Conversation
aalavandhan
force-pushed
the
xc-wampl
branch
3 times, most recently
from
7b7a7ca to
34ed3b8
Compare
| * the xc-ample contracts. | ||
| * | ||
| */ | ||
| contract MaticXCAmpleRebaseGateway is Layer2RebaseGateway, FxBaseChildTunnel { |
There was a problem hiding this comment.
Initially assumed we could abstract all L2s to implement the same gateway interface, however there are operational differences between L2s which prevent this.
eg) Matic doesn't charge a ETH fee to send data from ethereum to matic but Arbitrum does. Arbitrum expects the caller to "pre-pay" the storage and execution cost of the cross chain transaction and thus has some additional parameters.
Leaning toward defining separate gateway interfaces for each bridge.
There was a problem hiding this comment.
Yeah. Unfortunately I think it's unlikely these bridges will be similar enough to abstract away completely.
aalavandhan
force-pushed
the
xc-wampl
branch
from
acd29af to
77bd9b3
Compare
aalavandhan
force-pushed
the
xc-wampl
branch
from
77bd9b3 to
5b3d984
Compare
| IERC20(_l1Token).safeTransferFrom(from, address(this), _amount); | ||
| IERC20(_l1Token).approve(vault, _amount); | ||
|
|
||
| ITokenVault(vault).lock(_l1Token, address(this), _amount); |
There was a problem hiding this comment.
We do this to keep the audit trail on the vault. Explicitly calling lock over just transferring directly into the vault logs the GatewayLocked event.
https://github.com/ampleforth/cross-chain-ample/blob/master/contracts/base-chain/TokenVault.sol#L31
However I wonder if the GatewayLocked and GatewayUnlocked events are redundant as one could always reconstruct the same information using merely the token transfer logs.
If so we can safeTransferFrom the user's wallet directly into the vault which is cleaner and more gas efficient.
2 participants
AMPL Arbitrum Bridge gateways.
To test the UI: