p-token: Fix review comments

[febo]

Address Jon's comments from #28.

Closes #28

[joncinque]

Looks great! Mostly small things

[joncinque]

Just a couple of last bits, then this is good to go on my side

[joncinque]

Did you mean to remove the if *bytes.get_unchecked(1) == 1 portion to this?

We probably want to keep that to be safe

[febo]

The issue is that it disproportionally increases CUs to have it. We could do the check on the accessor method:

pub fn new_authority(&self) -> Option<&Pubkey> {
    if self.new_authority.0 == 1 {
        Option::Some(&self.new_authority.1)
    } else {
        Option::None
    }
}

[febo]

Or even return a Result?

[joncinque]

That's fine, but it has to be 1 or 0 or it errors, which means this would need to return Result<Option<&Pubkey>, ProgramError> instead

[febo]

Yeah, I think that is the best option. I will test it, but I can't imagine that it would add too much overhead.

[febo]

We can't do this and the current implementation has a different behaviour than the original one. We need to validate the instruction data before executing any code of the processor, otherwise we can get different results.

I fixed both SetAuthority and InitializeMint instruction data validation, adding the validation for the option byte.

[joncinque]

Should this have if *bytes.get_unchecked(33) == 1 like the other one?

[febo]

Same here, we could have the check on the accessor instead.

[joncinque]

Looks great!