Move `to != owner` check from `ERC721.approval` to `ERC721._approval`

[Amxx]

In ERC721.approve (public) we do

require(to != owner, "ERC721: approval to current owner");

we don't do it in ERC721._approve (internal)

That means its not ok to approve the owner, unless you do it internally ? IMO the check should move from the public to the internal function. (note that there is already a call to ERC721.ownerOf inside the internal version, so its not like we are saving an sload anyway)

[boxingcoder00]

This may be a foolish question(very new to the space) but what exactly would be the benefit? gas or optimization ?

[ernestognw]

This may be a foolish question(very new to the space) but what exactly would be the benefit? gas or optimization ?

It's just a matter of interface consistency, we expect developers to have reasonable guarantees when they use the library. I think it's reasonable to expect this same check in both the public and internal interface.

[rupak21]

hey i want to work on this issue`

function _approval(address owner, address to, uint256 tokenId) internal virtual {
require(to != owner, "ERC721: approval to current owner");
_tokenApprovals[tokenId] = to;
emit Approval(owner, to, tokenId);
}
`

[Amxx]

hey i want to work on this issue`

function _approval(address owner, address to, uint256 tokenId) internal virtual { require(to != owner, "ERC721: approval to current owner"); _tokenApprovals[tokenId] = to; emit Approval(owner, to, tokenId); } `

There are already 2 PRs for this issue.

Also, the code you shared above changes the internal interface, which is not something we want.

[Saksham010]

hey i want to work on this issuefunction _approval(address owner, address to, uint256 tokenId) internal virtual { require(to != owner, "ERC721: approval to current owner"); _tokenApprovals[tokenId] = to; emit Approval(owner, to, tokenId); }

There are already 2 PRs for this issue.

Also, the code you shared above changes the internal interface, which is not something we want.

Made a PR, it passes all required test

[kr-meet]

@Amxx I want to work on this issue, will you give me an overview

[Amxx]

@kr-meet There is a big refactor of ERC721 that is going on internally right now. IMO we should wait for that refactor to be done before changing other things, otherwize we'll multiply conflicts