2
* Phoebe DOM Implementation.
4
* This is a C++ approximation of the W3C DOM model, which follows
5
* fairly closely the specifications in the various .idl files, copies of
6
* which are provided for reference. Most important is this one:
8
* http://www.w3.org/TR/2004/REC-DOM-Level-3-Core-20040407/idl-definitions.html
13
* Copyright (C) 2005-2008 Bob Jamison
15
* This library is free software; you can redistribute it and/or
16
* modify it under the terms of the GNU Lesser General Public
17
* License as published by the Free Software Foundation; either
18
* version 2.1 of the License, or (at your option) any later version.
20
* This library is distributed in the hope that it will be useful,
21
* but WITHOUT ANY WARRANTY; without even the implied warranty of
22
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
23
* Lesser General Public License for more details.
25
* You should have received a copy of the GNU Lesser General Public
26
* License along with this library; if not, write to the Free Software
27
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
30
#include "charclass.h"
34
* (impl) LetterOrDigit ::=
37
bool isLetterOrDigit(int ch)
48
* BaseChar | Ideographic
54
if (isIdeographic(ch))
61
* Following are from unicode.org, in the UnicodeData file
62
* in the Unicode Database
70
bool isBaseChar(int ch)
73
if ( (0x0041 <= ch && ch <= 0x005A) ||
74
(0x0061 <= ch && ch <= 0x007A) ||
75
(0x00C0 <= ch && ch <= 0x00D6) ||
76
(0x00D8 <= ch && ch <= 0x00F6) ||
77
(0x00F8 <= ch && ch <= 0x00FF) ||
78
(0x0100 <= ch && ch <= 0x0131) ||
79
(0x0134 <= ch && ch <= 0x013E) ||
80
(0x0141 <= ch && ch <= 0x0148) ||
81
(0x014A <= ch && ch <= 0x017E) ||
82
(0x0180 <= ch && ch <= 0x01C3) ||
83
(0x01CD <= ch && ch <= 0x01F0) ||
84
(0x01F4 <= ch && ch <= 0x01F5) ||
85
(0x01FA <= ch && ch <= 0x0217) ||
86
(0x0250 <= ch && ch <= 0x02A8) ||
87
(0x02BB <= ch && ch <= 0x02C1) ||
89
(0x0388 <= ch && ch <= 0x038A) ||
91
(0x038E <= ch && ch <= 0x03A1) ||
92
(0x03A3 <= ch && ch <= 0x03CE) ||
93
(0x03D0 <= ch && ch <= 0x03D6) ||
98
(0x03E2 <= ch && ch <= 0x03F3) ||
99
(0x0401 <= ch && ch <= 0x040C) ||
100
(0x040E <= ch && ch <= 0x044F) ||
101
(0x0451 <= ch && ch <= 0x045C) ||
102
(0x045E <= ch && ch <= 0x0481) ||
103
(0x0490 <= ch && ch <= 0x04C4) ||
104
(0x04C7 <= ch && ch <= 0x04C8) ||
105
(0x04CB <= ch && ch <= 0x04CC) ||
106
(0x04D0 <= ch && ch <= 0x04EB) ||
107
(0x04EE <= ch && ch <= 0x04F5) ||
108
(0x04F8 <= ch && ch <= 0x04F9) ||
109
(0x0531 <= ch && ch <= 0x0556) ||
111
(0x0561 <= ch && ch <= 0x0586) ||
112
(0x05D0 <= ch && ch <= 0x05EA) ||
113
(0x05F0 <= ch && ch <= 0x05F2) ||
114
(0x0621 <= ch && ch <= 0x063A) ||
115
(0x0641 <= ch && ch <= 0x064A) ||
116
(0x0671 <= ch && ch <= 0x06B7) ||
117
(0x06BA <= ch && ch <= 0x06BE) ||
118
(0x06C0 <= ch && ch <= 0x06CE) ||
119
(0x06D0 <= ch && ch <= 0x06D3) ||
121
(0x06E5 <= ch && ch <= 0x06E6) ||
122
(0x0905 <= ch && ch <= 0x0939) ||
124
(0x0958 <= ch && ch <= 0x0961) ||
125
(0x0985 <= ch && ch <= 0x098C) ||
126
(0x098F <= ch && ch <= 0x0990) ||
127
(0x0993 <= ch && ch <= 0x09A8) ||
128
(0x09AA <= ch && ch <= 0x09B0) ||
130
(0x09B6 <= ch && ch <= 0x09B9) ||
131
(0x09DC <= ch && ch <= 0x09DD) ||
132
(0x09DF <= ch && ch <= 0x09E1) ||
133
(0x09F0 <= ch && ch <= 0x09F1) ||
134
(0x0A05 <= ch && ch <= 0x0A0A) ||
135
(0x0A0F <= ch && ch <= 0x0A10) ||
136
(0x0A13 <= ch && ch <= 0x0A28) ||
137
(0x0A2A <= ch && ch <= 0x0A30) ||
138
(0x0A32 <= ch && ch <= 0x0A33) ||
139
(0x0A35 <= ch && ch <= 0x0A36) ||
140
(0x0A38 <= ch && ch <= 0x0A39) ||
141
(0x0A59 <= ch && ch <= 0x0A5C) ||
143
(0x0A72 <= ch && ch <= 0x0A74) ||
144
(0x0A85 <= ch && ch <= 0x0A8B) ||
146
(0x0A8F <= ch && ch <= 0x0A91) ||
147
(0x0A93 <= ch && ch <= 0x0AA8) ||
148
(0x0AAA <= ch && ch <= 0x0AB0) ||
149
(0x0AB2 <= ch && ch <= 0x0AB3) ||
150
(0x0AB5 <= ch && ch <= 0x0AB9) ||
153
(0x0B05 <= ch && ch <= 0x0B0C) ||
154
(0x0B0F <= ch && ch <= 0x0B10) ||
155
(0x0B13 <= ch && ch <= 0x0B28) ||
156
(0x0B2A <= ch && ch <= 0x0B30) ||
157
(0x0B32 <= ch && ch <= 0x0B33) ||
158
(0x0B36 <= ch && ch <= 0x0B39) ||
160
(0x0B5C <= ch && ch <= 0x0B5D) ||
161
(0x0B5F <= ch && ch <= 0x0B61) ||
162
(0x0B85 <= ch && ch <= 0x0B8A) ||
163
(0x0B8E <= ch && ch <= 0x0B90) ||
164
(0x0B92 <= ch && ch <= 0x0B95) ||
165
(0x0B99 <= ch && ch <= 0x0B9A) ||
167
(0x0B9E <= ch && ch <= 0x0B9F) ||
168
(0x0BA3 <= ch && ch <= 0x0BA4) ||
169
(0x0BA8 <= ch && ch <= 0x0BAA) ||
170
(0x0BAE <= ch && ch <= 0x0BB5) ||
171
(0x0BB7 <= ch && ch <= 0x0BB9) ||
172
(0x0C05 <= ch && ch <= 0x0C0C) ||
173
(0x0C0E <= ch && ch <= 0x0C10) ||
174
(0x0C12 <= ch && ch <= 0x0C28) ||
175
(0x0C2A <= ch && ch <= 0x0C33) ||
176
(0x0C35 <= ch && ch <= 0x0C39) ||
177
(0x0C60 <= ch && ch <= 0x0C61) ||
178
(0x0C85 <= ch && ch <= 0x0C8C) ||
179
(0x0C8E <= ch && ch <= 0x0C90) ||
180
(0x0C92 <= ch && ch <= 0x0CA8) ||
181
(0x0CAA <= ch && ch <= 0x0CB3) ||
182
(0x0CB5 <= ch && ch <= 0x0CB9) ||
184
(0x0CE0 <= ch && ch <= 0x0CE1) ||
185
(0x0D05 <= ch && ch <= 0x0D0C) ||
186
(0x0D0E <= ch && ch <= 0x0D10) ||
187
(0x0D12 <= ch && ch <= 0x0D28) ||
188
(0x0D2A <= ch && ch <= 0x0D39) ||
189
(0x0D60 <= ch && ch <= 0x0D61) ||
190
(0x0E01 <= ch && ch <= 0x0E2E) ||
192
(0x0E32 <= ch && ch <= 0x0E33) ||
193
(0x0E40 <= ch && ch <= 0x0E45) ||
194
(0x0E81 <= ch && ch <= 0x0E82) ||
196
(0x0E87 <= ch && ch <= 0x0E88) ||
199
(0x0E94 <= ch && ch <= 0x0E97) ||
200
(0x0E99 <= ch && ch <= 0x0E9F) ||
201
(0x0EA1 <= ch && ch <= 0x0EA3) ||
204
(0x0EAA <= ch && ch <= 0x0EAB) ||
205
(0x0EAD <= ch && ch <= 0x0EAE) ||
207
(0x0EB2 <= ch && ch <= 0x0EB3) ||
209
(0x0EC0 <= ch && ch <= 0x0EC4) ||
210
(0x0F40 <= ch && ch <= 0x0F47) ||
211
(0x0F49 <= ch && ch <= 0x0F69) ||
212
(0x10A0 <= ch && ch <= 0x10C5) ||
213
(0x10D0 <= ch && ch <= 0x10F6) ||
215
(0x1102 <= ch && ch <= 0x1103) ||
216
(0x1105 <= ch && ch <= 0x1107) ||
218
(0x110B <= ch && ch <= 0x110C) ||
219
(0x110E <= ch && ch <= 0x1112) ||
226
(0x1154 <= ch && ch <= 0x1155) ||
228
(0x115F <= ch && ch <= 0x1161) ||
233
(0x116D <= ch && ch <= 0x116E) ||
234
(0x1172 <= ch && ch <= 0x1173) ||
239
(0x11AE <= ch && ch <= 0x11AF) ||
240
(0x11B7 <= ch && ch <= 0x11B8) ||
242
(0x11BC <= ch && ch <= 0x11C2) ||
246
(0x1E00 <= ch && ch <= 0x1E9B) ||
247
(0x1EA0 <= ch && ch <= 0x1EF9) ||
248
(0x1F00 <= ch && ch <= 0x1F15) ||
249
(0x1F18 <= ch && ch <= 0x1F1D) ||
250
(0x1F20 <= ch && ch <= 0x1F45) ||
251
(0x1F48 <= ch && ch <= 0x1F4D) ||
252
(0x1F50 <= ch && ch <= 0x1F57) ||
256
(0x1F5F <= ch && ch <= 0x1F7D) ||
257
(0x1F80 <= ch && ch <= 0x1FB4) ||
258
(0x1FB6 <= ch && ch <= 0x1FBC) ||
260
(0x1FC2 <= ch && ch <= 0x1FC4) ||
261
(0x1FC6 <= ch && ch <= 0x1FCC) ||
262
(0x1FD0 <= ch && ch <= 0x1FD3) ||
263
(0x1FD6 <= ch && ch <= 0x1FDB) ||
264
(0x1FE0 <= ch && ch <= 0x1FEC) ||
265
(0x1FF2 <= ch && ch <= 0x1FF4) ||
266
(0x1FF6 <= ch && ch <= 0x1FFC) ||
268
(0x212A <= ch && ch <= 0x212B) ||
270
(0x2180 <= ch && ch <= 0x2182) ||
271
(0x3041 <= ch && ch <= 0x3094) ||
272
(0x30A1 <= ch && ch <= 0x30FA) ||
273
(0x3105 <= ch && ch <= 0x312C) ||
274
(0xAC00 <= ch && ch <= 0xD7A3) )
282
* (86) Ideographic ::=
284
bool isIdeographic(int ch)
286
if ( (0x4E00 <= ch && ch <=0x9FA5) ||
288
(0x3021 <= ch && ch <=0x3029) )
294
* (87) CombiningChar ::=
296
bool isCombiningChar(int ch)
298
if ( (0x0300 <= ch && ch <= 0x0345) ||
299
(0x0360 <= ch && ch <= 0x0361) ||
300
(0x0483 <= ch && ch <= 0x0486) ||
301
(0x0591 <= ch && ch <= 0x05A1) ||
302
(0x05A3 <= ch && ch <= 0x05B9) ||
303
(0x05BB <= ch && ch <= 0x05BD) ||
305
(0x05C1 <= ch && ch <= 0x05C2) ||
307
(0x064B <= ch && ch <= 0x0652) ||
309
(0x06D6 <= ch && ch <= 0x06DC) ||
310
(0x06DD <= ch && ch <= 0x06DF) ||
311
(0x06E0 <= ch && ch <= 0x06E4) ||
312
(0x06E7 <= ch && ch <= 0x06E8) ||
313
(0x06EA <= ch && ch <= 0x06ED) ||
314
(0x0901 <= ch && ch <= 0x0903) ||
316
(0x093E <= ch && ch <= 0x094C) ||
318
(0x0951 <= ch && ch <= 0x0954) ||
319
(0x0962 <= ch && ch <= 0x0963) ||
320
(0x0981 <= ch && ch <= 0x0983) ||
324
(0x09C0 <= ch && ch <= 0x09C4) ||
325
(0x09C7 <= ch && ch <= 0x09C8) ||
326
(0x09CB <= ch && ch <= 0x09CD) ||
328
(0x09E2 <= ch && ch <= 0x09E3) ||
333
(0x0A40 <= ch && ch <= 0x0A42) ||
334
(0x0A47 <= ch && ch <= 0x0A48) ||
335
(0x0A4B <= ch && ch <= 0x0A4D) ||
336
(0x0A70 <= ch && ch <= 0x0A71) ||
337
(0x0A81 <= ch && ch <= 0x0A83) ||
339
(0x0ABE <= ch && ch <= 0x0AC5) ||
340
(0x0AC7 <= ch && ch <= 0x0AC9) ||
341
(0x0ACB <= ch && ch <= 0x0ACD) ||
342
(0x0B01 <= ch && ch <= 0x0B03) ||
344
(0x0B3E <= ch && ch <= 0x0B43) ||
345
(0x0B47 <= ch && ch <= 0x0B48) ||
346
(0x0B4B <= ch && ch <= 0x0B4D) ||
347
(0x0B56 <= ch && ch <= 0x0B57) ||
348
(0x0B82 <= ch && ch <= 0x0B83) ||
349
(0x0BBE <= ch && ch <= 0x0BC2) ||
350
(0x0BC6 <= ch && ch <= 0x0BC8) ||
351
(0x0BCA <= ch && ch <= 0x0BCD) ||
353
(0x0C01 <= ch && ch <= 0x0C03) ||
354
(0x0C3E <= ch && ch <= 0x0C44) ||
355
(0x0C46 <= ch && ch <= 0x0C48) ||
356
(0x0C4A <= ch && ch <= 0x0C4D) ||
357
(0x0C55 <= ch && ch <= 0x0C56) ||
358
(0x0C82 <= ch && ch <= 0x0C83) ||
359
(0x0CBE <= ch && ch <= 0x0CC4) ||
360
(0x0CC6 <= ch && ch <= 0x0CC8) ||
361
(0x0CCA <= ch && ch <= 0x0CCD) ||
362
(0x0CD5 <= ch && ch <= 0x0CD6) ||
363
(0x0D02 <= ch && ch <= 0x0D03) ||
364
(0x0D3E <= ch && ch <= 0x0D43) ||
365
(0x0D46 <= ch && ch <= 0x0D48) ||
366
(0x0D4A <= ch && ch <= 0x0D4D) ||
369
(0x0E34 <= ch && ch <= 0x0E3A) ||
370
(0x0E47 <= ch && ch <= 0x0E4E) ||
372
(0x0EB4 <= ch && ch <= 0x0EB9) ||
373
(0x0EBB <= ch && ch <= 0x0EBC) ||
374
(0x0EC8 <= ch && ch <= 0x0ECD) ||
375
(0x0F18 <= ch && ch <= 0x0F19) ||
381
(0x0F71 <= ch && ch <= 0x0F84) ||
382
(0x0F86 <= ch && ch <= 0x0F8B) ||
383
(0x0F90 <= ch && ch <= 0x0F95) ||
385
(0x0F99 <= ch && ch <= 0x0FAD) ||
386
(0x0FB1 <= ch && ch <= 0x0FB7) ||
388
(0x20D0 <= ch && ch <= 0x20DC) ||
390
(0x302A <= ch && ch <= 0x302F) ||
403
if ( (0x0030 <= ch && ch <= 0x0039) ||
404
(0x0660 <= ch && ch <= 0x0669) ||
405
(0x06F0 <= ch && ch <= 0x06F9) ||
406
(0x0966 <= ch && ch <= 0x096F) ||
407
(0x09E6 <= ch && ch <= 0x09EF) ||
408
(0x0A66 <= ch && ch <= 0x0A6F) ||
409
(0x0AE6 <= ch && ch <= 0x0AEF) ||
410
(0x0B66 <= ch && ch <= 0x0B6F) ||
411
(0x0BE7 <= ch && ch <= 0x0BEF) ||
412
(0x0C66 <= ch && ch <= 0x0C6F) ||
413
(0x0CE6 <= ch && ch <= 0x0CEF) ||
414
(0x0D66 <= ch && ch <= 0x0D6F) ||
415
(0x0E50 <= ch && ch <= 0x0E59) ||
416
(0x0ED0 <= ch && ch <= 0x0ED9) ||
417
(0x0F20 <= ch && ch <= 0x0F29) )
426
bool isExtender(int ch)
436
(0x3031 <= ch && ch <= 0x3035) ||
437
(0x309D <= ch && ch <= 0x309E) ||
438
(0x30FC <= ch && ch <= 0x30FE) )
449
* UNICODE general class Zs
451
bool isSpaceSeparator(int ch)
476
* UNICODE general class Zl
478
bool isLineSeparator(int ch)
486
* UNICODE general class Zp
488
bool isParagraphSeparator(int ch)
496
* The union of the 3 space types.
498
bool isSpaceChar(int ch)
500
if ( isSpaceSeparator(ch) ||
501
isLineSeparator(ch) ||
502
isParagraphSeparator(ch))
508
* 3 spaces in isSpaceChar() which don't break
510
bool isNonBreakingSpace(int ch)
512
if (ch == 0x00A0 || ch == 0x2007 || ch == 0x202F)
520
bool isWhitespace(int ch)
522
if (isSpaceChar(ch) && !isNonBreakingSpace(ch))
524
if (ch == 0x0009 || // HORIZONTAL TABULATION
525
ch == 0x000A || // LINE FEED.
526
ch == 0x000B || // VERTICAL TABULATION.
527
ch == 0x000C || // FORM FEED.
528
ch == 0x000D || // CARRIAGE RETURN.
529
ch == 0x001C || // FILE SEPARATOR.
530
ch == 0x001D || // GROUP SEPARATOR.
531
ch == 0x001E || // RECORD SEPARATOR.
532
ch == 0x001F) // UNIT SEPARATOR.
added
removed
src/dom/charclass.cpp
