moonbitlang / core / 6821

89%

push

github

perf(builtin): build String/StringView::to_array with a direct loop

`to_array` went through `self.iter().fold(...)`, which keeps the
`String::iter` closure and pays a call per character. A direct
`for c in self` is lowered to an inline surrogate-pair decode with no
closure. Same result, including lone surrogates (new test compares with
`iter()`).

Release (builtin/string_to_array_bench_test.mbt), main -> this:

| bench                       | native        | js            | wasm-gc       |
|-----------------------------|---------------|---------------|---------------|
| String 51 ASCII chars       | 150 -> 54 ns  | 259 -> 212 ns | 343 -> 111 ns |
| String 1 KB mixed/emoji     | 2.17 -> 0.83us| 9.00 -> 3.76us| 6.04 -> 1.96us|
| StringView 1 KB mixed/emoji | 2.08 -> 0.85us| 6.92 -> 3.77us| 5.59 -> 1.97us|

Codex review: "No blocking issues found. The direct loops preserve surrogate-pair decoding, lone surrogates, and view boundaries."

Signed-off-by: Codex CLI <codex@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

4 of 4 new or added lines in 2 files covered. (100.0%)

16495 of 18386 relevant lines covered (89.72%)

312224.3 hits per line