Add wstETH collateral support

[valiafetisov]

Goal

Participation in wstETH collateral auctions is possible via the UI

Context

wstETH collateral requires it's own callee contract and uses Uniswap v3 exchange, therefore:

• new callee address should be passed to the clipper contract (along with possibly different callee data)
• the price of the new collateral should be calculated separately

This issue will also be used to refactor callee logic to streamline collateral onboarding.

Tasks

• Refactor callee / collateral logic
  • Import callee addresses from a single file (proposed in Add UniswapV2 contract addresses makerdao/exchange-callees#22)
  • Make a declarative configuration: price calculation function + callee data generation per callee
  • Add collateral onboarding readme
• Add new WstETHCurveUniv3Callee callee support
  • Deploy callee to test networks if needed
  • Add uniswap v3 library
  • Generate callee data
  • Calculate price
• Add new collateral config
• Manually validate that new collateral is supported
  • Manually create a new auction on a supported test chain
  • Execute auth and bidding

[valiafetisov]

Add new callee support

The callee in question is https://github.com/makerdao/exchange-callees/blob/master/src/WstETHCurveUniv3Callee.sol

But I can not find its address, since action-demo-keeper doesn't seem to support this collateral.

Till we get the address from the PR (by merging the PR and asking them to fill the json), I can try to deploy it myself, but it's not clear for me, what constructor parameters should be used. The contract accepts:

• address curvePool_: supposedly CurveContractV1.vy contract from the official docs https://curve.readthedocs.io/ref-addresses.html, but it can not be find on test chains
• address uniV3Router_: supposedly 0xe592427a0aece92de3edee1f18e0157c05861564 (across networks)
• address daiJoin_: MCD_JOIN_DAI from the https://chainlog.makerdao.com
• address weth_: supposedly WETH token address, also from the https://chainlog.makerdao.com

[valiafetisov]

It turns out, I've been looking into a wrong config file inside action-demo-keeper and it actually supports wstETH (but only on the mainnet): https://github.com/makerdao/auction-demo-keeper/blob/bca09252af029448ff67823b0b7825ad8c6feff8/config/mainnet.json#L358-L373

Another interesting, but currently unrelated point: latest mainnet config uses uniswapV2Callee for YFI-A (while all other collaterals are traded via V3 callee) https://github.com/makerdao/auction-demo-keeper/blob/bca09252af029448ff67823b0b7825ad8c6feff8/config/mainnet.json#L189

[valiafetisov]

Based on the callee deployed on the main, constructor parameters are:

parameter	description	mainnet	kovan	goerli
curvePool	StableSwapSTETH.vy contract https://curve.readthedocs.io/...	0xdc24316b9ae028f1497c275eb9192a3ea0f67022	unknown	unknown
uniV3Router	Same on all networks	0xe592427a0aece92de3edee1f18e0157c05861564	0xe592427a0aece92de3edee1f18e0157c05861564	0xe592427a0aece92de3edee1f18e0157c05861564
daiJoin	MCD_JOIN_DAI from the https://chainlog.makerdao.com	0x9759A6Ac90977b93B58547b4A71c78317f391A28	0x5AA71a3ae1C0bd6ac27A1f28e1415fFFB6F15B8c	0x9759A6Ac90977b93B58547b4A71c78317f391A28 (not sure if it's a valid goerli address)
weth	ETH from the https://chainlog.makerdao.com	0xC02aaA39b223FE8D0A0e5C4F27eAD9083C756Cc2	0xd0A1E359811322d97991E03f863a0C30C2cF029C	0xC02aaA39b223FE8D0A0e5C4F27eAD9083C756Cc2

Unfortunately, StableSwapSTETH.vy doesn't seem to be deployed anywhere except main.

[valiafetisov]

@LukSteib can you maybe ask PE if they are aware of WstETHCurveUniv3Callee or the StableSwapSTETH addresses on the test chains? Or did they only test it on the mainnet/fork? Otherwise, we would need to prioritise #81 sooner than later to actually test something.

[LukSteib]

Or did they only test it on the mainnet/fork?

Yes, testing is based on the mainnet/fork approach and not on the testnet based approach.

Otherwise, we would need to prioritise #81 sooner than later to actually test something.

Given the above, I agree that we need to prioritise that. IMO the minimal testing setup that we requested from chaoslabs (ie. liquidation of one urn of a specific ilk at a time) would already help here.

[valiafetisov]

Yes, testing is based on the mainnet/fork approach and not on the testnet based approach.

I tried forking mainnet a few blocks before active wstETH auction was taken in block 14052147, but it doesn't seem to work so far (it's not being fetched from the clipper).