1 // Copyright 2007-2010 the V8 project authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
4
5 #ifndef V8_UNICODE_INL_H_
6 #define V8_UNICODE_INL_H_
7
8 #include "src/unicode.h"
9 #include "src/base/logging.h"
10 #include "src/utils.h"
11
12 namespace unibrow {
13
get(uchar code_point)14 template <class T, int s> bool Predicate<T, s>::get(uchar code_point) {
15 CacheEntry entry = entries_[code_point & kMask];
16 if (entry.code_point() == code_point) return entry.value();
17 return CalculateValue(code_point);
18 }
19
CalculateValue(uchar code_point)20 template <class T, int s> bool Predicate<T, s>::CalculateValue(
21 uchar code_point) {
22 bool result = T::Is(code_point);
23 entries_[code_point & kMask] = CacheEntry(code_point, result);
24 return result;
25 }
26
get(uchar c,uchar n,uchar * result)27 template <class T, int s> int Mapping<T, s>::get(uchar c, uchar n,
28 uchar* result) {
29 CacheEntry entry = entries_[c & kMask];
30 if (entry.code_point_ == c) {
31 if (entry.offset_ == 0) {
32 return 0;
33 } else {
34 result[0] = c + entry.offset_;
35 return 1;
36 }
37 } else {
38 return CalculateValue(c, n, result);
39 }
40 }
41
CalculateValue(uchar c,uchar n,uchar * result)42 template <class T, int s> int Mapping<T, s>::CalculateValue(uchar c, uchar n,
43 uchar* result) {
44 bool allow_caching = true;
45 int length = T::Convert(c, n, result, &allow_caching);
46 if (allow_caching) {
47 if (length == 1) {
48 entries_[c & kMask] = CacheEntry(c, result[0] - c);
49 return 1;
50 } else {
51 entries_[c & kMask] = CacheEntry(c, 0);
52 return 0;
53 }
54 } else {
55 return length;
56 }
57 }
58
59
EncodeOneByte(char * str,uint8_t c)60 unsigned Utf8::EncodeOneByte(char* str, uint8_t c) {
61 static const int kMask = ~(1 << 6);
62 if (c <= kMaxOneByteChar) {
63 str[0] = c;
64 return 1;
65 }
66 str[0] = 0xC0 | (c >> 6);
67 str[1] = 0x80 | (c & kMask);
68 return 2;
69 }
70
71 // Encode encodes the UTF-16 code units c and previous into the given str
72 // buffer, and combines surrogate code units into single code points. If
73 // replace_invalid is set to true, orphan surrogate code units will be replaced
74 // with kBadChar.
Encode(char * str,uchar c,int previous,bool replace_invalid)75 unsigned Utf8::Encode(char* str,
76 uchar c,
77 int previous,
78 bool replace_invalid) {
79 static const int kMask = ~(1 << 6);
80 if (c <= kMaxOneByteChar) {
81 str[0] = c;
82 return 1;
83 } else if (c <= kMaxTwoByteChar) {
84 str[0] = 0xC0 | (c >> 6);
85 str[1] = 0x80 | (c & kMask);
86 return 2;
87 } else if (c <= kMaxThreeByteChar) {
88 if (Utf16::IsSurrogatePair(previous, c)) {
89 const int kUnmatchedSize = kSizeOfUnmatchedSurrogate;
90 return Encode(str - kUnmatchedSize,
91 Utf16::CombineSurrogatePair(previous, c),
92 Utf16::kNoPreviousCharacter,
93 replace_invalid) - kUnmatchedSize;
94 } else if (replace_invalid &&
95 (Utf16::IsLeadSurrogate(c) ||
96 Utf16::IsTrailSurrogate(c))) {
97 c = kBadChar;
98 }
99 str[0] = 0xE0 | (c >> 12);
100 str[1] = 0x80 | ((c >> 6) & kMask);
101 str[2] = 0x80 | (c & kMask);
102 return 3;
103 } else {
104 str[0] = 0xF0 | (c >> 18);
105 str[1] = 0x80 | ((c >> 12) & kMask);
106 str[2] = 0x80 | ((c >> 6) & kMask);
107 str[3] = 0x80 | (c & kMask);
108 return 4;
109 }
110 }
111
112
ValueOf(const byte * bytes,size_t length,size_t * cursor)113 uchar Utf8::ValueOf(const byte* bytes, size_t length, size_t* cursor) {
114 if (length <= 0) return kBadChar;
115 byte first = bytes[0];
116 // Characters between 0000 and 0007F are encoded as a single character
117 if (first <= kMaxOneByteChar) {
118 *cursor += 1;
119 return first;
120 }
121 return CalculateValue(bytes, length, cursor);
122 }
123
Length(uchar c,int previous)124 unsigned Utf8::Length(uchar c, int previous) {
125 if (c <= kMaxOneByteChar) {
126 return 1;
127 } else if (c <= kMaxTwoByteChar) {
128 return 2;
129 } else if (c <= kMaxThreeByteChar) {
130 if (Utf16::IsTrailSurrogate(c) &&
131 Utf16::IsLeadSurrogate(previous)) {
132 return kSizeOfUnmatchedSurrogate - kBytesSavedByCombiningSurrogates;
133 }
134 return 3;
135 } else {
136 return 4;
137 }
138 }
139
IsValidCharacter(uchar c)140 bool Utf8::IsValidCharacter(uchar c) {
141 return c < 0xD800u || (c >= 0xE000u && c < 0xFDD0u) ||
142 (c > 0xFDEFu && c <= 0x10FFFFu && (c & 0xFFFEu) != 0xFFFEu &&
143 c != kBadChar);
144 }
145
146 } // namespace unibrow
147
148 #endif // V8_UNICODE_INL_H_
149