Skip proof validation when performing estimation #1471
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #1471 +/- ##
==========================================
+ Coverage 80.79% 80.81% +0.02%
==========================================
Files 20 20
Lines 854 855 +1
Branches 148 149 +1
==========================================
+ Hits 690 691 +1
Misses 141 141
Partials 23 23
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| $.inboundNonce.set(message.nonce); | ||
|
|
||
| // Produce the commitment (message root) by applying the leaf proof to the message leaf | ||
| bytes32 commitment = MerkleProof.processProof(leafProof, leafHash); | ||
| if (msg.sender != address(this)) { |
There was a problem hiding this comment.
This should be safe because of the nonreentrant modifier. There is no way to re-enter this code path with the sender address of the gateway proxy. But we may need a test to cover that this can never happen.
For instance, in a separate branch, remove the nonreentrant modifier and maybe try sending Ether to a contract address with receive() implemented to try and reenter the Gateway skip proof verification. If its possible without nonreentrant modifier, and will it still be possible after nonreentrant modifier will give us insight into whether we want to use this method to bypass verification.
Another note is that the choice of the GatewayProxy address is arbitrary and we could of chosen any keyless account for this purpose, but the GatewayProxy is probably the most sane choice.
More context and details in https://linear.app/snowfork/issue/SNO-1397#comment-5693a6c2