diff options
| author | Cuong Manh Le <cuong.manhle.vn@gmail.com> | 2024-09-13 00:48:11 +0700 |
|---|---|---|
| committer | Gopher Robot <gobot@golang.org> | 2024-09-12 19:10:46 +0000 |
| commit | 751fbf9e5a73991fad24a48f1f523e446eb78291 (patch) | |
| tree | 4a4f173ce4430dd5c7a676572dde3fcc4f2072db /src/unicode | |
| parent | 8efb5ebfd02a8e5cb6c79c5cd0e093fe896e6347 (diff) | |
| download | go-751fbf9e5a73991fad24a48f1f523e446eb78291.tar.xz | |
unicode/utf8: use range loop in RuneCountInString
CL 28490 speeded up non-ASCII rune decoding, and ASCII rune is also
decoded faster now.
Benchmark using:
perflock -governor 70% go test -run=NONE -bench=BenchmarkRuneCountInString -count=10
Result:
name old time/op new time/op delta
RuneCountInStringTenASCIIChars-8 10.2ns ± 0% 7.1ns ± 1% -30.53% (p=0.000 n=8+9)
RuneCountInStringTenJapaneseChars-8 49.3ns ± 2% 38.5ns ± 2% -21.84% (p=0.000 n=8+8)
Fixes #13162
Change-Id: Ifb01f3799c5c93e7f7c7af13a95becfde85ae807
Reviewed-on: https://go-review.googlesource.com/c/go/+/612617
Reviewed-by: Tim King <taking@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Cuong Manh Le <cuong.manhle.vn@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Tim King <taking@google.com>
Diffstat (limited to 'src/unicode')
| -rw-r--r-- | src/unicode/utf8/utf8.go | 31 |
1 files changed, 2 insertions, 29 deletions
diff --git a/src/unicode/utf8/utf8.go b/src/unicode/utf8/utf8.go index 1c1391d55b..9743b74258 100644 --- a/src/unicode/utf8/utf8.go +++ b/src/unicode/utf8/utf8.go @@ -449,35 +449,8 @@ func RuneCount(p []byte) int { // RuneCountInString is like [RuneCount] but its input is a string. func RuneCountInString(s string) (n int) { - ns := len(s) - for i := 0; i < ns; n++ { - c := s[i] - if c < RuneSelf { - // ASCII fast path - i++ - continue - } - x := first[c] - if x == xx { - i++ // invalid. - continue - } - size := int(x & 7) - if i+size > ns { - i++ // Short or invalid. - continue - } - accept := acceptRanges[x>>4] - if c := s[i+1]; c < accept.lo || accept.hi < c { - size = 1 - } else if size == 2 { - } else if c := s[i+2]; c < locb || hicb < c { - size = 1 - } else if size == 3 { - } else if c := s[i+3]; c < locb || hicb < c { - size = 1 - } - i += size + for range s { + n++ } return n } |