Should we just not revert when the array is empty in withdrawMultiple?

[PaulRBerg]

There would be a few benefits:

1. Reduced gas cost (due to there being one if statement fewer)
2. Easier to maintain the tests (due to there being one testing branch fewer)

I can't think of any major downside.

[razgraf]

Looks fine, empty arrays would probably be due to UI bugs, which shouldn't happen too often. The worse they'll do is consume some gas for the non-state-altering tx. Maybe it's worth looking into what others do when expecting this type of arrays as params?

[PaulRBerg]

Yeah I'll try to find an equivalent scenario in OpenZeppelin's code base.

[PaulRBerg]

@andreivladbrg it would be good to spend a little bit of time to see what OpenZeppelin does in this case.

[PaulRBerg]

@andreivladbrg could you please research what OpenZeppelin does with zero-length arrays so we can make a decision about this?

[PaulRBerg]

Yup, you're right, this discussion no longer applies.

Should we add tests for when the array length is zero?

[andreivladbrg]

We could

[PaulRBerg]

We will make a final decision on this topic once we make up our minds regarding the proxy (sablier-labs/v2-periphery#11).

[PaulRBerg]

@andreivladbrg I ended up writing a test in PR #341 for the case when the length of the array is zero in withdrawMultiple.

Once that PR is merged in, we can lock this discussion for good.

[PaulRBerg]

Locking, as #341 has been merged.