we would need to use the BtcSwapScript/TxV2 structs right?

The lockup side is BtcSwapScript, the claim side is LBtcSwapScript

dangeross · 2024-06-24T06:48:17Z

so maybe better for this function to take the ChainSwapResponse object and let the method handle which of the two ChainSwapDetails to use?

Could work, maybe also having two wrapper functions, one for claim and one for lockup. E.g. chain_claim_from_swap_resp and chain_lockup_from_swap_resp

dangeross · 2024-06-24T07:22:34Z

If it helps @i5hi, here is our current usage. I think having 2 functions is a better approach, then passing the Side is not required. But the Side or maybe details_type Claim/Lockup need to be stored to switch the order of the musig key aggregation
https://github.com/breez/breez-liquid-sdk/blob/bcb47432608f8b91e40a1806b47c1f496ccc76c6/lib/core/src/model.rs#L369-L403

i5hi · 2024-06-25T04:39:43Z

@dangeross The sign_claim methods now use a new cooperative Cooperative :

    pub fn sign_claim(
        &self,
        keys: &Keypair,
        preimage: &Preimage,
        absolute_fees: Amount,
        is_cooperative: Option<Cooperative>,
    ) -> Result<Transaction, Error>

The updated tests in liquid/bitcoin_v2.rs have been updated as follows:

                        let tx = claim_tx
                            .sign_claim(
                                &our_keys,
                                &preimage,
                                Amount::from_sat(1000),
                                Some(Cooperative {
                                    boltz_api: &boltz_api_v2,
                                    swap_id: swap_id.clone(),
                                    pub_nonce: None,
                                    partial_sig: None,
                                }),
                            )
                            .unwrap();

This would fail as pub_nonce and partial_sig are not provided.

I went through the breez code base to understand how to use this properly. Can you confirm that this would be correct usage?

                        let claim_tx = LBtcSwapTxV2::new_claim(
                           claim_script.clone(),
                           claim_address.clone(),
                           &ElectrumConfig::default_liquid(),
                           BOLTZ_TESTNET_URL_V2.to_string(),
                           swap_id.clone(),
                       )
                       .unwrap();
                       let refund_tx = BtcSwapTxV2::new_refund(
                           lockup_script.clone(),
                           &refund_address,
                           &ElectrumConfig::default_bitcoin(),
                       )
                       .unwrap();
                       let claim_tx_response =
                           boltz_api_v2.get_chain_claim_tx_details(&swap_id).unwrap();
                       let (partial_sig, pub_nonce) = refund_tx
                           .partial_sig(
                               &our_refund_keys,
                               &claim_tx_response.pub_nonce,
                               &claim_tx_response.transaction_hash,
                           )
                           .unwrap();
                       let tx = claim_tx
                           .sign_claim(
                               &our_claim_keys,
                               &preimage,
                               Amount::from_sat(1000),
                               Some(Cooperative {
                                   boltz_api: &boltz_api_v2,
                                   swap_id: swap_id.clone(),
                                   pub_nonce: Some(pub_nonce),
                                   partial_sig: Some(partial_sig),
                               }),
                           )
                           .unwrap();

i5hi · 2024-06-25T04:41:13Z

I've noticed this actually won't be an issue with LN swaps

            let partial_sig_resp = match self.swap_script.swap_type {
                SwapType::Chain => match (pub_nonce, partial_sig) {
                    (Some(pub_nonce), Some(partial_sig)) => boltz_api.post_chain_claim_tx_details(
                        &swap_id,
                        preimage,
                        pub_nonce,
                        partial_sig,
                        ToSign {
                            pub_nonce: claim_pub_nonce.serialize().to_lower_hex_string(),
                            transaction: claim_tx_hex,
                            index: 0,
                        },
                    ),
                    _ => Err(Error::Protocol(
                        "Chain swap claim needs a partial_sig".to_string(),
                    )),
                },
                SwapType::ReverseSubmarine => boltz_api.get_reverse_partial_sig(
                    &swap_id,
                    &preimage,
                    &claim_pub_nonce,
                    &claim_tx_hex,
                ),
                _ => Err(Error::Protocol(format!(
                    "Cannot get partial sig for {:?} Swap",
                    self.swap_script.swap_type
                ))),
            }?;

So i will not change the the LN swap tests. Let me know if the previous comment is the correct flow for chain swaps.

dangeross · 2024-06-25T04:50:27Z

@i5hi the pub_nonce and partial_sig are only used in the chain swap claim where there is this double exchange of nonce's/sig's. So all other swap types work without these set in Cooperative

i5hi · 2024-06-25T05:05:44Z

And the pub_nonce and partial_sig are from the refund_tx right?

i5hi · 2024-06-25T05:06:26Z

And when refunding they would be from the claim_tx?

dangeross · 2024-06-25T05:36:17Z

And the pub_nonce and partial_sig are from the refund_tx right?

Yes

And when refunding they would be from the claim_tx?

They are only needed for the claim flow

michael1011

Great stuff! Nice to see that you managed to make it work even with our lackluster documentation (coming soon 🤞)

Left some minor comments. Some test cases would be nice too

src/swaps/bitcoin.rs

src/swaps/boltzv2.rs

src/swaps/bitcoinv2.rs

michael1011 · 2024-06-25T15:04:50Z

src/swaps/bitcoinv2.rs

@@ -888,8 +1015,9 @@ impl BtcSwapTxV2 {
    /// Multiply the size by the fee_rate to get the absolute fees.
    pub fn size(&self, keys: &Keypair, preimage: &Preimage) -> Result<usize, Error> {
        let dummy_abs_fee = 5_000;
+        // Can only calculate non-coperative claims


In case there is a need for calculating those, schnorr signatures are always 64 bytes. Could be stubbed

michael1011 · 2024-06-25T15:07:57Z

src/swaps/boltz.rs

+                "transaction.server.confirmed".to_string()
+            }
+            ChainSwapStates::TransactionClaimed => "transaction.claimed".to_string(),
+            ChainSwapStates::TransactionClaimPending => "transaction.claim.pending".to_string(),


transaction.claim.pending is for submarine swaps, when we paid the invoice but wait before broadcasting the claim transaction in case you happen to come online to do a key path spend with us

In chain swaps you either:

key path spend with us and both claim transactions, ours and yours, are broadcasted

you enforce via script path and we do the same thing

So, that status doesn't need to be here

src/swaps/liquidv2.rs

michael1011 · 2024-06-25T15:15:10Z

src/swaps/liquidv2.rs

+            SwapType::Chain => {
+                boltz_client
+                    .get_chain_txs(swap_id)?
+                    .user_lock
+                    .ok_or(Error::Protocol(
+                        "No user_lock transaction for Chain Swap available".to_string(),
+                    ))?
+                    .transaction
+                    .hex
+            }
+            SwapType::ReverseSubmarine => boltz_client.get_reverse_tx(swap_id)?.hex,
+            SwapType::Submarine => boltz_client.get_submarine_tx(swap_id)?.hex,


This is getting a little confusing. For submarine and chain swaps this would fetch the lockup transaction of the user. For reverse swaps the one of the server. How about returning a struct with optional user and server transaction?

Updated the function name to fetch_lockup_utxo_boltz does that make it more clear?

Added chain_swaps.rs to tests

Minor fixes

Co-authored-by: michael1011 <[email protected]>

i5hi · 2024-06-27T11:50:44Z

Damn! I merged michael's update to the function name and it broke the test because its usage needs to be updated.

Sorry guys! I will just merge and fix it on trunk immediately.

Can I get an ACK on doing it this way?

Or @dangeross would you be able to fix it from your branch?

Or recommend another strategy?

dangeross · 2024-06-27T11:57:15Z

ACK @i5hi. Feel free to commit to the PR

i5hi · 2024-06-27T12:28:23Z

Patched on trunk!
Cheers!

dangeross added 3 commits June 14, 2024 14:35

Implement Chain Swaps

8bc250e

Remove redundant param

d61d690

Add side to SwapScriptV2

95a794c

dangeross marked this pull request as ready for review June 20, 2024 08:37

dangeross changed the title ~~[Draft] Chain swaps~~ Chain swaps Jun 20, 2024

i5hi self-assigned this Jun 23, 2024

i5hi requested a review from michael1011 June 24, 2024 05:09

michael1011 reviewed Jun 25, 2024

View reviewed changes

i5hi and others added 8 commits June 26, 2024 15:17

Added chain_swaps.rs to tests

0aed56f

Merge pull request #1 from i5hi/chain-swaps

2674163

Added chain_swaps.rs to tests

Merge upstream trunk

1db00ad

Minor fixes

1ea1c57

Merge pull request #2 from i5hi/chain-swaps

6b99bd1

Minor fixes

Update src/swaps/bitcoinv2.rs

bfcd94c

Co-authored-by: michael1011 <[email protected]>

Update src/swaps/bitcoinv2.rs

9ef1405

Co-authored-by: michael1011 <[email protected]>

Update src/swaps/bitcoinv2.rs

fc5a2c9

Co-authored-by: michael1011 <[email protected]>

i5hi and others added 2 commits June 27, 2024 17:06

Update src/swaps/bitcoinv2.rs

9219fa9

Co-authored-by: michael1011 <[email protected]>

Update src/swaps/liquidv2.rs

af95a98

Co-authored-by: michael1011 <[email protected]>

i5hi merged commit 2cbde83 into SatoshiPortal:trunk Jun 27, 2024
3 of 4 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Chain swaps #51

Chain swaps #51

dangeross commented Jun 14, 2024 •

edited

Loading

i5hi commented Jun 22, 2024

i5hi commented Jun 24, 2024

dangeross commented Jun 24, 2024

i5hi commented Jun 24, 2024

i5hi commented Jun 24, 2024 •

edited

Loading

dangeross commented Jun 24, 2024

dangeross commented Jun 24, 2024

dangeross commented Jun 24, 2024

i5hi commented Jun 25, 2024

i5hi commented Jun 25, 2024

dangeross commented Jun 25, 2024

i5hi commented Jun 25, 2024

i5hi commented Jun 25, 2024

dangeross commented Jun 25, 2024 •

edited

Loading

michael1011 left a comment

michael1011 Jun 25, 2024

michael1011 Jun 25, 2024

michael1011 Jun 25, 2024

i5hi Jun 26, 2024

i5hi commented Jun 27, 2024

dangeross commented Jun 27, 2024

i5hi commented Jun 27, 2024

Chain swaps #51

Chain swaps #51

Conversation

dangeross commented Jun 14, 2024 • edited Loading

i5hi commented Jun 22, 2024

i5hi commented Jun 24, 2024

dangeross commented Jun 24, 2024

i5hi commented Jun 24, 2024

i5hi commented Jun 24, 2024 • edited Loading

dangeross commented Jun 24, 2024

dangeross commented Jun 24, 2024

dangeross commented Jun 24, 2024

i5hi commented Jun 25, 2024

i5hi commented Jun 25, 2024

dangeross commented Jun 25, 2024

i5hi commented Jun 25, 2024

i5hi commented Jun 25, 2024

dangeross commented Jun 25, 2024 • edited Loading

michael1011 left a comment

Choose a reason for hiding this comment

michael1011 Jun 25, 2024

Choose a reason for hiding this comment

michael1011 Jun 25, 2024

Choose a reason for hiding this comment

michael1011 Jun 25, 2024

Choose a reason for hiding this comment

i5hi Jun 26, 2024

Choose a reason for hiding this comment

i5hi commented Jun 27, 2024

dangeross commented Jun 27, 2024

i5hi commented Jun 27, 2024

dangeross commented Jun 14, 2024 •

edited

Loading

i5hi commented Jun 24, 2024 •

edited

Loading

dangeross commented Jun 25, 2024 •

edited

Loading