| 1 |
747 |
jeremybenn |
// Copyright 2011 The Go Authors. All rights reserved.
|
| 2 |
|
|
// Use of this source code is governed by a BSD-style
|
| 3 |
|
|
// license that can be found in the LICENSE file.
|
| 4 |
|
|
|
| 5 |
|
|
package unicode
|
| 6 |
|
|
|
| 7 |
|
|
// Bit masks for each code point under U+0100, for fast lookup.
|
| 8 |
|
|
const (
|
| 9 |
|
|
pC = 1 << iota // a control character.
|
| 10 |
|
|
pP // a punctuation character.
|
| 11 |
|
|
pN // a numeral.
|
| 12 |
|
|
pS // a symbolic character.
|
| 13 |
|
|
pZ // a spacing character.
|
| 14 |
|
|
pLu // an upper-case letter.
|
| 15 |
|
|
pLl // a lower-case letter.
|
| 16 |
|
|
pp // a printable character according to Go's definition.
|
| 17 |
|
|
pg = pp | pZ // a graphical character according to the Unicode definition.
|
| 18 |
|
|
)
|
| 19 |
|
|
|
| 20 |
|
|
// GraphicRanges defines the set of graphic characters according to Unicode.
|
| 21 |
|
|
var GraphicRanges = []*RangeTable{
|
| 22 |
|
|
L, M, N, P, S, Zs,
|
| 23 |
|
|
}
|
| 24 |
|
|
|
| 25 |
|
|
// PrintRanges defines the set of printable characters according to Go.
|
| 26 |
|
|
// ASCII space, U+0020, is handled separately.
|
| 27 |
|
|
var PrintRanges = []*RangeTable{
|
| 28 |
|
|
L, M, N, P, S,
|
| 29 |
|
|
}
|
| 30 |
|
|
|
| 31 |
|
|
// IsGraphic reports whether the rune is defined as a Graphic by Unicode.
|
| 32 |
|
|
// Such characters include letters, marks, numbers, punctuation, symbols, and
|
| 33 |
|
|
// spaces, from categories L, M, N, P, S, Zs.
|
| 34 |
|
|
func IsGraphic(r rune) bool {
|
| 35 |
|
|
// We convert to uint32 to avoid the extra test for negative,
|
| 36 |
|
|
// and in the index we convert to uint8 to avoid the range check.
|
| 37 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 38 |
|
|
return properties[uint8(r)]&pg != 0
|
| 39 |
|
|
}
|
| 40 |
|
|
return IsOneOf(GraphicRanges, r)
|
| 41 |
|
|
}
|
| 42 |
|
|
|
| 43 |
|
|
// IsPrint reports whether the rune is defined as printable by Go. Such
|
| 44 |
|
|
// characters include letters, marks, numbers, punctuation, symbols, and the
|
| 45 |
|
|
// ASCII space character, from categories L, M, N, P, S and the ASCII space
|
| 46 |
|
|
// character. This categorization is the same as IsGraphic except that the
|
| 47 |
|
|
// only spacing character is ASCII space, U+0020.
|
| 48 |
|
|
func IsPrint(r rune) bool {
|
| 49 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 50 |
|
|
return properties[uint8(r)]&pp != 0
|
| 51 |
|
|
}
|
| 52 |
|
|
return IsOneOf(PrintRanges, r)
|
| 53 |
|
|
}
|
| 54 |
|
|
|
| 55 |
|
|
// IsOneOf reports whether the rune is a member of one of the ranges.
|
| 56 |
|
|
// The rune is known to be above Latin-1.
|
| 57 |
|
|
func IsOneOf(set []*RangeTable, r rune) bool {
|
| 58 |
|
|
for _, inside := range set {
|
| 59 |
|
|
if Is(inside, r) {
|
| 60 |
|
|
return true
|
| 61 |
|
|
}
|
| 62 |
|
|
}
|
| 63 |
|
|
return false
|
| 64 |
|
|
}
|
| 65 |
|
|
|
| 66 |
|
|
// IsControl reports whether the rune is a control character.
|
| 67 |
|
|
// The C (Other) Unicode category includes more code points
|
| 68 |
|
|
// such as surrogates; use Is(C, rune) to test for them.
|
| 69 |
|
|
func IsControl(r rune) bool {
|
| 70 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 71 |
|
|
return properties[uint8(r)]&pC != 0
|
| 72 |
|
|
}
|
| 73 |
|
|
// All control characters are < Latin1Max.
|
| 74 |
|
|
return false
|
| 75 |
|
|
}
|
| 76 |
|
|
|
| 77 |
|
|
// IsLetter reports whether the rune is a letter (category L).
|
| 78 |
|
|
func IsLetter(r rune) bool {
|
| 79 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 80 |
|
|
return properties[uint8(r)]&(pLu|pLl) != 0
|
| 81 |
|
|
}
|
| 82 |
|
|
return Is(Letter, r)
|
| 83 |
|
|
}
|
| 84 |
|
|
|
| 85 |
|
|
// IsMark reports whether the rune is a mark character (category M).
|
| 86 |
|
|
func IsMark(r rune) bool {
|
| 87 |
|
|
// There are no mark characters in Latin-1.
|
| 88 |
|
|
return Is(Mark, r)
|
| 89 |
|
|
}
|
| 90 |
|
|
|
| 91 |
|
|
// IsNumber reports whether the rune is a number (category N).
|
| 92 |
|
|
func IsNumber(r rune) bool {
|
| 93 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 94 |
|
|
return properties[uint8(r)]&pN != 0
|
| 95 |
|
|
}
|
| 96 |
|
|
return Is(Number, r)
|
| 97 |
|
|
}
|
| 98 |
|
|
|
| 99 |
|
|
// IsPunct reports whether the rune is a Unicode punctuation character
|
| 100 |
|
|
// (category P).
|
| 101 |
|
|
func IsPunct(r rune) bool {
|
| 102 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 103 |
|
|
return properties[uint8(r)]&pP != 0
|
| 104 |
|
|
}
|
| 105 |
|
|
return Is(Punct, r)
|
| 106 |
|
|
}
|
| 107 |
|
|
|
| 108 |
|
|
// IsSpace reports whether the rune is a space character as defined
|
| 109 |
|
|
// by Unicode's White Space property; in the Latin-1 space
|
| 110 |
|
|
// this is
|
| 111 |
|
|
// '\t', '\n', '\v', '\f', '\r', ' ', U+0085 (NEL), U+00A0 (NBSP).
|
| 112 |
|
|
// Other definitions of spacing characters are set by category
|
| 113 |
|
|
// Z and property Pattern_White_Space.
|
| 114 |
|
|
func IsSpace(r rune) bool {
|
| 115 |
|
|
// This property isn't the same as Z; special-case it.
|
| 116 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 117 |
|
|
switch r {
|
| 118 |
|
|
case '\t', '\n', '\v', '\f', '\r', ' ', 0x85, 0xA0:
|
| 119 |
|
|
return true
|
| 120 |
|
|
}
|
| 121 |
|
|
return false
|
| 122 |
|
|
}
|
| 123 |
|
|
return Is(White_Space, r)
|
| 124 |
|
|
}
|
| 125 |
|
|
|
| 126 |
|
|
// IsSymbol reports whether the rune is a symbolic character.
|
| 127 |
|
|
func IsSymbol(r rune) bool {
|
| 128 |
|
|
if uint32(r) <= MaxLatin1 {
|
| 129 |
|
|
return properties[uint8(r)]&pS != 0
|
| 130 |
|
|
}
|
| 131 |
|
|
return Is(Symbol, r)
|
| 132 |
|
|
}
|
