Uniswap V3 Split Router Callee

[KirillDogadin-std]

This PR adds new Uniswap Split Route callee contract and its tests.

Notes:

• The contract was developed using foundry
• Since Split route also requires data provided by the Uniswap AlphaRouter to operate we:
  • Hardcoded example data to the test contract
  • Provided node.js script to fetch fresh data and then pass it to the same test via env variables

Functional testing

1. Create .env file (in the root of the project) containing valid rpc url under FOUNDRY_ETH_RPC_URL name
2. Run tests
   • Using hardcoded alpha router data: forge test --match testTakeLinkUniswapSplitProfit
   • Using freshly fetched alpha router data: (cd scripts/uniswap-split-route-callee && npm ci && node index.js)

[rockyfour]

@KirillDogadin-std @valiafetisov in the 1Inch integration the actual collateral slice amount and owed debt to be paid back are re-encoded in the callee itself - https://github.com/makerdao/exchange-callees/blob/1f9ba7b0e5d202774212c8ccfb9b0ac929ee9124/src/OneInchCallee.sol#L131-L132

Here this is not done. So in practice, the slice parameter, which is calculated in the clipper is ignored, and you might end up not swapping all the collateral you got from the clipper.

Was the approach from 1Inch changed intentionally?

[KirillDogadin-std]

https://github.com/makerdao/exchange-callees/pull/35#issuecomment-1457771576

Was the approach from 1Inch changed intentionally?

the calldata that is generated by AlphaRouter has a non-trivial structure and is supposed to be (expected to be) highly optimised for the specific input amount. This means we should choose either or:

• do repackage the data, write several not-easy to understand and implement operations to unpack the data (of non-fixed size) and validate that slice amount matches the cumulative sum of routes' input.
• assume that the wallet that calls the contract is responsible for supplying the matching (matching to the Universal Router calldata) slice value, keep the contract simple and less expensive.

The latter was preferred according to https://github.com/makerdao/exchange-callees/pull/34#discussion_r1066712189

[rockyfour]

Almost there. I do see this failure (let's ignore other instabilities for the purpose of merging this): [FAIL. Reason: STF] testTakeLinkUniswapSplitProfit() (gas: 8260641125390436681) this was on block 16941906

Will approve after that test is fixed.

[KirillDogadin-std]

So this test fails sometimes due to the link price fluctuation (solidity test determines the link amount used in the test based on dust and link price at a block height). With the hardcoded amount there's a risk that eventually the test will fail either with STF or InsufficientProfit. To avoid this, the initial dynamic computation of the link amount was introduced.

I've adjusted the amount of link that the script provides, it works out now, but i am not even sure that when you will take a look at this next time it will pass.

If you have a preference on how to address this (if this should be addressed) please let me know.

[rockyfour]

I've adjusted the amount of link that the script provides, it works out now, but i am not even sure that when you will take a look at this next time it will pass.

Ok, you didn't update the test itself with the new txData, so I added a suggestion for it above.

With that change the test passes (block 16959704). Note that I didn't go deeply into the new amount you set and whether it could be made more robust. You can consider looking at that further and maybe making changes in a separate PR if needed.

Please commit the above suggestion when you can, then I will approve and merge the PR.

[KirillDogadin-std]

Committed, thanks for including suggestion.