feat(stdlib): Add Buffer.getChar

[spotandjake]

This adds Buffer.setChar and Buffer.getChar to complement their bytes counterparts.

Note: We have to read the first byte before getChar so we can correctly do the bounds check similar to how its done in Bytes.getChar.

Note: I realized we missed a piece of documentation on Bytes.getChar so I added it here.

[alex-snezhko]

Why is this 1 if it's UTF-8?

[spotandjake]

Characters can be between 1 and 4 bytes, we need to ensure the first byte exists so we can check the char size on line 406 and we do an additional length check on line 408 with the actual length.

This is why getChar needs to operate on the bytes directly rather than just using Bytes.getChar like our other helpers.

[alex-snezhko]

If the operation is intended to get a character starting at a byte index then if you point your index in the middle of a UTF-8 character then what would the expectation be?

[spotandjake]

I'm pretty sure it would either return a character in the case that the rest was a valid char (I can't recall if thats ever the case), but more likely MalformedUnicode.

[ospencer]

I'm somewhat against these get/set APIs (not just the ones introduced in this PR, the existing ones too). I think it really encourages the wrong use of Buffers; they're really just meant for appending data, and they add a bunch of mental burden of "am I supposed to use add or set?". I'm more in favor of removing the existing ones rather than adding more. What are the use cases here?

[spotandjake]

I'm somewhat against these get/set APIs (not just the ones introduced in this PR, the existing ones too). I think it really encourages the wrong use of Buffers; they're really just meant for appending data, and they add a bunch of mental burden of "am I supposed to use add or set?". I'm more in favor of removing the existing ones rather than adding more. What are the use cases here?

If we have any of the get and set commands I think these should exist to match. If we want to talk about removing them though I think its fair, most applications would possibly be better served by Bytes, and Bytes.resize.

[ospencer]

We only have them for integers (so they're missing for everything else), and they landed in the initial implementation... I think we were just kinda caught up in that implementation and didn't audit the full API.

[spotandjake]

I removed Buffer.setChar as discussed at the community meeting and will follow up with a pr removing the other Buffer.set apis in a pr closer to 0.8.