Remove further hardcoded function resolution

[frangio]

For some reason I missed these in #4299.

The one in ERC721Enumerable._removeTokenFromOwnerEnumeration is interesting and deserves attention. We were using ERC721.balanceOf because we are more confident that it's the length of the enumeration than when we use the potentially overridden balanceOf. I've just made it call balanceOf now because I don't see a way around this, other than introducing a storage inefficiency by storing the length which is redundant with the balance mapping.

---

For the record here is how I am checking that these changes are exhaustive. (rg = ripgrep)

rg '^(abstract )?contract (\w+).*' contracts -I -g '*.sol' --replace '\b$2\.\w+\(' > patterns
rg -f patterns -g '*.sol'

The only remaining ones are ERC2771Context._msgData() and ERC2771Context._msgSender() in a mock, but without this we get a compiler warning.

[changeset-bot]

⚠️ No Changeset found

Latest commit: 29ed898

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes changesets to release 1 package

Name	Type
openzeppelin-solidity	Major

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[Amxx]

LGTM.

There are many restrictions to combining the ERC721Enumerable module with "custom" balanceOf. This is why this module is not compatible with ERC721Concecutive (for example).
Maybe we will want to clarify the assumption that this module makes.

[ernestognw]

I also agree with with adding a warning about the risks of using custom balanceOf

[ernestognw]

Merging since it already implements the suggestion @Amxx and I mentioned.