DON'T MERGE: tmp/log increase #31
Conversation
There was a problem hiding this comment.
As we can´t processExits during the same block due to this, I'd suggest we protect the new unprotected predicate requiring the caller to be "us".
It is true that we are "implicitly removing its usage" here as part of the first batch, but just in case.
Then we can delay its removal to the following block after the exit has been processed.
wdyt?
I see what you mean. It should never happen that someone else calls this, but just to be extra sure I can require this predicate to be called by the multisig. |
There was a problem hiding this comment.
LGTM
Just a comment, in the fork test AFAIK we could assert this between both timelock calls to ensure we get the erc20predicate back before 2nd call, isn't it?
I mean this
address newOldPredicate = registry.erc20Predicate();
assertEq(currentPredicate, newOldPredicate);
|
There was a problem hiding this comment.
LGTM
Here some suggestions:
-
To add to this PR, a gas report this time (including estimate gas) considering the issue we had. This could become a best practice from now on as well if that is ok with you all.
-
To add a comment in the new predicate about why that specific funds need special release, we can link the proof and the errored tx or something. So it might be easier during signing process as well as for brand reputation.
|
I don't think gas is important this time, because the relevant contract is not being touched. |
|
Completed |
No description provided.