fix(protocol): fix bridge bugs in getMessageMinGasLimit

[dantaik]

Fix two bugs in Bridge.getMessageMinGasLimit reported by OZ:

The abi encoding of A = (Message calldata msg) is actually 10 * 32 bytes (not 7 * 32 as elements are not packed) + 32 bytes (A is a dynamic tuple, offset to first elements) + 32 bytes (offset to last bytes element of Message ) + 32 bytes (padded encoding of length of Message.data + dataLength (padded to 32 bytes) = 13 * 32 + (dataLength // 32 * 32) + 32.

For example, if dataLength = 4, the actual abi encoding is 448 bytes long, whereas dataLength + 256 which is the formula used in the code gives only 260 bytes long (code here)

I don't understand the division by 16: I assume the 16 comes from the gas cost of a nonzero calldata byte, but then why divide the number of bytes by 16 instead of multiplying ?

[adaki2004]

Seems OK, but we shall immediately try it out on testnet to see in action - as currently (even tho seems inaccurate, it works :) )

And 2 tests (related to gas) are failing.

[dantaik]

Seems OK, but we shall immediately try it out on testnet to see in action - as currently (even tho seems inaccurate, it works :) )

And 2 tests (related to gas) are failing.

Fixed

[cyberhorsey]

Seems OK, but we shall immediately try it out on testnet to see in action - as currently (even tho seems inaccurate, it works :) )

And 2 tests (related to gas) are failing.

it works in testnet because we are multiplying the return value offchain

[openzeppelin-code]

fix(protocol): fix bridge bugs in getMessageMinGasLimit

Generated at commit: 5daf562ed7d7fbeee08f04c9f2bd682e1012e7c0

🚨 Report Summary

	Severity Level	Results
Contracts	Critical High Medium Low Note Total	2 2 0 8 41 53
Dependencies	Critical High Medium Low Note Total	0 0 0 0 0 0

---

For more details view the full report in OpenZeppelin Code Inspector