Make determining whether a code point represents a combining mark faster (#1719)

This commit is contained in:
Philipp Emanuel Weidmann
2020-06-12 09:40:00 +05:30
committed by GitHub
parent efb38b8636
commit 5ce26cca71
3 changed files with 29 additions and 9 deletions

View File

@@ -79,7 +79,7 @@ bench-compare:
for i in 1 2 3; do \ for i in 1 2 3; do \
go test -bench=. ./internal/...; \ go test -bench=. ./internal/...; \
done > benchmark_results done > benchmark_results
benchstat benchmark_results_baseline benchmark_results benchstat -alpha 0.15 benchmark_results_baseline benchmark_results
clean: clean:
rm -f micro rm -f micro

View File

@@ -16,6 +16,16 @@ import (
// For rendering, micro will display the combining characters. It's not perfect // For rendering, micro will display the combining characters. It's not perfect
// but it's pretty good. // but it's pretty good.
var minMark = rune(unicode.Mark.R16[0].Lo)
func isMark(r rune) bool {
// Fast path
if r < minMark {
return false
}
return unicode.In(r, unicode.Mark)
}
// DecodeCharacter returns the next character from an array of bytes // DecodeCharacter returns the next character from an array of bytes
// A character is a rune along with any accompanying combining runes // A character is a rune along with any accompanying combining runes
func DecodeCharacter(b []byte) (rune, []rune, int) { func DecodeCharacter(b []byte) (rune, []rune, int) {
@@ -24,7 +34,7 @@ func DecodeCharacter(b []byte) (rune, []rune, int) {
c, s := utf8.DecodeRune(b) c, s := utf8.DecodeRune(b)
var combc []rune var combc []rune
for unicode.In(c, unicode.Mark) { for isMark(c) {
combc = append(combc, c) combc = append(combc, c)
size += s size += s
@@ -43,7 +53,7 @@ func DecodeCharacterInString(str string) (rune, []rune, int) {
c, s := utf8.DecodeRuneInString(str) c, s := utf8.DecodeRuneInString(str)
var combc []rune var combc []rune
for unicode.In(c, unicode.Mark) { for isMark(c) {
combc = append(combc, c) combc = append(combc, c)
size += s size += s
@@ -61,7 +71,7 @@ func CharacterCount(b []byte) int {
for len(b) > 0 { for len(b) > 0 {
r, size := utf8.DecodeRune(b) r, size := utf8.DecodeRune(b)
if !unicode.In(r, unicode.Mark) { if !isMark(r) {
s++ s++
} }
@@ -77,7 +87,7 @@ func CharacterCountInString(str string) int {
s := 0 s := 0
for _, r := range str { for _, r := range str {
if !unicode.In(r, unicode.Mark) { if !isMark(r) {
s++ s++
} }
} }

View File

@@ -5,6 +5,16 @@ import (
"unicode/utf8" "unicode/utf8"
) )
var minMark = rune(unicode.Mark.R16[0].Lo)
func isMark(r rune) bool {
// Fast path
if r < minMark {
return false
}
return unicode.In(r, unicode.Mark)
}
// DecodeCharacter returns the next character from an array of bytes // DecodeCharacter returns the next character from an array of bytes
// A character is a rune along with any accompanying combining runes // A character is a rune along with any accompanying combining runes
func DecodeCharacter(b []byte) (rune, []rune, int) { func DecodeCharacter(b []byte) (rune, []rune, int) {
@@ -13,7 +23,7 @@ func DecodeCharacter(b []byte) (rune, []rune, int) {
c, s := utf8.DecodeRune(b) c, s := utf8.DecodeRune(b)
var combc []rune var combc []rune
for unicode.In(c, unicode.Mark) { for isMark(c) {
combc = append(combc, c) combc = append(combc, c)
size += s size += s
@@ -32,7 +42,7 @@ func DecodeCharacterInString(str string) (rune, []rune, int) {
c, s := utf8.DecodeRuneInString(str) c, s := utf8.DecodeRuneInString(str)
var combc []rune var combc []rune
for unicode.In(c, unicode.Mark) { for isMark(c) {
combc = append(combc, c) combc = append(combc, c)
size += s size += s
@@ -50,7 +60,7 @@ func CharacterCount(b []byte) int {
for len(b) > 0 { for len(b) > 0 {
r, size := utf8.DecodeRune(b) r, size := utf8.DecodeRune(b)
if !unicode.In(r, unicode.Mark) { if !isMark(r) {
s++ s++
} }
@@ -66,7 +76,7 @@ func CharacterCountInString(str string) int {
s := 0 s := 0
for _, r := range str { for _, r := range str {
if !unicode.In(r, unicode.Mark) { if !isMark(r) {
s++ s++
} }
} }