Fix ERC20._update

[JulissaDantes]

There is a risk with the current implementation of the _update function of altering the tokens total supply when sending the zero address as the from and to parameters.

If someone does _update(address(0), address(0), amount), this will go into
if (from == address(0)) { ... }
that will increase the total supply and mint tokens to address(0), then because of the else it will not go into
else if (to == address(0)) { ... }
so the total supply amount would remain unbalanced.

PR Checklist

• Tests
• Documentation
• Changelog entry

[Amxx]

ERC20Snapshot should also be updated. Should that be done in a separate PR?

[JulissaDantes]

ERC20Snapshot should also be updated. Should that be done in a separate PR?

It will be done here.

[Amxx]

We should merge master into next-5.0.

This will enable hardhat-exposed, removing the needs for mocks.

[frangio]

Really confused by why the linter is still failing. It doesn't show any errors locally.

[Amxx]

Really confused by why the linter is still failing. It doesn't show any errors locally.

same here

[frangio]

Adding unchecked here is the only pending item for this PR.

[frangio]

I was somehow able to fix linting by running the latest version of Prettier with the new configuration.

[Amxx]

I was somehow able to fix linting by running the latest version of Prettier with the new configuration.

is it possible that github action has a cache with the old version ?

[frangio]

I tried deleting the caches and it didn't make a difference.

[Amxx]

Note that linting of the mocks is irrelevant as they will be deleted when we merge master.