OLD | NEW |
1 // Copyright 2007-2008 the V8 project authors. All rights reserved. | 1 // Copyright 2007-2008 the V8 project authors. All rights reserved. |
2 // Redistribution and use in source and binary forms, with or without | 2 // Redistribution and use in source and binary forms, with or without |
3 // modification, are permitted provided that the following conditions are | 3 // modification, are permitted provided that the following conditions are |
4 // met: | 4 // met: |
5 // | 5 // |
6 // * Redistributions of source code must retain the above copyright | 6 // * Redistributions of source code must retain the above copyright |
7 // notice, this list of conditions and the following disclaimer. | 7 // notice, this list of conditions and the following disclaimer. |
8 // * Redistributions in binary form must reproduce the above | 8 // * Redistributions in binary form must reproduce the above |
9 // copyright notice, this list of conditions and the following | 9 // copyright notice, this list of conditions and the following |
10 // disclaimer in the documentation and/or other materials provided | 10 // disclaimer in the documentation and/or other materials provided |
(...skipping 26 matching lines...) Expand all Loading... |
37 | 37 |
38 namespace unibrow { | 38 namespace unibrow { |
39 | 39 |
40 typedef unsigned int uchar; | 40 typedef unsigned int uchar; |
41 typedef unsigned char byte; | 41 typedef unsigned char byte; |
42 | 42 |
43 /** | 43 /** |
44 * The max length of the result of converting the case of a single | 44 * The max length of the result of converting the case of a single |
45 * character. | 45 * character. |
46 */ | 46 */ |
47 static const int kMaxCaseConvertedSize = 3; | 47 static const int kMaxMappingSize = 4; |
48 | 48 |
49 template <class T, int size = 256> | 49 template <class T, int size = 256> |
50 class Predicate { | 50 class Predicate { |
51 public: | 51 public: |
52 inline Predicate() { } | 52 inline Predicate() { } |
53 inline bool get(uchar c); | 53 inline bool get(uchar c); |
54 private: | 54 private: |
55 friend class Test; | 55 friend class Test; |
56 bool CalculateValue(uchar c); | 56 bool CalculateValue(uchar c); |
57 struct CacheEntry { | 57 struct CacheEntry { |
(...skipping 15 matching lines...) Expand all Loading... |
73 // map differently depending on context are always looked up. | 73 // map differently depending on context are always looked up. |
74 template <class T, int size = 256> | 74 template <class T, int size = 256> |
75 class Mapping { | 75 class Mapping { |
76 public: | 76 public: |
77 inline Mapping() { } | 77 inline Mapping() { } |
78 inline int get(uchar c, uchar n, uchar* result); | 78 inline int get(uchar c, uchar n, uchar* result); |
79 private: | 79 private: |
80 friend class Test; | 80 friend class Test; |
81 int CalculateValue(uchar c, uchar n, uchar* result); | 81 int CalculateValue(uchar c, uchar n, uchar* result); |
82 struct CacheEntry { | 82 struct CacheEntry { |
83 inline CacheEntry() : code_point_(0), offset_(0) { } | 83 inline CacheEntry() : code_point_(kNoChar), offset_(0) { } |
84 inline CacheEntry(uchar code_point, signed offset) | 84 inline CacheEntry(uchar code_point, signed offset) |
85 : code_point_(code_point), | 85 : code_point_(code_point), |
86 offset_(offset) { } | 86 offset_(offset) { } |
87 uchar code_point_ : 21; | 87 uchar code_point_; |
88 signed offset_ : 11; | 88 signed offset_; |
| 89 static const int kNoChar = (1 << 21) - 1; |
89 }; | 90 }; |
90 static const int kSize = size; | 91 static const int kSize = size; |
91 static const int kMask = kSize - 1; | 92 static const int kMask = kSize - 1; |
92 CacheEntry entries_[kSize]; | 93 CacheEntry entries_[kSize]; |
93 }; | 94 }; |
94 | 95 |
95 class UnicodeData { | 96 class UnicodeData { |
96 private: | 97 private: |
97 friend class Test; | 98 friend class Test; |
98 static int GetByteCount(); | 99 static int GetByteCount(); |
(...skipping 116 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
215 }; | 216 }; |
216 struct Lowercase { | 217 struct Lowercase { |
217 static bool Is(uchar c); | 218 static bool Is(uchar c); |
218 }; | 219 }; |
219 struct Letter { | 220 struct Letter { |
220 static bool Is(uchar c); | 221 static bool Is(uchar c); |
221 }; | 222 }; |
222 struct Space { | 223 struct Space { |
223 static bool Is(uchar c); | 224 static bool Is(uchar c); |
224 }; | 225 }; |
225 struct Titlecase { | |
226 static bool Is(uchar c); | |
227 }; | |
228 struct Number { | 226 struct Number { |
229 static bool Is(uchar c); | 227 static bool Is(uchar c); |
230 }; | 228 }; |
231 struct DecimalDigit { | |
232 static bool Is(uchar c); | |
233 }; | |
234 struct Ideographic { | |
235 static bool Is(uchar c); | |
236 }; | |
237 struct WhiteSpace { | 229 struct WhiteSpace { |
238 static bool Is(uchar c); | 230 static bool Is(uchar c); |
239 }; | 231 }; |
240 struct HexDigit { | |
241 static bool Is(uchar c); | |
242 }; | |
243 struct AsciiHexDigit { | |
244 static bool Is(uchar c); | |
245 }; | |
246 struct BidiControl { | |
247 static bool Is(uchar c); | |
248 }; | |
249 struct JoinControl { | |
250 static bool Is(uchar c); | |
251 }; | |
252 struct Dash { | |
253 static bool Is(uchar c); | |
254 }; | |
255 struct Hyphen { | |
256 static bool Is(uchar c); | |
257 }; | |
258 struct LineTerminator { | 232 struct LineTerminator { |
259 static bool Is(uchar c); | 233 static bool Is(uchar c); |
260 }; | 234 }; |
261 struct RegExpSpecialChar { | |
262 static bool Is(uchar c); | |
263 }; | |
264 struct CombiningMark { | 235 struct CombiningMark { |
265 static bool Is(uchar c); | 236 static bool Is(uchar c); |
266 }; | 237 }; |
267 struct ConnectorPunctuation { | 238 struct ConnectorPunctuation { |
268 static bool Is(uchar c); | 239 static bool Is(uchar c); |
269 }; | 240 }; |
270 struct ToLowercase { | 241 struct ToLowercase { |
| 242 static const int kMaxWidth = 3; |
271 static int Convert(uchar c, | 243 static int Convert(uchar c, |
272 uchar n, | 244 uchar n, |
273 uchar* result, | 245 uchar* result, |
274 bool* allow_caching_ptr); | 246 bool* allow_caching_ptr); |
275 }; | 247 }; |
276 struct ToUppercase { | 248 struct ToUppercase { |
| 249 static const int kMaxWidth = 3; |
| 250 static int Convert(uchar c, |
| 251 uchar n, |
| 252 uchar* result, |
| 253 bool* allow_caching_ptr); |
| 254 }; |
| 255 struct Ecma262Canonicalize { |
| 256 static const int kMaxWidth = 1; |
| 257 static int Convert(uchar c, |
| 258 uchar n, |
| 259 uchar* result, |
| 260 bool* allow_caching_ptr); |
| 261 }; |
| 262 struct Ecma262UnCanonicalize { |
| 263 static const int kMaxWidth = 4; |
| 264 static int Convert(uchar c, |
| 265 uchar n, |
| 266 uchar* result, |
| 267 bool* allow_caching_ptr); |
| 268 }; |
| 269 struct CanonicalizationRange { |
| 270 static const int kMaxWidth = 1; |
277 static int Convert(uchar c, | 271 static int Convert(uchar c, |
278 uchar n, | 272 uchar n, |
279 uchar* result, | 273 uchar* result, |
280 bool* allow_caching_ptr); | 274 bool* allow_caching_ptr); |
281 }; | 275 }; |
282 | 276 |
283 } // namespace unibrow | 277 } // namespace unibrow |
284 | 278 |
285 #endif // __UNIBROW_H__ | 279 #endif // __UNIBROW_H__ |
OLD | NEW |