* initial commit for splitting payouts
Collateral goes to slot's host address, while reward payouts go to the slot's host payoutAddress
* Add fillSlot overload to make payoutAddress "optional"
* add tests for payoutAddress
* add doc to patchFillSlotOverloads
* formatting
* remove optional payoutAddress parameter
* Move payoutAddress to freeSlot
- remove payoutAddress parameter from `fillSlot`
- remove `payoutAddress` from slot struct and storage
- add payoutAddress parameter to `freeSlot`, preventing the need for storage
* formatting
* update certora spec to match updated function signature
* Add withdrawAddress to withdrawFunds
- prevent erc20 msg.sender blacklisting
* Update tests for paying out to withdrawAddress
* formatting
* Add collateralRecipient
* refactor: change withdrawFunds and freeSlot overloads
- `withdrawFunds` now has an option withdrawRecipient parameter
- `freeSlot` now has two optional parameters: rewardRecipient, and collateralRecipient. Both or none must be specified.
* update certora spec for new sigs
This adds a rule to formally verify that the `SlotState` and
`RequestState` of any given `Slot` or `Request` does not change more
than once per function call.
Closes#129
This invariant ensures that the total supply of the used token in the
contract is always greater equal to the sum of all balances within the
token combined.
This commit adds CVL rule that formally verifies the state changes of
any given request in relation to the functions of the contract that can
cause them.
Closes#128
Rationale: subtracting 2000 from the provided gas seems
arbitrary, and doesn't provide any benefits. Whether
verify() fails with an out-of-gas error, or returns
'false', in both cases the proof is not verified.
Co-Authored-By: Balazs Komuves <bkomuves@gmail.com>
Reason: Circom and Ethereum EIP-197 both represent these
elements as arrays of two elements, but they do it in
reverse order of each other. This change makes it explicit
which number is the real part, and which number is the
imaginary part.
In Codex's integration tests we now create real ZK proofs,
which take a bit longer to generate. We therefore need a
period that remains the same while the proof is generated.
Andrea Franz
Eric
r4bbit
Adam Uhlíř
Slava
Giuliano Mega
Jaremy Creechley
Ben Bierens
Mark Spanbroek