Honor byte-order mark when parsing UTF-16 and UTF-8 input

[SAY-5]

Fixes #62. UTF-16 (and UTF-8 BOM) documents crashed with a BoundsError because the byte tokenizer was handed the raw bytes without any encoding detection, so the interleaved zero bytes desynced the </> scanning and walked off the end of the buffer.

This adds a small normalization step at the single Vector{UInt8} entry point: a leading BOM (FF FE, FE FF, or EF BB BF) is honored, UTF-16 is transcoded to UTF-8, and a UTF-8 BOM is stripped, so the tokenizer always sees UTF-8. Input without a BOM is returned unchanged. Added a regression test covering UTF-16 LE/BE and UTF-8 BOM round-trips.

[SAY-5]

Good call, added that. The loop now also writes each BOM-prefixed buffer to a temp file and reads it back through read(path, Node), so the mmap path is covered too.

[codecov-commenter]

Welcome to Codecov 🎉

Once you merge this PR into your default branch, you're all set! Codecov will compare coverage reports and display results in all future pull requests.

Thanks for integrating Codecov - We've got you covered ☂️

[mathieu17g]

Thanks Anything else you want to add, or shall I merge?