From: Rob Pike <r@golang.org>
Date: Wed, 21 Jul 2010 02:53:59 +0000 (-0700)
Subject: bytes: add Title
X-Git-Tag: weekly.2010-07-29~65
X-Git-Url: http://www.git.cypherpunks.su/?a=commitdiff_plain;h=de228c0e0c6e4f5736dcdde913afff527d0f76f5;p=gostls13.git

bytes: add Title

R=rsc
CC=golang-dev
https://golang.org/cl/1872042
---

diff --git a/src/pkg/bytes/bytes.go b/src/pkg/bytes/bytes.go
index bcf7b8609b..25a5738324 100644
--- a/src/pkg/bytes/bytes.go
+++ b/src/pkg/bytes/bytes.go
@@ -332,6 +332,52 @@ func ToLower(s []byte) []byte { return Map(unicode.ToLower, s) }
 // ToTitle returns a copy of the byte array s with all Unicode letters mapped to their title case.
 func ToTitle(s []byte) []byte { return Map(unicode.ToTitle, s) }
 
+// isSeparator reports whether the rune could mark a word boundary.
+// TODO: update when package unicode captures more of the properties.
+func isSeparator(rune int) bool {
+	// ASCII alphanumerics and underscore are not separators
+	if rune <= 0x7F {
+		switch {
+		case '0' <= rune && rune <= '9':
+			return false
+		case 'a' <= rune && rune <= 'z':
+			return false
+		case 'A' <= rune && rune <= 'Z':
+			return false
+		case rune == '_':
+			return false
+		}
+		return true
+	}
+	// Letters and digits are not separators
+	if unicode.IsLetter(rune) || unicode.IsDigit(rune) {
+		return false
+	}
+	// Otherwise, all we can do for now is treat spaces as separators.
+	return unicode.IsSpace(rune)
+}
+
+// BUG(r): The rule Title uses for word boundaries does not handle Unicode punctuation properly.
+
+// Title returns a copy of s with all Unicode letters that begin words
+// mapped to their title case.
+func Title(s []byte) []byte {
+	// Use a closure here to remember state.
+	// Hackish but effective. Depends on Map scanning in order and calling
+	// the closure once per rune.
+	prev := ' '
+	return Map(
+		func(r int) int {
+			if isSeparator(prev) {
+				prev = r
+				return unicode.ToTitle(r)
+			}
+			prev = r
+			return r
+		},
+		s)
+}
+
 // TrimLeftFunc returns a subslice of s by slicing off all leading UTF-8 encoded
 // Unicode code points c that satisfy f(c).
 func TrimLeftFunc(s []byte, f func(r int) bool) []byte {
diff --git a/src/pkg/bytes/bytes_test.go b/src/pkg/bytes/bytes_test.go
index 8197543dcf..de5edd120f 100644
--- a/src/pkg/bytes/bytes_test.go
+++ b/src/pkg/bytes/bytes_test.go
@@ -685,3 +685,25 @@ func TestReplace(t *testing.T) {
 		}
 	}
 }
+
+type TitleTest struct {
+	in, out string
+}
+
+var TitleTests = []TitleTest{
+	TitleTest{"", ""},
+	TitleTest{"a", "A"},
+	TitleTest{" aaa aaa aaa ", " Aaa Aaa Aaa "},
+	TitleTest{" Aaa Aaa Aaa ", " Aaa Aaa Aaa "},
+	TitleTest{"123a456", "123a456"},
+	TitleTest{"double-blind", "Double-Blind"},
+	TitleTest{"ÿøû", "Ÿøû"},
+}
+
+func TestTitle(t *testing.T) {
+	for _, tt := range TitleTests {
+		if s := string(Title([]byte(tt.in))); s != tt.out {
+			t.Errorf("Title(%q) = %q, want %q", tt.in, s, tt.out)
+		}
+	}
+}