Mint re-entrancy guard with optimizations.

[Vectorized]

Ok, so as we know, the _safeMint function is NOT safe from re-entrancy. #92

Ideally, most contracts should use _mint.

Although we can warn about it in the comments / docs, many devs won't read it and will happily use _safeMint as it is.
For security reasons, I won't cite any examples here.

Furthermore, the README example shows _safeMint being used in an unprotected manner.

From a usability point of view, I suggest that we make the _mint function re-entrancy safe regardless, to be foolproof.

This will also reduce the cognitive dissonance of _safeMint being unsafe, giving a cleaner API design.

The overhead is only 100 gas for an extra warm SLOAD.
Assuming a function using _safeMint costs ~60K gas, this is only 0.2% overhead that can protect against catastrophic consequences. For users who use _mint, the gas overhead is as good as none.

To ease the concerns for the increase in gas for _safeMint, the function is optimized even further, such that we can save more gas to cover for the overhead.

In practice, implementations that need to use _safeMint for some reason can save 2k gas (cold SLOAD) with this functionality, as they don't need to use OpenZepplin's Reentrancy guard.

Before:

·------------------------------------------|---------------------------|-------------|
|           Solc version: 0.8.11           ·  Optimizer enabled: true  ·  Runs: 800  ·
···········································|···························|·············|
|  Methods                                                                            
···························|···············|·············|·············|·············|
|  Contract                ·  Method       ·  Min        ·  Max        ·  Avg        ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  mintOne      ·      56502  ·      90702  ·      57186  ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  mintTen      ·      74696  ·     108896  ·      75380  ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  safeMintOne  ·      59256  ·      93456  ·      59940  ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  safeMintTen  ·      79246  ·     113446  ·      79930  ·
···························|···············|·············|·············|·············|
|  Deployments                             ·                                         ·
···········································|·············|·············|·············|
|  ERC721AGasReporterMock                  ·          -  ·          -  ·    1146497  ·
···········································|·············|·············|·············|

After re-entrancy guard + gas optimizations:

·------------------------------------------|---------------------------|-------------|
|           Solc version: 0.8.11           ·  Optimizer enabled: true  ·  Runs: 800  ·
···········································|···························|·············|
|  Methods                                                                            
···························|···············|·············|·············|·············|
|  Contract                ·  Method       ·  Min        ·  Max        ·  Avg        ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  mintOne      ·      56462  ·      90662  ·      57146  ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  mintTen      ·      74116  ·     108316  ·      74800  ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  safeMintOne  ·      59143  ·      93343  ·      59827  ·
···························|···············|·············|·············|·············|
|  ERC721AGasReporterMock  ·  safeMintTen  ·      76775  ·     110975  ·      77459  ·
···························|···············|·············|·············|·············|
|  Deployments                             ·                                         ·
···········································|·············|·············|·············|
|  ERC721AGasReporterMock                  ·          -  ·          -  ·    1168285  ·
·------------------------------------------|-------------|-------------|-------------|

@cygaar @ahbanavi

[cygaar]

I like it, let's land this after the start index PR like you said

[liarco]

Hi, thank you all for the great job with this project.

I just tried to get in touch with @Vectorized from other places but I couldn't, so I hope not to bother you asking this question here.

About what you said here:

Ideally, most contracts should use _mint.

Although we can warn about it in the comments / docs, many devs won't read it and will happily use _safeMint as it is.

Do you mean that using _safeMint() instead of _mint() is a bad practice in general, or do you refer to just using it without considering potential re-entrancy attacks?

Would you mind explaining why you would suggest using _mint() in most situations?

Thank you for your time.

[Vectorized]

@liarco Messaged you on twitter. I forgot to turn on direct messages. You can contact me there from now.

https://ethereum.stackexchange.com/questions/115280/mint-vs-safemint-which-is-best-for-erc721 <- by the lead author of ERC721.