[littledivy]

rust::String::latin1() pushed every byte as a char. Rust JSG converts every one-byte string argument this way, so a 64 KiB ASCII string took 24.8 µs vs 2.0 µs in C++ JSG.

In this patch the leading ASCII run is found 32 bytes at a time and copied in bulk, and only the rest is encoded per character. The exact UTF-8 length is reserved up front.

CPU time per call (bench-jsg from #8013, --config=benchmark, 5 processes × 5 reps):

	before	after	C++ JSG
ascii/16	46.9 ns	45.6 ns	28.5 ns
ascii/1024	445.5 ns	78.4 ns	54.8 ns
ascii/65536	24.8 µs	1.80 µs	2.04 µs
latin1/16	70.7 ns	51.3 ns	33.4 ns
latin1/65536	33.9 µs	34.9 µs	31.8 µs

[clanky-the-clanker]

✅ Approved — No issues found. The new Latin-1 to UTF-8 fast path is correct: the unchecked UTF-8 conversion only ever sees bytes that were checked to be ASCII, and the reserved capacity matches the exact output length.

_{The decision is advisory: this comment never approves, blocks or requests changes on the pull request.}

---

✅ All checks passed

📋 Code Quality: ✅ LGTM

I found no problems in the new string_from_latin1 path.

• Safe use of the unchecked conversion: leading_ascii_len only counts a 32-byte chunk when every byte in it is below 0x80. After the chunks, it finishes with a byte-by-byte is_ascii() scan. So the ascii slice holds only ASCII bytes, and str::from_utf8_unchecked is safe to call on it.
• Exact capacity: a byte at or above 0x80 becomes 2 UTF-8 bytes and every other byte stays 1 byte. That makes len + count(rest >= 0x80) the exact output size, so the string never has to grow while it is filled. The prefix is all ASCII, so counting only over rest is correct.
• Tests: the new test covers an empty input and ASCII prefixes just below, at and just above the 32-byte chunk size, and at several multiples of it. Each case is followed by non-ASCII bytes and an ASCII tail.

---

📝 Documentation: ✅ LGTM

I found no documentation problems. The new comments describe what the code does now, as the repository's comment guidelines ask.

---

_{Reviewed commit: 232dde4a}

[ask-bonk]

Since last review: 0 resolved, 0 still open, 0 new.
LGTM!

---

_{Reviewed commit: 232dde4a · github run}

[codecov-commenter]

Codecov Report

❌ Patch coverage is 92.30769% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 41.96%. Comparing base (2c37de0) to head (232dde4).
⚠️ Report is 40 commits behind head on main.

Files with missing lines	Patch %	Lines
src/rust/cxx/src/symbols/rust_string.rs	92.30%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8016      +/-   ##
==========================================
- Coverage   42.05%   41.96%   -0.10%     
==========================================
  Files        1061     1064       +3     
  Lines      301008   302174    +1166     
  Branches    29043    29082      +39     
==========================================
+ Hits       126599   126803     +204     
- Misses     157389   158340     +951     
- Partials    17020    17031      +11     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.