fix(fast-build): properly decode multi-byte UTF-8 in JSON string parser

[janechu]

Pull Request

📖 Description

Fixes a bug in the hand-rolled JSON string parser (json.rs) where non-ASCII characters were corrupted during parsing.

The parse_string function advanced through the input byte-by-byte and used b as char for unescaped characters. For ASCII bytes this is correct, but for multi-byte UTF-8 sequences it casts each byte independently to a Unicode scalar, producing garbage. For example, ⭐ (U+2B50, UTF-8: 0xE2 0xAD 0x90) was rendered as â (U+00E2) followed by two invisible control characters.

The fix determines the byte-length of each UTF-8 sequence from the leading byte, reads the full sequence, and decodes it with std::str::from_utf8. ASCII bytes (< 0x80) continue to use the fast b as char path.

This explains the â SELECTED corruption observed in the observer-map fixture output (issue seen in #7388).

📑 Test Plan

New unit tests in json.rs:

• test_parse_string_with_emoji — ⭐ round-trips correctly
• test_parse_string_with_multibyte_chars — 2-byte (é), 3-byte (✓), and 4-byte (🎉) sequences
• test_parse_string_emoji_in_object — emoji values inside a JSON object

New integration tests in tests/bindings.rs:

• test_binding_emoji_from_state — emoji from JSON state renders correctly in a binding
• test_binding_emoji_literal_in_template — emoji in template literal text is preserved verbatim
• test_binding_multibyte_chars — general multi-byte character round-trip

All existing tests pass (cargo test).

✅ Checklist

General

• I have included a change request file using $ npm run change
• I have added tests for my changes.
• I have tested my changes.
• I have updated the project documentation to reflect my changes.
• I have read the CONTRIBUTING documentation and followed the standards for this project.

[Copilot]

Pull request overview

Fixes crates/microsoft-fast-build’s hand-rolled JSON string parsing so that non-ASCII literal characters in JSON strings (e.g., emoji, accented letters) are decoded as proper multi-byte UTF-8 sequences instead of being corrupted by per-byte as char casting. This improves correctness for SSR/state-driven rendering paths that consume JSON.

Changes:

• Update parse_string to decode multi-byte UTF-8 sequences as a whole (keeping a fast ASCII path).
• Add unit tests in json.rs covering emoji and 2/3/4-byte UTF-8 sequences in strings and objects.
• Add integration tests in tests/bindings.rs to ensure multi-byte characters round-trip through bindings/rendering.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
crates/microsoft-fast-build/src/json.rs	Fix UTF-8 decoding in parse_string and add targeted unit tests.
crates/microsoft-fast-build/tests/bindings.rs	Add integration tests to verify correct rendering/binding with non-ASCII content.
crates/microsoft-fast-build/DESIGN.md	Document the updated parse_string behavior for literal non-ASCII UTF-8.