eth: v1 contract #2038

buck54321 · 2023-01-10T18:58:04Z

Replaces #1426. Not tested with simnet-trade-tests (#2037) or loadbot, so leaving in draft for now. Need to get some reviews done.

Implements the version 1 contracts for ethereum and tokens. Based
on feedback in #1426, everything is now encoded in the
"contract data". This "contract data", is the msgjson.Init.Contract
-> msgjson.Audit.Contract -> MatchMetaData.Proof.CounterContract,
AuditInfo.Contract -> Redemption.Spends.Contract.

A few new terms are introduced to differentiate various encodings and
data sets. The aforementioned contract data did encode a version
and a secret hash. It now encodes a version and a "locator", which is
a []byte whose length and content depend on the version. For
version 0, the locator is still just the secretHash[:]. For v1,
the locator encodes all of the immutable data that defines the
swap. This immutable data is now collected in something called
a "vector" (dexeth.SwapVector). For version 0, some vector data
is stored on-chain indexed by the secret hash. For version 1, all
vector data is encoded in the locator.

I've also made an effort to standardize the use of status/step,
and eliminated the use of ambiguous "ver" variables throughout.
A "status" is now the collection of mutable contract data: the step,
the init block height, and the secret. The status and vector
collectively fully characterize the swap state.

client/asset/eth:
New contractV1 and tokenContractorV1 interfaces. To avoid duplication,
the ERC20 parts of the tokenContractors are separated into a new type
erc20Contractor that is embedded by both versions. Getters for
status and vector are added in place of the old method "swap".

assetWallet and embedding types are updated to work with the new
version-dependent locators and the status and vector model.

dex/networks/{eth,erc20}:
New contracts added. New methods for dealing with locators. Simnet
entries added for eth and dextt.eth in the ContractAddresses and Tokens
maps. txDataHandler interace is replaced with versioned package-level
functions.

server/asset/eth:
Server is fully switched to version 1. No option to use version 0.
Translation to new version was straightforward, with one notable
difference that we can no longer get a block height from the
contract once the swap is redeemed.```

chappjc · 2023-01-10T20:03:56Z

client/asset/eth/eth.go

+	spent = status.Step >= dexeth.SSRedeemed
+	confs = uint32(hdr.Number.Uint64() - status.BlockHeight + 1)
+
+	// NOTE: confs will equal to the block number for the version 1 contract if spent == true,
+	// because BlockHeight will be zero. This is probably fine, since the caller
+	// will examine spent first? Otherwise, we could do this.


Interesting problem to have here, but I think it's right to be using what's recorded in the contract state like you're doing in this version of the PR, at least before it's spent. What you suggest about getting the tx confirmations after it is spent might be alright unless the tx was replaced.

chappjc · 2023-01-10T20:05:07Z

client/asset/eth/eth.go

+	// Or maybe it'd be better to set confs to some constant, large, round
+	// number so it doesn't look dumb in logs.


Yah, maybe like 32 or something?

martonp · 2023-04-11T17:05:01Z

There is a new ETH functionality that would allow us to have ETH redemptions for accounts without any ETH:
https://eips.ethereum.org/EIPS/eip-4337

Would be cool to have it in the v1 contract.

norwnd · 2023-07-21T11:21:03Z

There is a new ETH functionality that would allow us to have ETH redemptions for accounts without any ETH:
https://eips.ethereum.org/EIPS/eip-4337

It might also allow for removal of additional "approve" step since multiple user actions can be executed within single transaction (from https://youtu.be/iLf8qpOmxQc?t=2028).

But who knows how long it will take to get this Account Abstraction design working, I bet it's 1+ year down the road.

martonp

Swaps seem to be working fine, but I'm getting some errors on the server:

dex/networks/eth/contracts/ETHSwapV1.sol

martonp · 2024-03-11T19:25:19Z

dex/networks/eth/contracts/ETHSwapV1.sol

+        require(!secretValidates(record, v.secretHash), "swap already redeemed");
+
+        // Is it already refunded?
+        require(record != RefundRecord, "swap already refunded");


require(blockNum > 0 && blockNum <= block.number, "swap not active");

If this check passes then record != RefundRecord will definitely be true.

martonp · 2024-03-11T19:33:09Z

dex/networks/eth/contracts/ETHSwapV1.sol

+    // contractKey generates a key hash which commits to the contract data. The
+    // generated hash is used as a key in the swaps map.
+    function contractKey(Vector calldata v) public pure returns (bytes32) {
+        return sha256(bytes.concat(v.secretHash, bytes20(v.initiator), bytes20(v.participant), bytes32(v.value), bytes8(v.refundTimestamp)));


Suggested change

return sha256(bytes.concat(v.secretHash, bytes20(v.initiator), bytes20(v.participant), bytes32(v.value), bytes8(v.refundTimestamp)));

return keccak256(abi.encode(v));

I'm pretty sure keccak256 will be cheaper, and maybe abi.encode would also be cheaper than the bytes.concat.

I was originally using abi.encodePacked, but we decided to use bytes.encode because of some ambiguity in the implementation of abi encoding, iirc. I can't find the convo. I think it was on Matrix.

I still need to investigate keccak hash

client/asset/eth/eth.go

dex/networks/eth/params.go

server/asset/eth/coiner.go

martonp · 2024-03-13T20:58:01Z

server/asset/eth/coiner.go

+			LockTime:   uint64(init.LockTime.Unix()),
+		}
+
+		// if value < bc.vector.Value {


Should these checks be copied below to use with both versions?

The value check is moved and the others are completely pointless since we just pulled (v0) or validated (v1) the data on-chain. This is also not the purview of the backend, which reports chain data to the swapper for validation.

JoeGruffins

Glad you rebased all of this. The gas savings on everything but Refund is substantial. I think I have a valid comment about the RefundRecord being used as the Secret and maybe some unintended things can happen with that.

client/asset/eth/contractor.go

JoeGruffins · 2024-03-18T05:45:38Z

client/asset/eth/contractor.go

+// func (c *contractorV1) record(v *dexeth.SwapVector) (r [32]byte, err error) {
+// 	abiVec := dexeth.SwapVectorToAbigen(v)
+// 	ck, err := c.ContractKey(&bind.CallOpts{From: c.acctAddr}, abiVec)
+// 	if err != nil {
+// 		return r, fmt.Errorf("ContractKey error: %v", err)
+// 	}
+// 	return c.Swaps(&bind.CallOpts{From: c.acctAddr}, ck)
+// }


Ok to remove?

I though this might come in handy in an emergency for troubleshooting. The underlying Swaps method is generated automatically by abigen anyway.

client/asset/eth/contractor.go

client/asset/eth/eth.go

server/asset/eth/coiner.go

server/asset/eth/tokener.go

JoeGruffins · 2024-03-19T07:59:34Z

server/asset/eth/coiner.go

+			// From: ,
+			To:         init.Participant,


Leaving from out on purpose? Also here is another example of To as Participant, which I think it correct.

We simply don't have this information at this point, and luckily the server doesn't need it. I wasn't sure what else to do but leave it empty.

* tame logs about missing translations

Implements the version 1 contracts for ethereum and tokens. Based on feedback in decred#1426, everything is now encoded in the "contract data". This "contract data", is the msgjson.Init.Contract -> msgjson.Audit.Contract -> MatchMetaData.Proof.CounterContract, AuditInfo.Contract -> Redemption.Spends.Contract. A few new terms are introduced to differentiate various encodings and data sets. The aforementioned contract data did encode a version and a secret hash. It now encodes a version and a "locator", which is a []byte whose length and content depend on the version. For version 0, the locator is still just the secretHash[:]. For v1, the locator encodes all of the immutable data that defines the swap. This immutable data is now collected in something called a "vector" (dexeth.SwapVector). For version 0, some vector data is stored on-chain indexed by the secret hash. For version 1, all vector data is encoded in the locator. I've also made an effort to standardize the use of status/step, and eliminated the use of ambiguous "ver" variables throughout. A "status" is now the collection of mutable contract data: the step, the init block height, and the secret. The status and vector collectively fully characterize the swap. client/asset/eth: New contractV1 and tokenContractorV1 interfaces. To avoid duplication, the ERC20 parts of the tokenContractors are separated into a new type erc20Contractor that is embedded by both versions. Getters for status and vector are added in place of the old method "swap". assetWallet and embedding types are updated to work with the new version-dependent locators and the status and vector model. dex/networks/{eth,erc20}: New contracts added. New methods for dealing with locators. Simnet entries added for eth and dextt.eth in the ContractAddresses and Tokens maps. txDataHandler interace is replaced with versioned package-level functions. server/asset/eth: Server is fully switched to version 1. No option to use version 0. Translation to new version was straightforward, with one notable difference that we can no longer get a block height from the contract once the swap is redeemed.

buck54321 · 2024-03-22T21:24:26Z

server/cmd/dcrdex/evm-protocol-overrides.json

+    "eth": 0,
+    "usdc.eth": 0,
+    "polygon": 0,
+    "usdc.polygon": 0,
+    "wbtc.polygon": 0,
+    "weth.polygon": 0


This is annoying, but until we launch contracts, you'll have to delete these overrides to test v1 contracts.

JoeGruffins · 2024-03-28T07:46:51Z

Is it possible to add tests for the contracts like @martonp has done in #2717 ?

buck54321 changed the title ~~eth contract v1~~ eth: v1 contract Jan 10, 2023

buck54321 mentioned this pull request Jan 10, 2023

eth swap contract v1 #1426

Closed

chappjc reviewed Jan 10, 2023

View reviewed changes

chappjc added this to the 1.0 milestone Jan 11, 2023

chappjc added the ETH label Jan 12, 2023

buck54321 mentioned this pull request Feb 3, 2023

client/asset: clarify eth and token versioning #2094

Merged

buck54321 force-pushed the ethv1-take2 branch from 50ac869 to 47c5942 Compare July 19, 2023 18:39

buck54321 force-pushed the ethv1-take2 branch from 47c5942 to 6860081 Compare August 7, 2023 21:44

buck54321 force-pushed the ethv1-take2 branch from 6860081 to 9f934c8 Compare September 11, 2023 15:10

buck54321 force-pushed the ethv1-take2 branch from 9f934c8 to cd4866c Compare October 27, 2023 00:14

buck54321 force-pushed the ethv1-take2 branch from cd4866c to 3341df8 Compare March 3, 2024 22:42

buck54321 marked this pull request as ready for review March 3, 2024 22:46

martonp reviewed Mar 14, 2024

View reviewed changes

JoeGruffins reviewed Mar 19, 2024

View reviewed changes

buck54321 added 4 commits March 22, 2024 15:55

testing: tame logs about missing translations (decred#2571)

06583f0

* tame logs about missing translations

update server tests and enable per-asset protocol reversion

45c44b0

review followup

76d5a28

buck54321 force-pushed the ethv1-take2 branch from dd65c17 to 76d5a28 Compare March 22, 2024 21:23

buck54321 commented Mar 22, 2024

View reviewed changes

JoeGruffins approved these changes Mar 27, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

eth: v1 contract #2038

eth: v1 contract #2038

buck54321 commented Jan 10, 2023 •

edited

chappjc Jan 10, 2023

chappjc Jan 10, 2023

martonp commented Apr 11, 2023

norwnd commented Jul 21, 2023

martonp left a comment

martonp Mar 11, 2024

martonp Mar 11, 2024

buck54321 Mar 22, 2024

buck54321 Mar 26, 2024 •

edited

martonp Mar 13, 2024

buck54321 Mar 22, 2024

JoeGruffins left a comment

JoeGruffins Mar 18, 2024

buck54321 Mar 22, 2024

JoeGruffins Mar 19, 2024

buck54321 Mar 22, 2024

buck54321 Mar 22, 2024

JoeGruffins commented Mar 28, 2024

		// Or maybe it'd be better to set confs to some constant, large, round
		// number so it doesn't look dumb in logs.

	return sha256(bytes.concat(v.secretHash, bytes20(v.initiator), bytes20(v.participant), bytes32(v.value), bytes8(v.refundTimestamp)));
	return keccak256(abi.encode(v));

eth: v1 contract #2038

Are you sure you want to change the base?

eth: v1 contract #2038

Conversation

buck54321 commented Jan 10, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

martonp commented Apr 11, 2023

norwnd commented Jul 21, 2023

martonp left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

buck54321 Mar 26, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JoeGruffins left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JoeGruffins commented Mar 28, 2024

buck54321 commented Jan 10, 2023 •

edited

buck54321 Mar 26, 2024 •

edited