Fix text not serializing correctly when having utf8 boundaries

[bnchi]

Fixes #169

[johnlepikhin]

Hi there,

Any chance to merge it?

[bnchi]

I'm just waiting on the maintainers to review and merge this.

[wooorm]

I missed/forgot about this. Will get to it soon!

[Copilot]

Pull Request Overview

This PR fixes an issue where text was not serialized correctly when encountering UTF-8 boundaries. It adds a new test to verify proper handling of UTF-8 characters and updates state management to correctly handle multi-byte character slicing.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
mdast_util_to_markdown/tests/roundtrip.rs	Added a new test ensuring correct serialization of UTF-8 text.
mdast_util_to_markdown/src/state.rs	Updated character slicing to correctly extract the last UTF-8 character.

[Copilot]

Spelling error: please change 'boundries' to 'boundaries' for clarity.

Suggested change

	"should support utf8 in boundries when serializing"
	);
	"should support utf8 in boundaries when serializing"

[johnlepikhin]

I just found another case which is not resolved with suggested fix:

    let doc = "я_𝄞";
    assert_eq!(to(&from(doc, &Default::default()).unwrap()).unwrap(), doc);

---- roundtrip stdout ----
thread 'roundtrip' panicked at mdast_util_to_markdown/src/state.rs:540:51:
byte index 7 is not a char boundary; it is inside '𝄞' (bytes 4..8) of `
я_𝄞

[wooorm]

Found them, thanks! :)