Optimize deep compare for formatted text

[brrichards]

Summary

Optimize buildDeltaFromTree with cursor-based run detection. buildDeltaFromTree iterated through every character individually via charactersWithFormatting(), calling formatToQuillAttributes and JSON.stringify per character for format comparison. To help with operations like this two new cursor-based API's have been added to FormattedTextAsTree.Tree:

• getuniformRun(startIndex, maxLength) - returns length of consecutive characters with identical formating using raw cursor comparison. Handles StringLineAtom by always returning 1 (Quill only supports one per line by default).
• textString(startIndex, length) - returns a substring extracted via cursor, handling both StringTextAtom and StringLineAtom content types.

buildDeltaFromTree now walks the tree in runs rather than character-by-character, calling formatToQuillAttributes() once per run instead of once per character.

[Copilot]

Pull request overview

This PR optimizes formatted-text to Quill Delta generation by adding cursor-based “run” APIs to FormattedTextAsTree.Tree and updating buildDeltaFromTree to emit ops per uniform formatting run rather than per character.

Changes:

• Add getUniformRun(startIndex, maxLength?) and textString(startIndex, length) to FormattedTextAsTree.Tree (backed by a borrowed cursor implementation).
• Update buildDeltaFromTree to walk the tree by uniform runs and call formatToQuillAttributes() once per run.
• Add unit tests covering getUniformRun/textString, including behavior around line atoms.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 5 comments.

File	Description
packages/framework/react/src/text/formatted/quillFormattedView.tsx	Switch delta building to run-based traversal using the new tree APIs.
packages/dds/tree/src/text/textDomainFormatted.ts	Implement new cursor-based APIs (getUniformRun, textString) and expose them on the tree type.
packages/dds/tree/src/test/text/textDomainFormatted.spec.ts	Add tests for the new APIs, including line-atom scenarios.

[brrichards]

I have to assume that .join() is slower for small runs (a sentence, or a paragraph) compared to +=. Maybe the overhead cost is worth it?