A 0 may mean the function call will revert or that there is no liquidity available
For all intents and purposes, check if a quote returns 0 and if it does, do not execute it or expect a revert
A BadgerDAO sponsored repo of Open Source Contracts for:
- Calculating onChain Prices
- Executing the best onChain Swap
We understand that we cannot prove a optimal price because at any time a new source of liquidity may be available and the contract cannot adapt.
However we believe that given a set of constraints (available Dexes, handpicked), we can efficiently compute the best trade available to us
In exploring this issue we aim to:
- Find the most gas-efficient way to get the best executable price (currently 120 /150k per quote, from 1.6MLN)
- Finding the most reliable price we can, to determine if an offer is fair or unfair (Cowswap integration)
- Can we create a "trustless swap" that is provably not frontrun nor manipulated?
- How would such a "self-defending" contract act and how would it be able to defend itself, get the best quote, and be certain of it (with statistical certainty)
Given a tokenIn, tokenOut and AmountIn, returns a Quote from the most popular dexes
OnChainPricingMainnet-> Fully onChain math to find best, single source swap (no fragmented swaps yet)OnChainPricingMainnetLenient-> Slippage tollerant version of the Pricer
- Curve
- UniV2
- UniV3
- Balancer
- Sushi
Covering >80% TVL on Mainnet. (Prob even more)
V4 adds support for Chainlink Price Feeds, all feeds are supported via the Feeds Registry
Because V4 marks the separation between "executable prices" and "ideal prices" we added new functions.
Read below for full details
V4 functions are view
!!! V4 functions do not revert on error, they return 0 !!!
Returns true if the pricer will return a non-zero quote NOTE: This is not proof of optimality
/// @dev Given tokenIn, out and amountIn, returns true if a quote will be non-zero
/// @notice Doesn't guarantee optimality, just non-zero
function isPairSupported(address tokenIn, address tokenOut, uint256 amountIn) external returns (bool)In Brownie
quote = pricer.isPairSupported(t_in, t_out, amt_in)Finds the best quote available between the sources Prioritizes price feeds
function findOptimalSwap(address tokenIn, address tokenOut, uint256 amountIn) external virtual returns (Quote memory)In Brownie
quote = pricer.findOptimalSwap(t_in, t_out, amt_in)Finds the best executable quote Uses PriceFeeds (if available) to verify the quote is better than the feed
function findExecutableSwap(address tokenIn, address tokenOut, uint256 amountIn) external virtual returns (Quote memory)In Brownie
quote = pricer.findExecutableSwap(t_in, t_out, amt_in)Finds the best executable quote Doesn't check price feeds, use at your own risk
function unsafeFindExecutableSwap(address tokenIn, address tokenOut, uint256 amountIn) external virtual returns (Quote memory)In Brownie
quote = pricer.unsafeFindExecutableSwap(t_in, t_out, amt_in)Variation of Pricer with a slippage tollerance, mostly used to allow a multisig enough wiggle room for operation
These tests compare the PricerV3 (150k per quote) against V2 (1.6MLN per quote)
brownie test tests/heuristic_equivalency/test_heuristic_equivalency.py
TODO: Improve to just use the specific quote
brownie test tests/gas_benchmark/benchmark_pricer_gas.py --gas
TODO: Add like 200 tokens TODO: Compare against Coingecko API or smth
brownie test tests/gas_benchmark/benchmark_token_coverage.py --gas
Run V3 Pricer against V2, to confirm results are correct, but with gas savings
brownie test tests/heuristic_equivalency/test_heuristic_equivalency.py