Refactor constructor safety

[asmello]

Although a lot more inconvenient for us, some of our functions have invariants that have to be held externally, so they should be unsafe fn. This also makes it justifiable to expose them publicly (as then users can opt-in for manual checking of the invariants).

Notable changes:

• Expose Pointer::new as a public (but unsafe) constructor (now called Pointer::new_unchecked). This gives users a way to bypass validation if they're sure they've got a valid string. Akin to String::from_utf8_unchecked and family.
• Add PointerBuf::new_unchecked as a new unsafe constructor, to match Pointer::new_unchecked.
• Made Token::from_encoded_unchecked unsafe. This one I didn't make public because I don't see much use for it. We may revisit this later.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 99.0%. Comparing base (c96007e) to head (aeb33b1).

Additional details and impacted files

| [Files with missing lines](https://app.codecov.io/gh/chanced/jsonptr/pull/76?dropdown=coverage&src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Chance) | Coverage Δ | | |---|---|---| | [src/pointer.rs](https://app.codecov.io/gh/chanced/jsonptr/pull/76?src=pr&el=tree&filepath=src%2Fpointer.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Chance#diff-c3JjL3BvaW50ZXIucnM=) | `98.1% <100.0%> (+<0.1%)` | :arrow_up: | | [src/token.rs](https://app.codecov.io/gh/chanced/jsonptr/pull/76?src=pr&el=tree&filepath=src%2Ftoken.rs&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Chance#diff-c3JjL3Rva2VuLnJz) | `100.0% <100.0%> (ø)` | |

[asmello]

I think the sanitizer test is failing due to a bug in the sanitizer itself (given the obscure error referencing its source code).

My guess is https://github.com/rust-lang/rust/issues/111073

[chanced]

I'm at the fair with my family but noticed this come through and just wanted to say that I'm stoked about this one. I considered suggesting exposing Pointer::new but was hesitant.

I'll review both tomorrow morning.

Also, you're awesome dude.

[asmello]

No rush, just came with a few ideas recently.

I noticed json-patch recently upgraded to 0.6.0, which is great (and I feel bad I let them do it, meant to contribute a PR myself).

[chanced]

Hah, you did! Perhaps not to hatch directly but the state of this crate would be in much worse shape without your contributions. You brought it up a few levels! :)

[asmello]

To be honest the unsafe for internal uses might be a bit overkill, I'm not 100% sure this is a great idea long term, but since it's an internal detail I think it's ok to experiment with.