Update treesitter parsing with text buffer chunk api

[rebornix]

Introduced a new getNearChunk api to get the minimal text after index offset. Adopted this in branch https://github.com/microsoft/vscode/tree/alexr00/gettingStartedWithTreeSitter and I can see the perf difference for editing

Before

After

---

@alexr00 I tried to adopt this change in #213565 but I didn't see the incremental parsing, even a single type will trigger the whole document to be re-parsed (from index: 0). We can catch up and see what's going on and what else you potentially need from the text buffer side.

[alexr00]

@rebornix this looks great! Would you consider merging it to main? #213565 needs some work still but I can adopt your chunk API there when I merge main into that PR. Or, if you'd prefer, I can just cherry-pick your changes into that PR.

[rebornix]

@alexr00 it's already in main so the only change you need in @213565 is https://github.com/microsoft/vscode/pull/215471/files#diff-57fe96385cfa45838fd3f40c7e745756d14de8201b073b8d46bd8bd8edb2e43e once you merge main into it.

[rebornix]

@alexr00 I updated the PR against alexr00/treeSitterParserService, we can merge it anytime. I'll do some benchmark and see if incremental parsing is still working.

[rebornix]

It's actually doing great with the latest code

I'm merging this into your PR @alexr00 first and we can see what else we can improve.

[alexr00]

Great, thanks for doing that!