⚡️Possible ERC-721 safe_mint Gas Optimisation

[jaglinux]

Sorry for nitpick, I was going through the code and spotted gas saving optimization.
https://github.com/pcaversaccio/snekmate/blob/main/src/tokens/ERC721.vy#L546
Using token_id for incrementing self._counter saves gas, even though self._counter qualifies for warm touch,
reading state variable is costly than reading from local variable.

We can further optimize the code. As per the comment in next line, we can use unsafe_add() since we don't expect token_ids to exceed 2**256. However testSafeMintOverflow() fails if we add unsafe_add.

Also, the comments needs to move before L546 ?

I have a draft PR at jaglinux@4ea3521

[pcaversaccio]

Correct - we can use token_id to increment the counter. Want to open a PR?

"we can use unsafe_add()" -> no since it's always possible that someone changes the upper bound of _all_tokens. I want to include the checked arithmetic here, so we don't offer a foot gun! That's why my overflow test throws ;-) - I'm testing for this possibility.

"The comments needs to move before L546" -> no since the comment relates to the line that could potentially overflow (in this case an internal function call).

[jaglinux]

Yes, I will create the PR.