perf(runtime): rewrite case-change in C++ to skip UTF-8 round-trip

[cirospaciari]

Summary

Follow-up to #26772, which introduced the 11 case-changing utility methods in Zig.

• Rewrites the implementation from Zig to C++, eliminating two unnecessary allocations and transcoding steps per call
• The Zig implementation converted every JS string to UTF-8 via bunstr.toUTF8(allocator), processed codepoints, then converted back via bun.String.cloneUTF8(result_bytes). The new C++ implementation works directly with JSC's native string encoding (Latin1 or UTF-16) using StringView, StringBuilder, and ICU — same pattern as stripANSI.cpp
• New CaseChange.cpp + CaseChange.h with the algorithm templated on Latin1Character/UChar; 11 functions registered directly in bunObjectTable as C++ host functions; deleted string_case.zig and removed the icu_toUpper/icu_toLower C-extern bridge wrappers

Test plan

• All 2098 existing case-change tests pass (bun bd test test/js/bun/util/case-change.test.ts)
• No regressions in other tests

Changelog

[robobun]

^{Updated 1:46 AM PT - Feb 6th, 2026}

❌ @cirospaciari, your commit a669587 has 3 failures in Build #36644 (All Failures):

• test/js/node/http/node-http-backpressure.test.ts - SIGKILL on 🐧 3.23 x64-baseline
• test/js/node/http/node-http-backpressure.test.ts - SIGKILL on 🐧 25.04 x64-baseline
• test/integration/next-pages/test/dev-server.test.ts - 1 failing on 🍎 14 aarch64
• test/integration/next-pages/test/dev-server.test.ts - 1 failing on 🍎 13 aarch64
• test/js/bun/util/case-change.test.ts - 1 failing on 🍎 13 aarch64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🍎 14 x64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🍎 14 aarch64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🍎 13 x64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🪟 2019 x64-baseline (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🪟 2019 x64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 13 x64-asan (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 13 aarch64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 25.04 aarch64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 3.23 aarch64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 3.23 x64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 3.23 x64-baseline (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 13 x64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 25.04 x64 (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 25.04 x64-baseline (new)
• test/js/bun/util/case-change.test.ts - 1 failing on 🐧 13 x64-baseline (new)

---

🧪   To try this PR locally:

bunx bun-pr 26773

That installs a local version of the PR into your bun-26773 executable, so you can run:

bun-26773 --bun

[coderabbitai]

Walkthrough

Adds eleven string case-transformation functions (camelCase, pascalCase, snakeCase, kebabCase, constantCase, dotCase, capitalCase, trainCase, pathCase, sentenceCase, noCase) to Bun. Includes TypeScript declarations, native C++ implementations with character classification and separator logic, registration in the Bun object, and comprehensive test suite validating cross-compatibility.

Changes

Cohort / File(s)	Summary
TypeScript Type Declarations packages/bun-types/bun.d.ts	Added type signatures for 11 string case-transformation functions to the Bun module, each accepting a string input and returning a string output.
C++ Implementation src/bun.js/bindings/CaseChange.h, src/bun.js/bindings/CaseChange.cpp	Implemented case conversion logic with character classification, word boundary detection, and per-style separators. Created 11 host functions wrapping a templated convertCase function supporting both Latin1 and UTF-16 string inputs.
Bun Object Integration src/bun.js/bindings/BunObject.cpp	Registered 11 new case-transformation functions in the BunObject's export lookup table and included the CaseChange header.
Test Suite test/js/bun/util/case-change.test.ts	Added comprehensive test coverage including compatibility matrix against the change-case library, round-trip conversions, idempotency checks, edge cases (empty strings, unicode, acronyms), and error handling validation.

🚥 Pre-merge checks | ✅ 2

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately describes the main change: rewriting case-change functions from Zig to C++ to eliminate UTF-8 round-trips, which aligns with the substantial refactoring across multiple files (CaseChange.cpp/h, BunObject.cpp, and test additions).
Description check	✅ Passed	PR description includes both required sections with clear explanation of changes and test plan, though changelog placeholder is empty.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Fix all issues with AI agents

In `@src/bun.js/bindings/CaseChange.cpp`:
- Around line 236-266: The per-codepoint mapping using u_toupper/u_tolower
inside the loop (see getTransform, WordTransform, u_toupper, u_tolower,
builder.append, CharType/Latin1Character) must be replaced with ICU full-string
case mappings to handle multi-codepoint expansions and context/locale rules;
extract the word slice from input, call u_strToUpper/u_strToLower or use
UCaseMap to transform the whole word (or transform first character + remainder
for Capitalize) into a temporary UTF-16/UTF-32 buffer, then append that
transformed string to builder (handling length changes and encoding conversion)
instead of appending per-codepoint results. Ensure Capitalize semantics use
full-mapping for the first grapheme/character and full-mapping lowercase for the
rest, and thread locale/flags through the UCaseMap calls if applicable.