1 | /* |
2 | * Copyright (C) 1999-2000 Harri Porten ([email protected]) |
3 | * Copyright (C) 2006-2019 Apple Inc. All Rights Reserved. |
4 | * Copyright (C) 2007 Cameron Zwarich ([email protected]) |
5 | * Copyright (C) 2010 Zoltan Herczeg ([email protected]) |
6 | * Copyright (C) 2012 Mathias Bynens ([email protected]) |
7 | * |
8 | * This library is free software; you can redistribute it and/or |
9 | * modify it under the terms of the GNU Library General Public |
10 | * License as published by the Free Software Foundation; either |
11 | * version 2 of the License, or (at your option) any later version. |
12 | * |
13 | * This library is distributed in the hope that it will be useful, |
14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
16 | * Library General Public License for more details. |
17 | * |
18 | * You should have received a copy of the GNU Library General Public License |
19 | * along with this library; see the file COPYING.LIB. If not, write to |
20 | * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, |
21 | * Boston, MA 02110-1301, USA. |
22 | * |
23 | */ |
24 | |
25 | #include "config.h" |
26 | #include "Lexer.h" |
27 | |
28 | #include "BuiltinNames.h" |
29 | #include "Identifier.h" |
30 | #include "JSCInlines.h" |
31 | #include "JSFunctionInlines.h" |
32 | #include "KeywordLookup.h" |
33 | #include "Lexer.lut.h" |
34 | #include "Nodes.h" |
35 | #include "ParseInt.h" |
36 | #include "Parser.h" |
37 | #include <ctype.h> |
38 | #include <limits.h> |
39 | #include <string.h> |
40 | #include <wtf/Assertions.h> |
41 | #include <wtf/HexNumber.h> |
42 | #include <wtf/Variant.h> |
43 | #include <wtf/dtoa.h> |
44 | |
45 | namespace JSC { |
46 | |
47 | bool isLexerKeyword(const Identifier& identifier) |
48 | { |
49 | return JSC::mainTable.entry(identifier); |
50 | } |
51 | |
52 | enum CharacterType { |
53 | // Types for the main switch |
54 | |
55 | // The first three types are fixed, and also used for identifying |
56 | // ASCII alpha and alphanumeric characters (see isIdentStart and isIdentPart). |
57 | CharacterIdentifierStart, |
58 | CharacterZero, |
59 | CharacterNumber, |
60 | |
61 | // For single-byte characters grandfathered into Other_ID_Continue -- namely just U+00B7 MIDDLE DOT. |
62 | // (http://unicode.org/reports/tr31/#Backward_Compatibility) |
63 | CharacterOtherIdentifierPart, |
64 | |
65 | CharacterInvalid, |
66 | CharacterLineTerminator, |
67 | CharacterExclamationMark, |
68 | CharacterOpenParen, |
69 | CharacterCloseParen, |
70 | CharacterOpenBracket, |
71 | CharacterCloseBracket, |
72 | CharacterComma, |
73 | CharacterColon, |
74 | CharacterQuestion, |
75 | CharacterTilde, |
76 | CharacterQuote, |
77 | CharacterBackQuote, |
78 | CharacterDot, |
79 | CharacterSlash, |
80 | CharacterBackSlash, |
81 | CharacterSemicolon, |
82 | CharacterOpenBrace, |
83 | CharacterCloseBrace, |
84 | |
85 | CharacterAdd, |
86 | CharacterSub, |
87 | CharacterMultiply, |
88 | CharacterModulo, |
89 | CharacterAnd, |
90 | CharacterXor, |
91 | CharacterOr, |
92 | CharacterLess, |
93 | CharacterGreater, |
94 | CharacterEqual, |
95 | |
96 | // Other types (only one so far) |
97 | CharacterWhiteSpace, |
98 | CharacterHash, |
99 | CharacterPrivateIdentifierStart |
100 | }; |
101 | |
102 | // 256 Latin-1 codes |
103 | static constexpr const unsigned short typesOfLatin1Characters[256] = { |
104 | /* 0 - Null */ CharacterInvalid, |
105 | /* 1 - Start of Heading */ CharacterInvalid, |
106 | /* 2 - Start of Text */ CharacterInvalid, |
107 | /* 3 - End of Text */ CharacterInvalid, |
108 | /* 4 - End of Transm. */ CharacterInvalid, |
109 | /* 5 - Enquiry */ CharacterInvalid, |
110 | /* 6 - Acknowledgment */ CharacterInvalid, |
111 | /* 7 - Bell */ CharacterInvalid, |
112 | /* 8 - Back Space */ CharacterInvalid, |
113 | /* 9 - Horizontal Tab */ CharacterWhiteSpace, |
114 | /* 10 - Line Feed */ CharacterLineTerminator, |
115 | /* 11 - Vertical Tab */ CharacterWhiteSpace, |
116 | /* 12 - Form Feed */ CharacterWhiteSpace, |
117 | /* 13 - Carriage Return */ CharacterLineTerminator, |
118 | /* 14 - Shift Out */ CharacterInvalid, |
119 | /* 15 - Shift In */ CharacterInvalid, |
120 | /* 16 - Data Line Escape */ CharacterInvalid, |
121 | /* 17 - Device Control 1 */ CharacterInvalid, |
122 | /* 18 - Device Control 2 */ CharacterInvalid, |
123 | /* 19 - Device Control 3 */ CharacterInvalid, |
124 | /* 20 - Device Control 4 */ CharacterInvalid, |
125 | /* 21 - Negative Ack. */ CharacterInvalid, |
126 | /* 22 - Synchronous Idle */ CharacterInvalid, |
127 | /* 23 - End of Transmit */ CharacterInvalid, |
128 | /* 24 - Cancel */ CharacterInvalid, |
129 | /* 25 - End of Medium */ CharacterInvalid, |
130 | /* 26 - Substitute */ CharacterInvalid, |
131 | /* 27 - Escape */ CharacterInvalid, |
132 | /* 28 - File Separator */ CharacterInvalid, |
133 | /* 29 - Group Separator */ CharacterInvalid, |
134 | /* 30 - Record Separator */ CharacterInvalid, |
135 | /* 31 - Unit Separator */ CharacterInvalid, |
136 | /* 32 - Space */ CharacterWhiteSpace, |
137 | /* 33 - ! */ CharacterExclamationMark, |
138 | /* 34 - " */ CharacterQuote, |
139 | /* 35 - # */ CharacterHash, |
140 | /* 36 - $ */ CharacterIdentifierStart, |
141 | /* 37 - % */ CharacterModulo, |
142 | /* 38 - & */ CharacterAnd, |
143 | /* 39 - ' */ CharacterQuote, |
144 | /* 40 - ( */ CharacterOpenParen, |
145 | /* 41 - ) */ CharacterCloseParen, |
146 | /* 42 - * */ CharacterMultiply, |
147 | /* 43 - + */ CharacterAdd, |
148 | /* 44 - , */ CharacterComma, |
149 | /* 45 - - */ CharacterSub, |
150 | /* 46 - . */ CharacterDot, |
151 | /* 47 - / */ CharacterSlash, |
152 | /* 48 - 0 */ CharacterZero, |
153 | /* 49 - 1 */ CharacterNumber, |
154 | /* 50 - 2 */ CharacterNumber, |
155 | /* 51 - 3 */ CharacterNumber, |
156 | /* 52 - 4 */ CharacterNumber, |
157 | /* 53 - 5 */ CharacterNumber, |
158 | /* 54 - 6 */ CharacterNumber, |
159 | /* 55 - 7 */ CharacterNumber, |
160 | /* 56 - 8 */ CharacterNumber, |
161 | /* 57 - 9 */ CharacterNumber, |
162 | /* 58 - : */ CharacterColon, |
163 | /* 59 - ; */ CharacterSemicolon, |
164 | /* 60 - < */ CharacterLess, |
165 | /* 61 - = */ CharacterEqual, |
166 | /* 62 - > */ CharacterGreater, |
167 | /* 63 - ? */ CharacterQuestion, |
168 | /* 64 - @ */ CharacterPrivateIdentifierStart, |
169 | /* 65 - A */ CharacterIdentifierStart, |
170 | /* 66 - B */ CharacterIdentifierStart, |
171 | /* 67 - C */ CharacterIdentifierStart, |
172 | /* 68 - D */ CharacterIdentifierStart, |
173 | /* 69 - E */ CharacterIdentifierStart, |
174 | /* 70 - F */ CharacterIdentifierStart, |
175 | /* 71 - G */ CharacterIdentifierStart, |
176 | /* 72 - H */ CharacterIdentifierStart, |
177 | /* 73 - I */ CharacterIdentifierStart, |
178 | /* 74 - J */ CharacterIdentifierStart, |
179 | /* 75 - K */ CharacterIdentifierStart, |
180 | /* 76 - L */ CharacterIdentifierStart, |
181 | /* 77 - M */ CharacterIdentifierStart, |
182 | /* 78 - N */ CharacterIdentifierStart, |
183 | /* 79 - O */ CharacterIdentifierStart, |
184 | /* 80 - P */ CharacterIdentifierStart, |
185 | /* 81 - Q */ CharacterIdentifierStart, |
186 | /* 82 - R */ CharacterIdentifierStart, |
187 | /* 83 - S */ CharacterIdentifierStart, |
188 | /* 84 - T */ CharacterIdentifierStart, |
189 | /* 85 - U */ CharacterIdentifierStart, |
190 | /* 86 - V */ CharacterIdentifierStart, |
191 | /* 87 - W */ CharacterIdentifierStart, |
192 | /* 88 - X */ CharacterIdentifierStart, |
193 | /* 89 - Y */ CharacterIdentifierStart, |
194 | /* 90 - Z */ CharacterIdentifierStart, |
195 | /* 91 - [ */ CharacterOpenBracket, |
196 | /* 92 - \ */ CharacterBackSlash, |
197 | /* 93 - ] */ CharacterCloseBracket, |
198 | /* 94 - ^ */ CharacterXor, |
199 | /* 95 - _ */ CharacterIdentifierStart, |
200 | /* 96 - ` */ CharacterBackQuote, |
201 | /* 97 - a */ CharacterIdentifierStart, |
202 | /* 98 - b */ CharacterIdentifierStart, |
203 | /* 99 - c */ CharacterIdentifierStart, |
204 | /* 100 - d */ CharacterIdentifierStart, |
205 | /* 101 - e */ CharacterIdentifierStart, |
206 | /* 102 - f */ CharacterIdentifierStart, |
207 | /* 103 - g */ CharacterIdentifierStart, |
208 | /* 104 - h */ CharacterIdentifierStart, |
209 | /* 105 - i */ CharacterIdentifierStart, |
210 | /* 106 - j */ CharacterIdentifierStart, |
211 | /* 107 - k */ CharacterIdentifierStart, |
212 | /* 108 - l */ CharacterIdentifierStart, |
213 | /* 109 - m */ CharacterIdentifierStart, |
214 | /* 110 - n */ CharacterIdentifierStart, |
215 | /* 111 - o */ CharacterIdentifierStart, |
216 | /* 112 - p */ CharacterIdentifierStart, |
217 | /* 113 - q */ CharacterIdentifierStart, |
218 | /* 114 - r */ CharacterIdentifierStart, |
219 | /* 115 - s */ CharacterIdentifierStart, |
220 | /* 116 - t */ CharacterIdentifierStart, |
221 | /* 117 - u */ CharacterIdentifierStart, |
222 | /* 118 - v */ CharacterIdentifierStart, |
223 | /* 119 - w */ CharacterIdentifierStart, |
224 | /* 120 - x */ CharacterIdentifierStart, |
225 | /* 121 - y */ CharacterIdentifierStart, |
226 | /* 122 - z */ CharacterIdentifierStart, |
227 | /* 123 - { */ CharacterOpenBrace, |
228 | /* 124 - | */ CharacterOr, |
229 | /* 125 - } */ CharacterCloseBrace, |
230 | /* 126 - ~ */ CharacterTilde, |
231 | /* 127 - Delete */ CharacterInvalid, |
232 | /* 128 - Cc category */ CharacterInvalid, |
233 | /* 129 - Cc category */ CharacterInvalid, |
234 | /* 130 - Cc category */ CharacterInvalid, |
235 | /* 131 - Cc category */ CharacterInvalid, |
236 | /* 132 - Cc category */ CharacterInvalid, |
237 | /* 133 - Cc category */ CharacterInvalid, |
238 | /* 134 - Cc category */ CharacterInvalid, |
239 | /* 135 - Cc category */ CharacterInvalid, |
240 | /* 136 - Cc category */ CharacterInvalid, |
241 | /* 137 - Cc category */ CharacterInvalid, |
242 | /* 138 - Cc category */ CharacterInvalid, |
243 | /* 139 - Cc category */ CharacterInvalid, |
244 | /* 140 - Cc category */ CharacterInvalid, |
245 | /* 141 - Cc category */ CharacterInvalid, |
246 | /* 142 - Cc category */ CharacterInvalid, |
247 | /* 143 - Cc category */ CharacterInvalid, |
248 | /* 144 - Cc category */ CharacterInvalid, |
249 | /* 145 - Cc category */ CharacterInvalid, |
250 | /* 146 - Cc category */ CharacterInvalid, |
251 | /* 147 - Cc category */ CharacterInvalid, |
252 | /* 148 - Cc category */ CharacterInvalid, |
253 | /* 149 - Cc category */ CharacterInvalid, |
254 | /* 150 - Cc category */ CharacterInvalid, |
255 | /* 151 - Cc category */ CharacterInvalid, |
256 | /* 152 - Cc category */ CharacterInvalid, |
257 | /* 153 - Cc category */ CharacterInvalid, |
258 | /* 154 - Cc category */ CharacterInvalid, |
259 | /* 155 - Cc category */ CharacterInvalid, |
260 | /* 156 - Cc category */ CharacterInvalid, |
261 | /* 157 - Cc category */ CharacterInvalid, |
262 | /* 158 - Cc category */ CharacterInvalid, |
263 | /* 159 - Cc category */ CharacterInvalid, |
264 | /* 160 - Zs category (nbsp) */ CharacterWhiteSpace, |
265 | /* 161 - Po category */ CharacterInvalid, |
266 | /* 162 - Sc category */ CharacterInvalid, |
267 | /* 163 - Sc category */ CharacterInvalid, |
268 | /* 164 - Sc category */ CharacterInvalid, |
269 | /* 165 - Sc category */ CharacterInvalid, |
270 | /* 166 - So category */ CharacterInvalid, |
271 | /* 167 - So category */ CharacterInvalid, |
272 | /* 168 - Sk category */ CharacterInvalid, |
273 | /* 169 - So category */ CharacterInvalid, |
274 | /* 170 - Ll category */ CharacterIdentifierStart, |
275 | /* 171 - Pi category */ CharacterInvalid, |
276 | /* 172 - Sm category */ CharacterInvalid, |
277 | /* 173 - Cf category */ CharacterInvalid, |
278 | /* 174 - So category */ CharacterInvalid, |
279 | /* 175 - Sk category */ CharacterInvalid, |
280 | /* 176 - So category */ CharacterInvalid, |
281 | /* 177 - Sm category */ CharacterInvalid, |
282 | /* 178 - No category */ CharacterInvalid, |
283 | /* 179 - No category */ CharacterInvalid, |
284 | /* 180 - Sk category */ CharacterInvalid, |
285 | /* 181 - Ll category */ CharacterIdentifierStart, |
286 | /* 182 - So category */ CharacterInvalid, |
287 | /* 183 - Po category */ CharacterOtherIdentifierPart, |
288 | /* 184 - Sk category */ CharacterInvalid, |
289 | /* 185 - No category */ CharacterInvalid, |
290 | /* 186 - Ll category */ CharacterIdentifierStart, |
291 | /* 187 - Pf category */ CharacterInvalid, |
292 | /* 188 - No category */ CharacterInvalid, |
293 | /* 189 - No category */ CharacterInvalid, |
294 | /* 190 - No category */ CharacterInvalid, |
295 | /* 191 - Po category */ CharacterInvalid, |
296 | /* 192 - Lu category */ CharacterIdentifierStart, |
297 | /* 193 - Lu category */ CharacterIdentifierStart, |
298 | /* 194 - Lu category */ CharacterIdentifierStart, |
299 | /* 195 - Lu category */ CharacterIdentifierStart, |
300 | /* 196 - Lu category */ CharacterIdentifierStart, |
301 | /* 197 - Lu category */ CharacterIdentifierStart, |
302 | /* 198 - Lu category */ CharacterIdentifierStart, |
303 | /* 199 - Lu category */ CharacterIdentifierStart, |
304 | /* 200 - Lu category */ CharacterIdentifierStart, |
305 | /* 201 - Lu category */ CharacterIdentifierStart, |
306 | /* 202 - Lu category */ CharacterIdentifierStart, |
307 | /* 203 - Lu category */ CharacterIdentifierStart, |
308 | /* 204 - Lu category */ CharacterIdentifierStart, |
309 | /* 205 - Lu category */ CharacterIdentifierStart, |
310 | /* 206 - Lu category */ CharacterIdentifierStart, |
311 | /* 207 - Lu category */ CharacterIdentifierStart, |
312 | /* 208 - Lu category */ CharacterIdentifierStart, |
313 | /* 209 - Lu category */ CharacterIdentifierStart, |
314 | /* 210 - Lu category */ CharacterIdentifierStart, |
315 | /* 211 - Lu category */ CharacterIdentifierStart, |
316 | /* 212 - Lu category */ CharacterIdentifierStart, |
317 | /* 213 - Lu category */ CharacterIdentifierStart, |
318 | /* 214 - Lu category */ CharacterIdentifierStart, |
319 | /* 215 - Sm category */ CharacterInvalid, |
320 | /* 216 - Lu category */ CharacterIdentifierStart, |
321 | /* 217 - Lu category */ CharacterIdentifierStart, |
322 | /* 218 - Lu category */ CharacterIdentifierStart, |
323 | /* 219 - Lu category */ CharacterIdentifierStart, |
324 | /* 220 - Lu category */ CharacterIdentifierStart, |
325 | /* 221 - Lu category */ CharacterIdentifierStart, |
326 | /* 222 - Lu category */ CharacterIdentifierStart, |
327 | /* 223 - Ll category */ CharacterIdentifierStart, |
328 | /* 224 - Ll category */ CharacterIdentifierStart, |
329 | /* 225 - Ll category */ CharacterIdentifierStart, |
330 | /* 226 - Ll category */ CharacterIdentifierStart, |
331 | /* 227 - Ll category */ CharacterIdentifierStart, |
332 | /* 228 - Ll category */ CharacterIdentifierStart, |
333 | /* 229 - Ll category */ CharacterIdentifierStart, |
334 | /* 230 - Ll category */ CharacterIdentifierStart, |
335 | /* 231 - Ll category */ CharacterIdentifierStart, |
336 | /* 232 - Ll category */ CharacterIdentifierStart, |
337 | /* 233 - Ll category */ CharacterIdentifierStart, |
338 | /* 234 - Ll category */ CharacterIdentifierStart, |
339 | /* 235 - Ll category */ CharacterIdentifierStart, |
340 | /* 236 - Ll category */ CharacterIdentifierStart, |
341 | /* 237 - Ll category */ CharacterIdentifierStart, |
342 | /* 238 - Ll category */ CharacterIdentifierStart, |
343 | /* 239 - Ll category */ CharacterIdentifierStart, |
344 | /* 240 - Ll category */ CharacterIdentifierStart, |
345 | /* 241 - Ll category */ CharacterIdentifierStart, |
346 | /* 242 - Ll category */ CharacterIdentifierStart, |
347 | /* 243 - Ll category */ CharacterIdentifierStart, |
348 | /* 244 - Ll category */ CharacterIdentifierStart, |
349 | /* 245 - Ll category */ CharacterIdentifierStart, |
350 | /* 246 - Ll category */ CharacterIdentifierStart, |
351 | /* 247 - Sm category */ CharacterInvalid, |
352 | /* 248 - Ll category */ CharacterIdentifierStart, |
353 | /* 249 - Ll category */ CharacterIdentifierStart, |
354 | /* 250 - Ll category */ CharacterIdentifierStart, |
355 | /* 251 - Ll category */ CharacterIdentifierStart, |
356 | /* 252 - Ll category */ CharacterIdentifierStart, |
357 | /* 253 - Ll category */ CharacterIdentifierStart, |
358 | /* 254 - Ll category */ CharacterIdentifierStart, |
359 | /* 255 - Ll category */ CharacterIdentifierStart |
360 | }; |
361 | |
362 | // This table provides the character that results from \X where X is the index in the table beginning |
363 | // with SPACE. A table value of 0 means that more processing needs to be done. |
364 | static constexpr const LChar singleCharacterEscapeValuesForASCII[128] = { |
365 | /* 0 - Null */ 0, |
366 | /* 1 - Start of Heading */ 0, |
367 | /* 2 - Start of Text */ 0, |
368 | /* 3 - End of Text */ 0, |
369 | /* 4 - End of Transm. */ 0, |
370 | /* 5 - Enquiry */ 0, |
371 | /* 6 - Acknowledgment */ 0, |
372 | /* 7 - Bell */ 0, |
373 | /* 8 - Back Space */ 0, |
374 | /* 9 - Horizontal Tab */ 0, |
375 | /* 10 - Line Feed */ 0, |
376 | /* 11 - Vertical Tab */ 0, |
377 | /* 12 - Form Feed */ 0, |
378 | /* 13 - Carriage Return */ 0, |
379 | /* 14 - Shift Out */ 0, |
380 | /* 15 - Shift In */ 0, |
381 | /* 16 - Data Line Escape */ 0, |
382 | /* 17 - Device Control 1 */ 0, |
383 | /* 18 - Device Control 2 */ 0, |
384 | /* 19 - Device Control 3 */ 0, |
385 | /* 20 - Device Control 4 */ 0, |
386 | /* 21 - Negative Ack. */ 0, |
387 | /* 22 - Synchronous Idle */ 0, |
388 | /* 23 - End of Transmit */ 0, |
389 | /* 24 - Cancel */ 0, |
390 | /* 25 - End of Medium */ 0, |
391 | /* 26 - Substitute */ 0, |
392 | /* 27 - Escape */ 0, |
393 | /* 28 - File Separator */ 0, |
394 | /* 29 - Group Separator */ 0, |
395 | /* 30 - Record Separator */ 0, |
396 | /* 31 - Unit Separator */ 0, |
397 | /* 32 - Space */ ' ', |
398 | /* 33 - ! */ '!', |
399 | /* 34 - " */ '"', |
400 | /* 35 - # */ '#', |
401 | /* 36 - $ */ '$', |
402 | /* 37 - % */ '%', |
403 | /* 38 - & */ '&', |
404 | /* 39 - ' */ '\'', |
405 | /* 40 - ( */ '(', |
406 | /* 41 - ) */ ')', |
407 | /* 42 - * */ '*', |
408 | /* 43 - + */ '+', |
409 | /* 44 - , */ ',', |
410 | /* 45 - - */ '-', |
411 | /* 46 - . */ '.', |
412 | /* 47 - / */ '/', |
413 | /* 48 - 0 */ 0, |
414 | /* 49 - 1 */ 0, |
415 | /* 50 - 2 */ 0, |
416 | /* 51 - 3 */ 0, |
417 | /* 52 - 4 */ 0, |
418 | /* 53 - 5 */ 0, |
419 | /* 54 - 6 */ 0, |
420 | /* 55 - 7 */ 0, |
421 | /* 56 - 8 */ 0, |
422 | /* 57 - 9 */ 0, |
423 | /* 58 - : */ ':', |
424 | /* 59 - ; */ ';', |
425 | /* 60 - < */ '<', |
426 | /* 61 - = */ '=', |
427 | /* 62 - > */ '>', |
428 | /* 63 - ? */ '?', |
429 | /* 64 - @ */ '@', |
430 | /* 65 - A */ 'A', |
431 | /* 66 - B */ 'B', |
432 | /* 67 - C */ 'C', |
433 | /* 68 - D */ 'D', |
434 | /* 69 - E */ 'E', |
435 | /* 70 - F */ 'F', |
436 | /* 71 - G */ 'G', |
437 | /* 72 - H */ 'H', |
438 | /* 73 - I */ 'I', |
439 | /* 74 - J */ 'J', |
440 | /* 75 - K */ 'K', |
441 | /* 76 - L */ 'L', |
442 | /* 77 - M */ 'M', |
443 | /* 78 - N */ 'N', |
444 | /* 79 - O */ 'O', |
445 | /* 80 - P */ 'P', |
446 | /* 81 - Q */ 'Q', |
447 | /* 82 - R */ 'R', |
448 | /* 83 - S */ 'S', |
449 | /* 84 - T */ 'T', |
450 | /* 85 - U */ 'U', |
451 | /* 86 - V */ 'V', |
452 | /* 87 - W */ 'W', |
453 | /* 88 - X */ 'X', |
454 | /* 89 - Y */ 'Y', |
455 | /* 90 - Z */ 'Z', |
456 | /* 91 - [ */ '[', |
457 | /* 92 - \ */ '\\', |
458 | /* 93 - ] */ ']', |
459 | /* 94 - ^ */ '^', |
460 | /* 95 - _ */ '_', |
461 | /* 96 - ` */ '`', |
462 | /* 97 - a */ 'a', |
463 | /* 98 - b */ 0x08, |
464 | /* 99 - c */ 'c', |
465 | /* 100 - d */ 'd', |
466 | /* 101 - e */ 'e', |
467 | /* 102 - f */ 0x0C, |
468 | /* 103 - g */ 'g', |
469 | /* 104 - h */ 'h', |
470 | /* 105 - i */ 'i', |
471 | /* 106 - j */ 'j', |
472 | /* 107 - k */ 'k', |
473 | /* 108 - l */ 'l', |
474 | /* 109 - m */ 'm', |
475 | /* 110 - n */ 0x0A, |
476 | /* 111 - o */ 'o', |
477 | /* 112 - p */ 'p', |
478 | /* 113 - q */ 'q', |
479 | /* 114 - r */ 0x0D, |
480 | /* 115 - s */ 's', |
481 | /* 116 - t */ 0x09, |
482 | /* 117 - u */ 0, |
483 | /* 118 - v */ 0x0B, |
484 | /* 119 - w */ 'w', |
485 | /* 120 - x */ 0, |
486 | /* 121 - y */ 'y', |
487 | /* 122 - z */ 'z', |
488 | /* 123 - { */ '{', |
489 | /* 124 - | */ '|', |
490 | /* 125 - } */ '}', |
491 | /* 126 - ~ */ '~', |
492 | /* 127 - Delete */ 0 |
493 | }; |
494 | |
495 | template <typename T> |
496 | Lexer<T>::Lexer(VM& vm, JSParserBuiltinMode builtinMode, JSParserScriptMode scriptMode) |
497 | : m_isReparsingFunction(false) |
498 | , m_vm(vm) |
499 | , m_parsingBuiltinFunction(builtinMode == JSParserBuiltinMode::Builtin) |
500 | , m_scriptMode(scriptMode) |
501 | { |
502 | } |
503 | |
504 | static inline JSTokenType tokenTypeForIntegerLikeToken(double doubleValue) |
505 | { |
506 | if ((doubleValue || !std::signbit(doubleValue)) && static_cast<int64_t>(doubleValue) == doubleValue) |
507 | return INTEGER; |
508 | return DOUBLE; |
509 | } |
510 | |
511 | template <typename T> |
512 | Lexer<T>::~Lexer() |
513 | { |
514 | } |
515 | |
516 | template <typename T> |
517 | String Lexer<T>::invalidCharacterMessage() const |
518 | { |
519 | switch (m_current) { |
520 | case 0: |
521 | return "Invalid character: '\\0'"_s ; |
522 | case 10: |
523 | return "Invalid character: '\\n'"_s ; |
524 | case 11: |
525 | return "Invalid character: '\\v'"_s ; |
526 | case 13: |
527 | return "Invalid character: '\\r'"_s ; |
528 | case 35: |
529 | return "Invalid character: '#'"_s ; |
530 | case 64: |
531 | return "Invalid character: '@'"_s ; |
532 | case 96: |
533 | return "Invalid character: '`'"_s ; |
534 | default: |
535 | return makeString("Invalid character '\\u" , hex(m_current, 4, Lowercase), '\''); |
536 | } |
537 | } |
538 | |
539 | template <typename T> |
540 | ALWAYS_INLINE const T* Lexer<T>::currentSourcePtr() const |
541 | { |
542 | ASSERT(m_code <= m_codeEnd); |
543 | return m_code; |
544 | } |
545 | |
546 | template <typename T> |
547 | void Lexer<T>::setCode(const SourceCode& source, ParserArena* arena) |
548 | { |
549 | m_arena = &arena->identifierArena(); |
550 | |
551 | m_lineNumber = source.firstLine().oneBasedInt(); |
552 | m_lastToken = -1; |
553 | |
554 | StringView sourceString = source.provider()->source(); |
555 | |
556 | if (!sourceString.isNull()) |
557 | setCodeStart(sourceString); |
558 | else |
559 | m_codeStart = 0; |
560 | |
561 | m_source = &source; |
562 | m_sourceOffset = source.startOffset(); |
563 | m_codeStartPlusOffset = m_codeStart + source.startOffset(); |
564 | m_code = m_codeStartPlusOffset; |
565 | m_codeEnd = m_codeStart + source.endOffset(); |
566 | m_error = false; |
567 | m_atLineStart = true; |
568 | m_lineStart = m_code; |
569 | m_lexErrorMessage = String(); |
570 | m_sourceURLDirective = String(); |
571 | m_sourceMappingURLDirective = String(); |
572 | |
573 | m_buffer8.reserveInitialCapacity(initialReadBufferCapacity); |
574 | m_buffer16.reserveInitialCapacity(initialReadBufferCapacity); |
575 | m_bufferForRawTemplateString16.reserveInitialCapacity(initialReadBufferCapacity); |
576 | |
577 | if (LIKELY(m_code < m_codeEnd)) |
578 | m_current = *m_code; |
579 | else |
580 | m_current = 0; |
581 | ASSERT(currentOffset() == source.startOffset()); |
582 | } |
583 | |
584 | template <typename T> |
585 | template <int shiftAmount> ALWAYS_INLINE void Lexer<T>::internalShift() |
586 | { |
587 | m_code += shiftAmount; |
588 | ASSERT(currentOffset() >= currentLineStartOffset()); |
589 | m_current = *m_code; |
590 | } |
591 | |
592 | template <typename T> |
593 | ALWAYS_INLINE void Lexer<T>::shift() |
594 | { |
595 | // At one point timing showed that setting m_current to 0 unconditionally was faster than an if-else sequence. |
596 | m_current = 0; |
597 | ++m_code; |
598 | if (LIKELY(m_code < m_codeEnd)) |
599 | m_current = *m_code; |
600 | } |
601 | |
602 | template <typename T> |
603 | ALWAYS_INLINE bool Lexer<T>::atEnd() const |
604 | { |
605 | ASSERT(!m_current || m_code < m_codeEnd); |
606 | return UNLIKELY(UNLIKELY(!m_current) && m_code == m_codeEnd); |
607 | } |
608 | |
609 | template <typename T> |
610 | ALWAYS_INLINE T Lexer<T>::peek(int offset) const |
611 | { |
612 | ASSERT(offset > 0 && offset < 5); |
613 | const T* code = m_code + offset; |
614 | return (code < m_codeEnd) ? *code : 0; |
615 | } |
616 | |
617 | struct ParsedUnicodeEscapeValue { |
618 | ParsedUnicodeEscapeValue(UChar32 value) |
619 | : m_value(value) |
620 | { |
621 | ASSERT(isValid()); |
622 | } |
623 | |
624 | enum SpecialValueType { Incomplete = -2, Invalid = -1 }; |
625 | ParsedUnicodeEscapeValue(SpecialValueType type) |
626 | : m_value(type) |
627 | { |
628 | } |
629 | |
630 | bool isValid() const { return m_value >= 0; } |
631 | bool isIncomplete() const { return m_value == Incomplete; } |
632 | |
633 | UChar32 value() const |
634 | { |
635 | ASSERT(isValid()); |
636 | return m_value; |
637 | } |
638 | |
639 | private: |
640 | UChar32 m_value; |
641 | }; |
642 | |
643 | template<typename CharacterType> |
644 | ParsedUnicodeEscapeValue Lexer<CharacterType>::parseUnicodeEscape() |
645 | { |
646 | if (m_current == '{') { |
647 | shift(); |
648 | UChar32 codePoint = 0; |
649 | do { |
650 | if (!isASCIIHexDigit(m_current)) |
651 | return m_current ? ParsedUnicodeEscapeValue::Invalid : ParsedUnicodeEscapeValue::Incomplete; |
652 | codePoint = (codePoint << 4) | toASCIIHexValue(m_current); |
653 | if (codePoint > UCHAR_MAX_VALUE) { |
654 | // For raw template literal syntax, we consume `NotEscapeSequence`. |
655 | // Here, we consume NotCodePoint's HexDigits. |
656 | // |
657 | // NotEscapeSequence :: |
658 | // u { [lookahread not one of HexDigit] |
659 | // u { NotCodePoint |
660 | // u { CodePoint [lookahead != }] |
661 | // |
662 | // NotCodePoint :: |
663 | // HexDigits but not if MV of HexDigits <= 0x10FFFF |
664 | // |
665 | // CodePoint :: |
666 | // HexDigits but not if MV of HexDigits > 0x10FFFF |
667 | shift(); |
668 | while (isASCIIHexDigit(m_current)) |
669 | shift(); |
670 | |
671 | return atEnd() ? ParsedUnicodeEscapeValue::Incomplete : ParsedUnicodeEscapeValue::Invalid; |
672 | } |
673 | shift(); |
674 | } while (m_current != '}'); |
675 | shift(); |
676 | return codePoint; |
677 | } |
678 | |
679 | auto character2 = peek(1); |
680 | auto character3 = peek(2); |
681 | auto character4 = peek(3); |
682 | if (UNLIKELY(!isASCIIHexDigit(m_current) || !isASCIIHexDigit(character2) || !isASCIIHexDigit(character3) || !isASCIIHexDigit(character4))) { |
683 | auto result = (m_code + 4) >= m_codeEnd ? ParsedUnicodeEscapeValue::Incomplete : ParsedUnicodeEscapeValue::Invalid; |
684 | |
685 | // For raw template literal syntax, we consume `NotEscapeSequence`. |
686 | // |
687 | // NotEscapeSequence :: |
688 | // u [lookahead not one of HexDigit][lookahead != {] |
689 | // u HexDigit [lookahead not one of HexDigit] |
690 | // u HexDigit HexDigit [lookahead not one of HexDigit] |
691 | // u HexDigit HexDigit HexDigit [lookahead not one of HexDigit] |
692 | while (isASCIIHexDigit(m_current)) |
693 | shift(); |
694 | |
695 | return result; |
696 | } |
697 | |
698 | auto result = convertUnicode(m_current, character2, character3, character4); |
699 | shift(); |
700 | shift(); |
701 | shift(); |
702 | shift(); |
703 | return result; |
704 | } |
705 | |
706 | template <typename T> |
707 | void Lexer<T>::shiftLineTerminator() |
708 | { |
709 | ASSERT(isLineTerminator(m_current)); |
710 | |
711 | m_positionBeforeLastNewline = currentPosition(); |
712 | T prev = m_current; |
713 | shift(); |
714 | |
715 | if (prev == '\r' && m_current == '\n') |
716 | shift(); |
717 | |
718 | ++m_lineNumber; |
719 | } |
720 | |
721 | template <typename T> |
722 | ALWAYS_INLINE bool Lexer<T>::lastTokenWasRestrKeyword() const |
723 | { |
724 | return m_lastToken == CONTINUE || m_lastToken == BREAK || m_lastToken == RETURN || m_lastToken == THROW; |
725 | } |
726 | |
727 | template <typename T> |
728 | ALWAYS_INLINE void Lexer<T>::skipWhitespace() |
729 | { |
730 | while (isWhiteSpace(m_current)) |
731 | shift(); |
732 | } |
733 | |
734 | static NEVER_INLINE bool isNonLatin1IdentStart(UChar c) |
735 | { |
736 | return u_hasBinaryProperty(c, UCHAR_ID_START); |
737 | } |
738 | |
739 | static inline bool isIdentStart(LChar c) |
740 | { |
741 | return typesOfLatin1Characters[c] == CharacterIdentifierStart; |
742 | } |
743 | |
744 | static inline bool isIdentStart(UChar32 c) |
745 | { |
746 | return isLatin1(c) ? isIdentStart(static_cast<LChar>(c)) : isNonLatin1IdentStart(c); |
747 | } |
748 | |
749 | static NEVER_INLINE bool isNonLatin1IdentPart(UChar32 c) |
750 | { |
751 | return u_hasBinaryProperty(c, UCHAR_ID_CONTINUE) || c == 0x200C || c == 0x200D; |
752 | } |
753 | |
754 | static ALWAYS_INLINE bool isIdentPart(LChar c) |
755 | { |
756 | // Character types are divided into two groups depending on whether they can be part of an |
757 | // identifier or not. Those whose type value is less or equal than CharacterOtherIdentifierPart can be |
758 | // part of an identifier. (See the CharacterType definition for more details.) |
759 | return typesOfLatin1Characters[c] <= CharacterOtherIdentifierPart; |
760 | } |
761 | |
762 | static ALWAYS_INLINE bool isIdentPart(UChar32 c) |
763 | { |
764 | return isLatin1(c) ? isIdentPart(static_cast<LChar>(c)) : isNonLatin1IdentPart(c); |
765 | } |
766 | |
767 | static ALWAYS_INLINE bool isIdentPart(UChar c) |
768 | { |
769 | return isIdentPart(static_cast<UChar32>(c)); |
770 | } |
771 | |
772 | template<typename CharacterType> ALWAYS_INLINE bool isIdentPartIncludingEscapeTemplate(const CharacterType* code, const CharacterType* codeEnd) |
773 | { |
774 | if (isIdentPart(code[0])) |
775 | return true; |
776 | |
777 | // Shortest sequence handled below is \u{0}, which is 5 characters. |
778 | if (!(code[0] == '\\' && codeEnd - code >= 5 && code[1] == 'u')) |
779 | return false; |
780 | |
781 | if (code[2] == '{') { |
782 | UChar32 codePoint = 0; |
783 | const CharacterType* pointer; |
784 | for (pointer = &code[3]; pointer < codeEnd; ++pointer) { |
785 | auto digit = *pointer; |
786 | if (!isASCIIHexDigit(digit)) |
787 | break; |
788 | codePoint = (codePoint << 4) | toASCIIHexValue(digit); |
789 | if (codePoint > UCHAR_MAX_VALUE) |
790 | return false; |
791 | } |
792 | return isIdentPart(codePoint) && pointer < codeEnd && *pointer == '}'; |
793 | } |
794 | |
795 | // Shortest sequence handled below is \uXXXX, which is 6 characters. |
796 | if (codeEnd - code < 6) |
797 | return false; |
798 | |
799 | auto character1 = code[2]; |
800 | auto character2 = code[3]; |
801 | auto character3 = code[4]; |
802 | auto character4 = code[5]; |
803 | return isASCIIHexDigit(character1) && isASCIIHexDigit(character2) && isASCIIHexDigit(character3) && isASCIIHexDigit(character4) |
804 | && isIdentPart(Lexer<LChar>::convertUnicode(character1, character2, character3, character4)); |
805 | } |
806 | |
807 | static ALWAYS_INLINE bool isIdentPartIncludingEscape(const LChar* code, const LChar* codeEnd) |
808 | { |
809 | return isIdentPartIncludingEscapeTemplate(code, codeEnd); |
810 | } |
811 | |
812 | static ALWAYS_INLINE bool isIdentPartIncludingEscape(const UChar* code, const UChar* codeEnd) |
813 | { |
814 | return isIdentPartIncludingEscapeTemplate(code, codeEnd); |
815 | } |
816 | |
817 | template<typename CharacterType> |
818 | static inline bool isASCIIDigitOrSeparator(CharacterType character) |
819 | { |
820 | return isASCIIDigit(character) || character == '_'; |
821 | } |
822 | |
823 | template<typename CharacterType> |
824 | static inline bool isASCIIHexDigitOrSeparator(CharacterType character) |
825 | { |
826 | return isASCIIHexDigit(character) || character == '_'; |
827 | } |
828 | |
829 | template<typename CharacterType> |
830 | static inline bool isASCIIBinaryDigitOrSeparator(CharacterType character) |
831 | { |
832 | return isASCIIBinaryDigit(character) || character == '_'; |
833 | } |
834 | |
835 | template<typename CharacterType> |
836 | static inline bool isASCIIOctalDigitOrSeparator(CharacterType character) |
837 | { |
838 | return isASCIIOctalDigit(character) || character == '_'; |
839 | } |
840 | |
841 | static inline LChar singleEscape(int c) |
842 | { |
843 | if (c < 128) { |
844 | ASSERT(static_cast<size_t>(c) < WTF_ARRAY_LENGTH(singleCharacterEscapeValuesForASCII)); |
845 | return singleCharacterEscapeValuesForASCII[c]; |
846 | } |
847 | return 0; |
848 | } |
849 | |
850 | template <typename T> |
851 | inline void Lexer<T>::record8(int c) |
852 | { |
853 | ASSERT(isLatin1(c)); |
854 | m_buffer8.append(static_cast<LChar>(c)); |
855 | } |
856 | |
857 | template <typename T> |
858 | inline void Lexer<T>::append8(const T* p, size_t length) |
859 | { |
860 | size_t currentSize = m_buffer8.size(); |
861 | m_buffer8.grow(currentSize + length); |
862 | LChar* rawBuffer = m_buffer8.data() + currentSize; |
863 | |
864 | for (size_t i = 0; i < length; i++) { |
865 | T c = p[i]; |
866 | ASSERT(isLatin1(c)); |
867 | rawBuffer[i] = c; |
868 | } |
869 | } |
870 | |
871 | template <typename T> |
872 | inline void Lexer<T>::append16(const LChar* p, size_t length) |
873 | { |
874 | size_t currentSize = m_buffer16.size(); |
875 | m_buffer16.grow(currentSize + length); |
876 | UChar* rawBuffer = m_buffer16.data() + currentSize; |
877 | |
878 | for (size_t i = 0; i < length; i++) |
879 | rawBuffer[i] = p[i]; |
880 | } |
881 | |
882 | template <typename T> |
883 | inline void Lexer<T>::record16(T c) |
884 | { |
885 | m_buffer16.append(c); |
886 | } |
887 | |
888 | template <typename T> |
889 | inline void Lexer<T>::record16(int c) |
890 | { |
891 | ASSERT(c >= 0); |
892 | ASSERT(c <= static_cast<int>(USHRT_MAX)); |
893 | m_buffer16.append(static_cast<UChar>(c)); |
894 | } |
895 | |
896 | template<typename CharacterType> inline void Lexer<CharacterType>::recordUnicodeCodePoint(UChar32 codePoint) |
897 | { |
898 | ASSERT(codePoint >= 0); |
899 | ASSERT(codePoint <= UCHAR_MAX_VALUE); |
900 | if (U_IS_BMP(codePoint)) |
901 | record16(codePoint); |
902 | else { |
903 | UChar codeUnits[2] = { U16_LEAD(codePoint), U16_TRAIL(codePoint) }; |
904 | append16(codeUnits, 2); |
905 | } |
906 | } |
907 | |
908 | #if !ASSERT_DISABLED |
909 | bool isSafeBuiltinIdentifier(VM& vm, const Identifier* ident) |
910 | { |
911 | if (!ident) |
912 | return true; |
913 | /* Just block any use of suspicious identifiers. This is intended to |
914 | * be used as a safety net while implementing builtins. |
915 | */ |
916 | // FIXME: How can a debug-only assertion be a safety net? |
917 | if (*ident == vm.propertyNames->builtinNames().callPublicName()) |
918 | return false; |
919 | if (*ident == vm.propertyNames->builtinNames().applyPublicName()) |
920 | return false; |
921 | if (*ident == vm.propertyNames->eval) |
922 | return false; |
923 | if (*ident == vm.propertyNames->Function) |
924 | return false; |
925 | return true; |
926 | } |
927 | #endif |
928 | |
929 | template <> |
930 | template <bool shouldCreateIdentifier> ALWAYS_INLINE JSTokenType Lexer<LChar>::parseIdentifier(JSTokenData* tokenData, OptionSet<LexerFlags> lexerFlags, bool strictMode) |
931 | { |
932 | tokenData->escaped = false; |
933 | const ptrdiff_t remaining = m_codeEnd - m_code; |
934 | if ((remaining >= maxTokenLength) && !lexerFlags.contains(LexerFlags::IgnoreReservedWords)) { |
935 | JSTokenType keyword = parseKeyword<shouldCreateIdentifier>(tokenData); |
936 | if (keyword != IDENT) { |
937 | ASSERT((!shouldCreateIdentifier) || tokenData->ident); |
938 | return keyword == RESERVED_IF_STRICT && !strictMode ? IDENT : keyword; |
939 | } |
940 | } |
941 | |
942 | bool isPrivateName = m_current == '@' && m_parsingBuiltinFunction; |
943 | if (isPrivateName) |
944 | shift(); |
945 | |
946 | const LChar* identifierStart = currentSourcePtr(); |
947 | unsigned identifierLineStart = currentLineStartOffset(); |
948 | |
949 | while (isIdentPart(m_current)) |
950 | shift(); |
951 | |
952 | if (UNLIKELY(m_current == '\\')) { |
953 | setOffsetFromSourcePtr(identifierStart, identifierLineStart); |
954 | return parseIdentifierSlowCase<shouldCreateIdentifier>(tokenData, lexerFlags, strictMode); |
955 | } |
956 | |
957 | const Identifier* ident = nullptr; |
958 | |
959 | if (shouldCreateIdentifier || m_parsingBuiltinFunction) { |
960 | int identifierLength = currentSourcePtr() - identifierStart; |
961 | ident = makeIdentifier(identifierStart, identifierLength); |
962 | if (m_parsingBuiltinFunction) { |
963 | if (!isSafeBuiltinIdentifier(m_vm, ident) && !isPrivateName) { |
964 | m_lexErrorMessage = makeString("The use of '" , ident->string(), "' is disallowed in builtin functions." ); |
965 | return ERRORTOK; |
966 | } |
967 | if (isPrivateName) |
968 | ident = &m_arena->makeIdentifier(m_vm, m_vm.propertyNames->lookUpPrivateName(*ident)); |
969 | else if (*ident == m_vm.propertyNames->undefinedKeyword) |
970 | tokenData->ident = &m_vm.propertyNames->undefinedPrivateName; |
971 | if (!ident) |
972 | return INVALID_PRIVATE_NAME_ERRORTOK; |
973 | } |
974 | tokenData->ident = ident; |
975 | } else |
976 | tokenData->ident = nullptr; |
977 | |
978 | if (UNLIKELY((remaining < maxTokenLength) && !lexerFlags.contains(LexerFlags::IgnoreReservedWords)) && !isPrivateName) { |
979 | ASSERT(shouldCreateIdentifier); |
980 | if (remaining < maxTokenLength) { |
981 | const HashTableValue* entry = JSC::mainTable.entry(*ident); |
982 | ASSERT((remaining < maxTokenLength) || !entry); |
983 | if (!entry) |
984 | return IDENT; |
985 | JSTokenType token = static_cast<JSTokenType>(entry->lexerValue()); |
986 | return (token != RESERVED_IF_STRICT) || strictMode ? token : IDENT; |
987 | } |
988 | return IDENT; |
989 | } |
990 | |
991 | return IDENT; |
992 | } |
993 | |
994 | template <> |
995 | template <bool shouldCreateIdentifier> ALWAYS_INLINE JSTokenType Lexer<UChar>::parseIdentifier(JSTokenData* tokenData, OptionSet<LexerFlags> lexerFlags, bool strictMode) |
996 | { |
997 | tokenData->escaped = false; |
998 | const ptrdiff_t remaining = m_codeEnd - m_code; |
999 | if ((remaining >= maxTokenLength) && !lexerFlags.contains(LexerFlags::IgnoreReservedWords)) { |
1000 | JSTokenType keyword = parseKeyword<shouldCreateIdentifier>(tokenData); |
1001 | if (keyword != IDENT) { |
1002 | ASSERT((!shouldCreateIdentifier) || tokenData->ident); |
1003 | return keyword == RESERVED_IF_STRICT && !strictMode ? IDENT : keyword; |
1004 | } |
1005 | } |
1006 | |
1007 | bool isPrivateName = m_current == '@' && m_parsingBuiltinFunction; |
1008 | if (isPrivateName) |
1009 | shift(); |
1010 | |
1011 | const UChar* identifierStart = currentSourcePtr(); |
1012 | int identifierLineStart = currentLineStartOffset(); |
1013 | |
1014 | UChar orAllChars = 0; |
1015 | |
1016 | while (isIdentPart(m_current)) { |
1017 | orAllChars |= m_current; |
1018 | shift(); |
1019 | } |
1020 | |
1021 | if (UNLIKELY(m_current == '\\')) { |
1022 | ASSERT(!isPrivateName); |
1023 | setOffsetFromSourcePtr(identifierStart, identifierLineStart); |
1024 | return parseIdentifierSlowCase<shouldCreateIdentifier>(tokenData, lexerFlags, strictMode); |
1025 | } |
1026 | |
1027 | bool isAll8Bit = false; |
1028 | |
1029 | if (!(orAllChars & ~0xff)) |
1030 | isAll8Bit = true; |
1031 | |
1032 | const Identifier* ident = nullptr; |
1033 | |
1034 | if (shouldCreateIdentifier || m_parsingBuiltinFunction) { |
1035 | int identifierLength = currentSourcePtr() - identifierStart; |
1036 | if (isAll8Bit) |
1037 | ident = makeIdentifierLCharFromUChar(identifierStart, identifierLength); |
1038 | else |
1039 | ident = makeIdentifier(identifierStart, identifierLength); |
1040 | if (m_parsingBuiltinFunction) { |
1041 | if (!isSafeBuiltinIdentifier(m_vm, ident) && !isPrivateName) { |
1042 | m_lexErrorMessage = makeString("The use of '" , ident->string(), "' is disallowed in builtin functions." ); |
1043 | return ERRORTOK; |
1044 | } |
1045 | if (isPrivateName) |
1046 | ident = &m_arena->makeIdentifier(m_vm, m_vm.propertyNames->lookUpPrivateName(*ident)); |
1047 | else if (*ident == m_vm.propertyNames->undefinedKeyword) |
1048 | tokenData->ident = &m_vm.propertyNames->undefinedPrivateName; |
1049 | if (!ident) |
1050 | return INVALID_PRIVATE_NAME_ERRORTOK; |
1051 | } |
1052 | tokenData->ident = ident; |
1053 | } else |
1054 | tokenData->ident = nullptr; |
1055 | |
1056 | if (UNLIKELY((remaining < maxTokenLength) && !lexerFlags.contains(LexerFlags::IgnoreReservedWords)) && !isPrivateName) { |
1057 | ASSERT(shouldCreateIdentifier); |
1058 | if (remaining < maxTokenLength) { |
1059 | const HashTableValue* entry = JSC::mainTable.entry(*ident); |
1060 | ASSERT((remaining < maxTokenLength) || !entry); |
1061 | if (!entry) |
1062 | return IDENT; |
1063 | JSTokenType token = static_cast<JSTokenType>(entry->lexerValue()); |
1064 | return (token != RESERVED_IF_STRICT) || strictMode ? token : IDENT; |
1065 | } |
1066 | return IDENT; |
1067 | } |
1068 | |
1069 | return IDENT; |
1070 | } |
1071 | |
1072 | template<typename CharacterType> template<bool shouldCreateIdentifier> JSTokenType Lexer<CharacterType>::parseIdentifierSlowCase(JSTokenData* tokenData, OptionSet<LexerFlags> lexerFlags, bool strictMode) |
1073 | { |
1074 | tokenData->escaped = true; |
1075 | auto identifierStart = currentSourcePtr(); |
1076 | bool bufferRequired = false; |
1077 | |
1078 | while (true) { |
1079 | if (LIKELY(isIdentPart(m_current))) { |
1080 | shift(); |
1081 | continue; |
1082 | } |
1083 | if (LIKELY(m_current != '\\')) |
1084 | break; |
1085 | |
1086 | // \uXXXX unicode characters. |
1087 | bufferRequired = true; |
1088 | if (identifierStart != currentSourcePtr()) |
1089 | m_buffer16.append(identifierStart, currentSourcePtr() - identifierStart); |
1090 | shift(); |
1091 | if (UNLIKELY(m_current != 'u')) |
1092 | return atEnd() ? UNTERMINATED_IDENTIFIER_ESCAPE_ERRORTOK : INVALID_IDENTIFIER_ESCAPE_ERRORTOK; |
1093 | shift(); |
1094 | auto character = parseUnicodeEscape(); |
1095 | if (UNLIKELY(!character.isValid())) |
1096 | return character.isIncomplete() ? UNTERMINATED_IDENTIFIER_UNICODE_ESCAPE_ERRORTOK : INVALID_IDENTIFIER_UNICODE_ESCAPE_ERRORTOK; |
1097 | if (UNLIKELY(m_buffer16.size() ? !isIdentPart(character.value()) : !isIdentStart(character.value()))) |
1098 | return INVALID_IDENTIFIER_UNICODE_ESCAPE_ERRORTOK; |
1099 | if (shouldCreateIdentifier) |
1100 | recordUnicodeCodePoint(character.value()); |
1101 | identifierStart = currentSourcePtr(); |
1102 | } |
1103 | |
1104 | int identifierLength; |
1105 | const Identifier* ident = nullptr; |
1106 | if (shouldCreateIdentifier) { |
1107 | if (!bufferRequired) { |
1108 | identifierLength = currentSourcePtr() - identifierStart; |
1109 | ident = makeIdentifier(identifierStart, identifierLength); |
1110 | } else { |
1111 | if (identifierStart != currentSourcePtr()) |
1112 | m_buffer16.append(identifierStart, currentSourcePtr() - identifierStart); |
1113 | ident = makeIdentifier(m_buffer16.data(), m_buffer16.size()); |
1114 | } |
1115 | |
1116 | tokenData->ident = ident; |
1117 | } else |
1118 | tokenData->ident = nullptr; |
1119 | |
1120 | m_buffer16.shrink(0); |
1121 | |
1122 | if (LIKELY(!lexerFlags.contains(LexerFlags::IgnoreReservedWords))) { |
1123 | ASSERT(shouldCreateIdentifier); |
1124 | const HashTableValue* entry = JSC::mainTable.entry(*ident); |
1125 | if (!entry) |
1126 | return IDENT; |
1127 | JSTokenType token = static_cast<JSTokenType>(entry->lexerValue()); |
1128 | if ((token != RESERVED_IF_STRICT) || strictMode) |
1129 | return bufferRequired ? UNEXPECTED_ESCAPE_ERRORTOK : token; |
1130 | } |
1131 | |
1132 | return IDENT; |
1133 | } |
1134 | |
1135 | static ALWAYS_INLINE bool characterRequiresParseStringSlowCase(LChar character) |
1136 | { |
1137 | return character < 0xE; |
1138 | } |
1139 | |
1140 | static ALWAYS_INLINE bool characterRequiresParseStringSlowCase(UChar character) |
1141 | { |
1142 | return character < 0xE || !isLatin1(character); |
1143 | } |
1144 | |
1145 | template <typename T> |
1146 | template <bool shouldBuildStrings> ALWAYS_INLINE typename Lexer<T>::StringParseResult Lexer<T>::parseString(JSTokenData* tokenData, bool strictMode) |
1147 | { |
1148 | int startingOffset = currentOffset(); |
1149 | int startingLineStartOffset = currentLineStartOffset(); |
1150 | int startingLineNumber = lineNumber(); |
1151 | T stringQuoteCharacter = m_current; |
1152 | shift(); |
1153 | |
1154 | const T* stringStart = currentSourcePtr(); |
1155 | |
1156 | while (m_current != stringQuoteCharacter) { |
1157 | if (UNLIKELY(m_current == '\\')) { |
1158 | if (stringStart != currentSourcePtr() && shouldBuildStrings) |
1159 | append8(stringStart, currentSourcePtr() - stringStart); |
1160 | shift(); |
1161 | |
1162 | LChar escape = singleEscape(m_current); |
1163 | |
1164 | // Most common escape sequences first. |
1165 | if (escape) { |
1166 | if (shouldBuildStrings) |
1167 | record8(escape); |
1168 | shift(); |
1169 | } else if (UNLIKELY(isLineTerminator(m_current))) |
1170 | shiftLineTerminator(); |
1171 | else if (m_current == 'x') { |
1172 | shift(); |
1173 | if (!isASCIIHexDigit(m_current) || !isASCIIHexDigit(peek(1))) { |
1174 | m_lexErrorMessage = "\\x can only be followed by a hex character sequence"_s ; |
1175 | return (atEnd() || (isASCIIHexDigit(m_current) && (m_code + 1 == m_codeEnd))) ? StringUnterminated : StringCannotBeParsed; |
1176 | } |
1177 | T prev = m_current; |
1178 | shift(); |
1179 | if (shouldBuildStrings) |
1180 | record8(convertHex(prev, m_current)); |
1181 | shift(); |
1182 | } else { |
1183 | setOffset(startingOffset, startingLineStartOffset); |
1184 | setLineNumber(startingLineNumber); |
1185 | m_buffer8.shrink(0); |
1186 | return parseStringSlowCase<shouldBuildStrings>(tokenData, strictMode); |
1187 | } |
1188 | stringStart = currentSourcePtr(); |
1189 | continue; |
1190 | } |
1191 | |
1192 | if (UNLIKELY(characterRequiresParseStringSlowCase(m_current))) { |
1193 | setOffset(startingOffset, startingLineStartOffset); |
1194 | setLineNumber(startingLineNumber); |
1195 | m_buffer8.shrink(0); |
1196 | return parseStringSlowCase<shouldBuildStrings>(tokenData, strictMode); |
1197 | } |
1198 | |
1199 | shift(); |
1200 | } |
1201 | |
1202 | if (currentSourcePtr() != stringStart && shouldBuildStrings) |
1203 | append8(stringStart, currentSourcePtr() - stringStart); |
1204 | if (shouldBuildStrings) { |
1205 | tokenData->ident = makeIdentifier(m_buffer8.data(), m_buffer8.size()); |
1206 | m_buffer8.shrink(0); |
1207 | } else |
1208 | tokenData->ident = 0; |
1209 | |
1210 | return StringParsedSuccessfully; |
1211 | } |
1212 | |
1213 | template <typename T> |
1214 | template <bool shouldBuildStrings, LexerEscapeParseMode escapeParseMode> ALWAYS_INLINE auto Lexer<T>::parseComplexEscape(bool strictMode, T stringQuoteCharacter) -> StringParseResult |
1215 | { |
1216 | if (m_current == 'x') { |
1217 | shift(); |
1218 | if (!isASCIIHexDigit(m_current) || !isASCIIHexDigit(peek(1))) { |
1219 | // For raw template literal syntax, we consume `NotEscapeSequence`. |
1220 | // |
1221 | // NotEscapeSequence :: |
1222 | // x [lookahread not one of HexDigit] |
1223 | // x HexDigit [lookahread not one of HexDigit] |
1224 | if (isASCIIHexDigit(m_current)) |
1225 | shift(); |
1226 | ASSERT(!isASCIIHexDigit(m_current)); |
1227 | |
1228 | m_lexErrorMessage = "\\x can only be followed by a hex character sequence"_s ; |
1229 | return atEnd() ? StringUnterminated : StringCannotBeParsed; |
1230 | } |
1231 | |
1232 | T prev = m_current; |
1233 | shift(); |
1234 | if (shouldBuildStrings) |
1235 | record16(convertHex(prev, m_current)); |
1236 | shift(); |
1237 | |
1238 | return StringParsedSuccessfully; |
1239 | } |
1240 | |
1241 | if (m_current == 'u') { |
1242 | shift(); |
1243 | |
1244 | if (escapeParseMode == LexerEscapeParseMode::String && m_current == stringQuoteCharacter) { |
1245 | if (shouldBuildStrings) |
1246 | record16('u'); |
1247 | return StringParsedSuccessfully; |
1248 | } |
1249 | |
1250 | auto character = parseUnicodeEscape(); |
1251 | if (character.isValid()) { |
1252 | if (shouldBuildStrings) |
1253 | recordUnicodeCodePoint(character.value()); |
1254 | return StringParsedSuccessfully; |
1255 | } |
1256 | |
1257 | m_lexErrorMessage = "\\u can only be followed by a Unicode character sequence"_s ; |
1258 | return atEnd() ? StringUnterminated : StringCannotBeParsed; |
1259 | } |
1260 | |
1261 | if (strictMode) { |
1262 | if (isASCIIDigit(m_current)) { |
1263 | // The only valid numeric escape in strict mode is '\0', and this must not be followed by a decimal digit. |
1264 | int character1 = m_current; |
1265 | shift(); |
1266 | if (character1 != '0' || isASCIIDigit(m_current)) { |
1267 | // For raw template literal syntax, we consume `NotEscapeSequence`. |
1268 | // |
1269 | // NotEscapeSequence :: |
1270 | // 0 DecimalDigit |
1271 | // DecimalDigit but not 0 |
1272 | if (character1 == '0') |
1273 | shift(); |
1274 | |
1275 | m_lexErrorMessage = "The only valid numeric escape in strict mode is '\\0'"_s ; |
1276 | return atEnd() ? StringUnterminated : StringCannotBeParsed; |
1277 | } |
1278 | if (shouldBuildStrings) |
1279 | record16(0); |
1280 | return StringParsedSuccessfully; |
1281 | } |
1282 | } else { |
1283 | if (isASCIIOctalDigit(m_current)) { |
1284 | // Octal character sequences |
1285 | T character1 = m_current; |
1286 | shift(); |
1287 | if (isASCIIOctalDigit(m_current)) { |
1288 | // Two octal characters |
1289 | T character2 = m_current; |
1290 | shift(); |
1291 | if (character1 >= '0' && character1 <= '3' && isASCIIOctalDigit(m_current)) { |
1292 | if (shouldBuildStrings) |
1293 | record16((character1 - '0') * 64 + (character2 - '0') * 8 + m_current - '0'); |
1294 | shift(); |
1295 | } else { |
1296 | if (shouldBuildStrings) |
1297 | record16((character1 - '0') * 8 + character2 - '0'); |
1298 | } |
1299 | } else { |
1300 | if (shouldBuildStrings) |
1301 | record16(character1 - '0'); |
1302 | } |
1303 | return StringParsedSuccessfully; |
1304 | } |
1305 | } |
1306 | |
1307 | if (!atEnd()) { |
1308 | if (shouldBuildStrings) |
1309 | record16(m_current); |
1310 | shift(); |
1311 | return StringParsedSuccessfully; |
1312 | } |
1313 | |
1314 | m_lexErrorMessage = "Unterminated string constant"_s ; |
1315 | return StringUnterminated; |
1316 | } |
1317 | |
1318 | template <typename T> |
1319 | template <bool shouldBuildStrings> auto Lexer<T>::parseStringSlowCase(JSTokenData* tokenData, bool strictMode) -> StringParseResult |
1320 | { |
1321 | T stringQuoteCharacter = m_current; |
1322 | shift(); |
1323 | |
1324 | const T* stringStart = currentSourcePtr(); |
1325 | |
1326 | while (m_current != stringQuoteCharacter) { |
1327 | if (UNLIKELY(m_current == '\\')) { |
1328 | if (stringStart != currentSourcePtr() && shouldBuildStrings) |
1329 | append16(stringStart, currentSourcePtr() - stringStart); |
1330 | shift(); |
1331 | |
1332 | LChar escape = singleEscape(m_current); |
1333 | |
1334 | // Most common escape sequences first |
1335 | if (escape) { |
1336 | if (shouldBuildStrings) |
1337 | record16(escape); |
1338 | shift(); |
1339 | } else if (UNLIKELY(isLineTerminator(m_current))) |
1340 | shiftLineTerminator(); |
1341 | else { |
1342 | StringParseResult result = parseComplexEscape<shouldBuildStrings, LexerEscapeParseMode::String>(strictMode, stringQuoteCharacter); |
1343 | if (result != StringParsedSuccessfully) |
1344 | return result; |
1345 | } |
1346 | |
1347 | stringStart = currentSourcePtr(); |
1348 | continue; |
1349 | } |
1350 | // Fast check for characters that require special handling. |
1351 | // Catches 0, \n, and \r as efficiently as possible, and lets through all common ASCII characters. |
1352 | static_assert(std::is_unsigned<T>::value, "Lexer expects an unsigned character type" ); |
1353 | if (UNLIKELY(m_current < 0xE)) { |
1354 | // New-line or end of input is not allowed |
1355 | if (atEnd() || m_current == '\r' || m_current == '\n') { |
1356 | m_lexErrorMessage = "Unexpected EOF"_s ; |
1357 | return atEnd() ? StringUnterminated : StringCannotBeParsed; |
1358 | } |
1359 | // Anything else is just a normal character |
1360 | } |
1361 | shift(); |
1362 | } |
1363 | |
1364 | if (currentSourcePtr() != stringStart && shouldBuildStrings) |
1365 | append16(stringStart, currentSourcePtr() - stringStart); |
1366 | if (shouldBuildStrings) |
1367 | tokenData->ident = makeIdentifier(m_buffer16.data(), m_buffer16.size()); |
1368 | else |
1369 | tokenData->ident = 0; |
1370 | |
1371 | m_buffer16.shrink(0); |
1372 | return StringParsedSuccessfully; |
1373 | } |
1374 | |
1375 | template <typename T> |
1376 | typename Lexer<T>::StringParseResult Lexer<T>::parseTemplateLiteral(JSTokenData* tokenData, RawStringsBuildMode rawStringsBuildMode) |
1377 | { |
1378 | bool parseCookedFailed = false; |
1379 | const T* stringStart = currentSourcePtr(); |
1380 | const T* rawStringStart = currentSourcePtr(); |
1381 | |
1382 | while (m_current != '`') { |
1383 | if (UNLIKELY(m_current == '\\')) { |
1384 | if (stringStart != currentSourcePtr()) |
1385 | append16(stringStart, currentSourcePtr() - stringStart); |
1386 | shift(); |
1387 | |
1388 | LChar escape = singleEscape(m_current); |
1389 | |
1390 | // Most common escape sequences first. |
1391 | if (escape) { |
1392 | record16(escape); |
1393 | shift(); |
1394 | } else if (UNLIKELY(isLineTerminator(m_current))) { |
1395 | // Normalize <CR>, <CR><LF> to <LF>. |
1396 | if (m_current == '\r') { |
1397 | ASSERT_WITH_MESSAGE(rawStringStart != currentSourcePtr(), "We should have at least shifted the escape." ); |
1398 | |
1399 | if (rawStringsBuildMode == RawStringsBuildMode::BuildRawStrings) { |
1400 | m_bufferForRawTemplateString16.append(rawStringStart, currentSourcePtr() - rawStringStart); |
1401 | m_bufferForRawTemplateString16.append('\n'); |
1402 | } |
1403 | |
1404 | shiftLineTerminator(); |
1405 | rawStringStart = currentSourcePtr(); |
1406 | } else |
1407 | shiftLineTerminator(); |
1408 | } else { |
1409 | bool strictMode = true; |
1410 | StringParseResult result = parseComplexEscape<true, LexerEscapeParseMode::Template>(strictMode, '`'); |
1411 | if (result != StringParsedSuccessfully) { |
1412 | if (rawStringsBuildMode == RawStringsBuildMode::BuildRawStrings && result == StringCannotBeParsed) |
1413 | parseCookedFailed = true; |
1414 | else |
1415 | return result; |
1416 | } |
1417 | } |
1418 | |
1419 | stringStart = currentSourcePtr(); |
1420 | continue; |
1421 | } |
1422 | |
1423 | if (m_current == '$' && peek(1) == '{') |
1424 | break; |
1425 | |
1426 | // Fast check for characters that require special handling. |
1427 | // Catches 0, \n, \r, 0x2028, and 0x2029 as efficiently |
1428 | // as possible, and lets through all common ASCII characters. |
1429 | if (UNLIKELY(((static_cast<unsigned>(m_current) - 0xE) & 0x2000))) { |
1430 | // End of input is not allowed. |
1431 | // Unlike String, line terminator is allowed. |
1432 | if (atEnd()) { |
1433 | m_lexErrorMessage = "Unexpected EOF"_s ; |
1434 | return StringUnterminated; |
1435 | } |
1436 | |
1437 | if (isLineTerminator(m_current)) { |
1438 | if (m_current == '\r') { |
1439 | // Normalize <CR>, <CR><LF> to <LF>. |
1440 | if (stringStart != currentSourcePtr()) |
1441 | append16(stringStart, currentSourcePtr() - stringStart); |
1442 | if (rawStringStart != currentSourcePtr() && rawStringsBuildMode == RawStringsBuildMode::BuildRawStrings) |
1443 | m_bufferForRawTemplateString16.append(rawStringStart, currentSourcePtr() - rawStringStart); |
1444 | |
1445 | record16('\n'); |
1446 | if (rawStringsBuildMode == RawStringsBuildMode::BuildRawStrings) |
1447 | m_bufferForRawTemplateString16.append('\n'); |
1448 | shiftLineTerminator(); |
1449 | stringStart = currentSourcePtr(); |
1450 | rawStringStart = currentSourcePtr(); |
1451 | } else |
1452 | shiftLineTerminator(); |
1453 | continue; |
1454 | } |
1455 | // Anything else is just a normal character |
1456 | } |
1457 | |
1458 | shift(); |
1459 | } |
1460 | |
1461 | bool isTail = m_current == '`'; |
1462 | |
1463 | if (currentSourcePtr() != stringStart) |
1464 | append16(stringStart, currentSourcePtr() - stringStart); |
1465 | if (rawStringStart != currentSourcePtr() && rawStringsBuildMode == RawStringsBuildMode::BuildRawStrings) |
1466 | m_bufferForRawTemplateString16.append(rawStringStart, currentSourcePtr() - rawStringStart); |
1467 | |
1468 | if (!parseCookedFailed) |
1469 | tokenData->cooked = makeIdentifier(m_buffer16.data(), m_buffer16.size()); |
1470 | else |
1471 | tokenData->cooked = nullptr; |
1472 | |
1473 | // Line terminator normalization (e.g. <CR> => <LF>) should be applied to both the raw and cooked representations. |
1474 | if (rawStringsBuildMode == RawStringsBuildMode::BuildRawStrings) |
1475 | tokenData->raw = makeIdentifier(m_bufferForRawTemplateString16.data(), m_bufferForRawTemplateString16.size()); |
1476 | else |
1477 | tokenData->raw = nullptr; |
1478 | |
1479 | tokenData->isTail = isTail; |
1480 | |
1481 | m_buffer16.shrink(0); |
1482 | m_bufferForRawTemplateString16.shrink(0); |
1483 | |
1484 | if (isTail) { |
1485 | // Skip ` |
1486 | shift(); |
1487 | } else { |
1488 | // Skip $ and { |
1489 | shift(); |
1490 | shift(); |
1491 | } |
1492 | |
1493 | return StringParsedSuccessfully; |
1494 | } |
1495 | |
1496 | template <typename T> |
1497 | ALWAYS_INLINE auto Lexer<T>::parseHex() -> Optional<NumberParseResult> |
1498 | { |
1499 | ASSERT(isASCIIHexDigit(m_current)); |
1500 | |
1501 | // Optimization: most hexadecimal values fit into 4 bytes. |
1502 | uint32_t hexValue = 0; |
1503 | int maximumDigits = 7; |
1504 | |
1505 | do { |
1506 | if (m_current == '_') { |
1507 | if (UNLIKELY(!isASCIIHexDigit(peek(1)))) |
1508 | return WTF::nullopt; |
1509 | |
1510 | shift(); |
1511 | } |
1512 | |
1513 | hexValue = (hexValue << 4) + toASCIIHexValue(m_current); |
1514 | shift(); |
1515 | --maximumDigits; |
1516 | } while (isASCIIHexDigitOrSeparator(m_current) && maximumDigits >= 0); |
1517 | |
1518 | if (LIKELY(maximumDigits >= 0 && m_current != 'n')) |
1519 | return NumberParseResult { hexValue }; |
1520 | |
1521 | // No more place in the hexValue buffer. |
1522 | // The values are shifted out and placed into the m_buffer8 vector. |
1523 | for (int i = 0; i < 8; ++i) { |
1524 | int digit = hexValue >> 28; |
1525 | if (digit < 10) |
1526 | record8(digit + '0'); |
1527 | else |
1528 | record8(digit - 10 + 'a'); |
1529 | hexValue <<= 4; |
1530 | } |
1531 | |
1532 | while (isASCIIHexDigitOrSeparator(m_current)) { |
1533 | if (m_current == '_') { |
1534 | if (UNLIKELY(!isASCIIHexDigit(peek(1)))) |
1535 | return WTF::nullopt; |
1536 | |
1537 | shift(); |
1538 | } |
1539 | |
1540 | record8(m_current); |
1541 | shift(); |
1542 | } |
1543 | |
1544 | if (UNLIKELY(Options::useBigInt() && m_current == 'n')) |
1545 | return NumberParseResult { makeIdentifier(m_buffer8.data(), m_buffer8.size()) }; |
1546 | |
1547 | return NumberParseResult { parseIntOverflow(m_buffer8.data(), m_buffer8.size(), 16) }; |
1548 | } |
1549 | |
1550 | template <typename T> |
1551 | ALWAYS_INLINE auto Lexer<T>::parseBinary() -> Optional<NumberParseResult> |
1552 | { |
1553 | ASSERT(isASCIIBinaryDigit(m_current)); |
1554 | |
1555 | // Optimization: most binary values fit into 4 bytes. |
1556 | uint32_t binaryValue = 0; |
1557 | const unsigned maximumDigits = 32; |
1558 | int digit = maximumDigits - 1; |
1559 | // Temporary buffer for the digits. Makes easier |
1560 | // to reconstruct the input characters when needed. |
1561 | LChar digits[maximumDigits]; |
1562 | |
1563 | do { |
1564 | if (m_current == '_') { |
1565 | if (UNLIKELY(!isASCIIBinaryDigit(peek(1)))) |
1566 | return WTF::nullopt; |
1567 | |
1568 | shift(); |
1569 | } |
1570 | |
1571 | binaryValue = (binaryValue << 1) + (m_current - '0'); |
1572 | digits[digit] = m_current; |
1573 | shift(); |
1574 | --digit; |
1575 | } while (isASCIIBinaryDigitOrSeparator(m_current) && digit >= 0); |
1576 | |
1577 | if (LIKELY(!isASCIIDigitOrSeparator(m_current) && digit >= 0 && m_current != 'n')) |
1578 | return NumberParseResult { binaryValue }; |
1579 | |
1580 | for (int i = maximumDigits - 1; i > digit; --i) |
1581 | record8(digits[i]); |
1582 | |
1583 | while (isASCIIBinaryDigitOrSeparator(m_current)) { |
1584 | if (m_current == '_') { |
1585 | if (UNLIKELY(!isASCIIBinaryDigit(peek(1)))) |
1586 | return WTF::nullopt; |
1587 | |
1588 | shift(); |
1589 | } |
1590 | |
1591 | record8(m_current); |
1592 | shift(); |
1593 | } |
1594 | |
1595 | if (UNLIKELY(Options::useBigInt() && m_current == 'n')) |
1596 | return NumberParseResult { makeIdentifier(m_buffer8.data(), m_buffer8.size()) }; |
1597 | |
1598 | if (isASCIIDigit(m_current)) |
1599 | return WTF::nullopt; |
1600 | |
1601 | return NumberParseResult { parseIntOverflow(m_buffer8.data(), m_buffer8.size(), 2) }; |
1602 | } |
1603 | |
1604 | template <typename T> |
1605 | ALWAYS_INLINE auto Lexer<T>::parseOctal() -> Optional<NumberParseResult> |
1606 | { |
1607 | ASSERT(isASCIIOctalDigit(m_current)); |
1608 | ASSERT(!m_buffer8.size() || (m_buffer8.size() == 1 && m_buffer8[0] == '0')); |
1609 | bool isLegacyLiteral = m_buffer8.size(); |
1610 | |
1611 | // Optimization: most octal values fit into 4 bytes. |
1612 | uint32_t octalValue = 0; |
1613 | const unsigned maximumDigits = 10; |
1614 | int digit = maximumDigits - 1; |
1615 | // Temporary buffer for the digits. Makes easier |
1616 | // to reconstruct the input characters when needed. |
1617 | LChar digits[maximumDigits]; |
1618 | |
1619 | do { |
1620 | if (m_current == '_') { |
1621 | if (UNLIKELY(!isASCIIOctalDigit(peek(1)) || isLegacyLiteral)) |
1622 | return WTF::nullopt; |
1623 | |
1624 | shift(); |
1625 | } |
1626 | |
1627 | octalValue = octalValue * 8 + (m_current - '0'); |
1628 | digits[digit] = m_current; |
1629 | shift(); |
1630 | --digit; |
1631 | } while (isASCIIOctalDigitOrSeparator(m_current) && digit >= 0); |
1632 | |
1633 | if (LIKELY(!isASCIIDigitOrSeparator(m_current) && digit >= 0 && m_current != 'n')) |
1634 | return NumberParseResult { octalValue }; |
1635 | |
1636 | for (int i = maximumDigits - 1; i > digit; --i) |
1637 | record8(digits[i]); |
1638 | |
1639 | while (isASCIIOctalDigitOrSeparator(m_current)) { |
1640 | if (m_current == '_') { |
1641 | if (UNLIKELY(!isASCIIOctalDigit(peek(1)) || isLegacyLiteral)) |
1642 | return WTF::nullopt; |
1643 | |
1644 | shift(); |
1645 | } |
1646 | |
1647 | record8(m_current); |
1648 | shift(); |
1649 | } |
1650 | |
1651 | if (UNLIKELY(Options::useBigInt() && m_current == 'n') && !isLegacyLiteral) |
1652 | return NumberParseResult { makeIdentifier(m_buffer8.data(), m_buffer8.size()) }; |
1653 | |
1654 | if (isASCIIDigit(m_current)) |
1655 | return WTF::nullopt; |
1656 | |
1657 | return NumberParseResult { parseIntOverflow(m_buffer8.data(), m_buffer8.size(), 8) }; |
1658 | } |
1659 | |
1660 | template <typename T> |
1661 | ALWAYS_INLINE auto Lexer<T>::parseDecimal() -> Optional<NumberParseResult> |
1662 | { |
1663 | ASSERT(isASCIIDigit(m_current) || m_buffer8.size()); |
1664 | bool isLegacyLiteral = m_buffer8.size() && isASCIIDigitOrSeparator(m_current); |
1665 | |
1666 | // Optimization: most decimal values fit into 4 bytes. |
1667 | uint32_t decimalValue = 0; |
1668 | |
1669 | // Since parseOctal may be executed before parseDecimal, |
1670 | // the m_buffer8 may hold ascii digits. |
1671 | if (!m_buffer8.size()) { |
1672 | const unsigned maximumDigits = 10; |
1673 | int digit = maximumDigits - 1; |
1674 | // Temporary buffer for the digits. Makes easier |
1675 | // to reconstruct the input characters when needed. |
1676 | LChar digits[maximumDigits]; |
1677 | |
1678 | do { |
1679 | if (m_current == '_') { |
1680 | if (UNLIKELY(!isASCIIDigit(peek(1)) || isLegacyLiteral)) |
1681 | return WTF::nullopt; |
1682 | |
1683 | shift(); |
1684 | } |
1685 | |
1686 | decimalValue = decimalValue * 10 + (m_current - '0'); |
1687 | digits[digit] = m_current; |
1688 | shift(); |
1689 | --digit; |
1690 | } while (isASCIIDigitOrSeparator(m_current) && digit >= 0); |
1691 | |
1692 | if (digit >= 0 && m_current != '.' && !isASCIIAlphaCaselessEqual(m_current, 'e') && m_current != 'n') |
1693 | return NumberParseResult { decimalValue }; |
1694 | |
1695 | for (int i = maximumDigits - 1; i > digit; --i) |
1696 | record8(digits[i]); |
1697 | } |
1698 | |
1699 | while (isASCIIDigitOrSeparator(m_current)) { |
1700 | if (m_current == '_') { |
1701 | if (UNLIKELY(!isASCIIDigit(peek(1)) || isLegacyLiteral)) |
1702 | return WTF::nullopt; |
1703 | |
1704 | shift(); |
1705 | } |
1706 | |
1707 | record8(m_current); |
1708 | shift(); |
1709 | } |
1710 | |
1711 | if (UNLIKELY(Options::useBigInt() && m_current == 'n' && !isLegacyLiteral)) |
1712 | return NumberParseResult { makeIdentifier(m_buffer8.data(), m_buffer8.size()) }; |
1713 | |
1714 | return WTF::nullopt; |
1715 | } |
1716 | |
1717 | template <typename T> |
1718 | ALWAYS_INLINE bool Lexer<T>::parseNumberAfterDecimalPoint() |
1719 | { |
1720 | ASSERT(isASCIIDigit(m_current)); |
1721 | record8('.'); |
1722 | |
1723 | do { |
1724 | if (m_current == '_') { |
1725 | if (UNLIKELY(!isASCIIDigit(peek(1)))) |
1726 | return false; |
1727 | |
1728 | shift(); |
1729 | } |
1730 | |
1731 | record8(m_current); |
1732 | shift(); |
1733 | } while (isASCIIDigitOrSeparator(m_current)); |
1734 | |
1735 | return true; |
1736 | } |
1737 | |
1738 | template <typename T> |
1739 | ALWAYS_INLINE bool Lexer<T>::parseNumberAfterExponentIndicator() |
1740 | { |
1741 | record8('e'); |
1742 | shift(); |
1743 | if (m_current == '+' || m_current == '-') { |
1744 | record8(m_current); |
1745 | shift(); |
1746 | } |
1747 | |
1748 | if (!isASCIIDigit(m_current)) |
1749 | return false; |
1750 | |
1751 | do { |
1752 | if (m_current == '_') { |
1753 | if (UNLIKELY(!isASCIIDigit(peek(1)))) |
1754 | return false; |
1755 | |
1756 | shift(); |
1757 | } |
1758 | |
1759 | record8(m_current); |
1760 | shift(); |
1761 | } while (isASCIIDigitOrSeparator(m_current)); |
1762 | |
1763 | return true; |
1764 | } |
1765 | |
1766 | template <typename T> |
1767 | ALWAYS_INLINE bool Lexer<T>::() |
1768 | { |
1769 | while (true) { |
1770 | while (UNLIKELY(m_current == '*')) { |
1771 | shift(); |
1772 | if (m_current == '/') { |
1773 | shift(); |
1774 | return true; |
1775 | } |
1776 | } |
1777 | |
1778 | if (atEnd()) |
1779 | return false; |
1780 | |
1781 | if (isLineTerminator(m_current)) { |
1782 | shiftLineTerminator(); |
1783 | m_hasLineTerminatorBeforeToken = true; |
1784 | } else |
1785 | shift(); |
1786 | } |
1787 | } |
1788 | |
1789 | template <typename T> |
1790 | ALWAYS_INLINE void Lexer<T>::() |
1791 | { |
1792 | // sourceURL and sourceMappingURL directives. |
1793 | if (!consume("source" )) |
1794 | return; |
1795 | |
1796 | if (consume("URL=" )) { |
1797 | m_sourceURLDirective = parseCommentDirectiveValue(); |
1798 | return; |
1799 | } |
1800 | |
1801 | if (consume("MappingURL=" )) { |
1802 | m_sourceMappingURLDirective = parseCommentDirectiveValue(); |
1803 | return; |
1804 | } |
1805 | } |
1806 | |
1807 | template <typename T> |
1808 | ALWAYS_INLINE String Lexer<T>::() |
1809 | { |
1810 | skipWhitespace(); |
1811 | const T* stringStart = currentSourcePtr(); |
1812 | while (!isWhiteSpace(m_current) && !isLineTerminator(m_current) && m_current != '"' && m_current != '\'' && !atEnd()) |
1813 | shift(); |
1814 | const T* stringEnd = currentSourcePtr(); |
1815 | skipWhitespace(); |
1816 | |
1817 | if (!isLineTerminator(m_current) && !atEnd()) |
1818 | return String(); |
1819 | |
1820 | append8(stringStart, stringEnd - stringStart); |
1821 | String result = String(m_buffer8.data(), m_buffer8.size()); |
1822 | m_buffer8.shrink(0); |
1823 | return result; |
1824 | } |
1825 | |
1826 | template <typename T> |
1827 | template <unsigned length> |
1828 | ALWAYS_INLINE bool Lexer<T>::consume(const char (&input)[length]) |
1829 | { |
1830 | unsigned lengthToCheck = length - 1; // Ignore the ending NULL byte in the string literal. |
1831 | |
1832 | unsigned i = 0; |
1833 | for (; i < lengthToCheck && m_current == input[i]; i++) |
1834 | shift(); |
1835 | |
1836 | return i == lengthToCheck; |
1837 | } |
1838 | |
1839 | template <typename T> |
1840 | bool Lexer<T>::nextTokenIsColon() |
1841 | { |
1842 | const T* code = m_code; |
1843 | while (code < m_codeEnd && (isWhiteSpace(*code) || isLineTerminator(*code))) |
1844 | code++; |
1845 | |
1846 | return code < m_codeEnd && *code == ':'; |
1847 | } |
1848 | |
1849 | template <typename T> |
1850 | void Lexer<T>::fillTokenInfo(JSToken* tokenRecord, JSTokenType token, int lineNumber, int endOffset, int lineStartOffset, JSTextPosition endPosition) |
1851 | { |
1852 | JSTokenLocation* tokenLocation = &tokenRecord->m_location; |
1853 | tokenLocation->line = lineNumber; |
1854 | tokenLocation->endOffset = endOffset; |
1855 | tokenLocation->lineStartOffset = lineStartOffset; |
1856 | ASSERT(tokenLocation->endOffset >= tokenLocation->lineStartOffset); |
1857 | tokenRecord->m_endPosition = endPosition; |
1858 | m_lastToken = token; |
1859 | } |
1860 | |
1861 | template <typename T> |
1862 | JSTokenType Lexer<T>::lexWithoutClearingLineTerminator(JSToken* tokenRecord, OptionSet<LexerFlags> lexerFlags, bool strictMode) |
1863 | { |
1864 | JSTokenData* tokenData = &tokenRecord->m_data; |
1865 | JSTokenLocation* tokenLocation = &tokenRecord->m_location; |
1866 | m_lastTokenLocation = JSTokenLocation(tokenRecord->m_location); |
1867 | |
1868 | ASSERT(!m_error); |
1869 | ASSERT(m_buffer8.isEmpty()); |
1870 | ASSERT(m_buffer16.isEmpty()); |
1871 | |
1872 | JSTokenType token = ERRORTOK; |
1873 | |
1874 | start: |
1875 | skipWhitespace(); |
1876 | |
1877 | tokenLocation->startOffset = currentOffset(); |
1878 | ASSERT(currentOffset() >= currentLineStartOffset()); |
1879 | tokenRecord->m_startPosition = currentPosition(); |
1880 | |
1881 | if (atEnd()) { |
1882 | token = EOFTOK; |
1883 | goto returnToken; |
1884 | } |
1885 | |
1886 | CharacterType type; |
1887 | if (LIKELY(isLatin1(m_current))) |
1888 | type = static_cast<CharacterType>(typesOfLatin1Characters[m_current]); |
1889 | else if (isNonLatin1IdentStart(m_current)) |
1890 | type = CharacterIdentifierStart; |
1891 | else if (isLineTerminator(m_current)) |
1892 | type = CharacterLineTerminator; |
1893 | else |
1894 | type = CharacterInvalid; |
1895 | |
1896 | switch (type) { |
1897 | case CharacterGreater: |
1898 | shift(); |
1899 | if (m_current == '>') { |
1900 | shift(); |
1901 | if (m_current == '>') { |
1902 | shift(); |
1903 | if (m_current == '=') { |
1904 | shift(); |
1905 | token = URSHIFTEQUAL; |
1906 | break; |
1907 | } |
1908 | token = URSHIFT; |
1909 | break; |
1910 | } |
1911 | if (m_current == '=') { |
1912 | shift(); |
1913 | token = RSHIFTEQUAL; |
1914 | break; |
1915 | } |
1916 | token = RSHIFT; |
1917 | break; |
1918 | } |
1919 | if (m_current == '=') { |
1920 | shift(); |
1921 | token = GE; |
1922 | break; |
1923 | } |
1924 | token = GT; |
1925 | break; |
1926 | case CharacterEqual: { |
1927 | if (peek(1) == '>') { |
1928 | token = ARROWFUNCTION; |
1929 | tokenData->line = lineNumber(); |
1930 | tokenData->offset = currentOffset(); |
1931 | tokenData->lineStartOffset = currentLineStartOffset(); |
1932 | ASSERT(tokenData->offset >= tokenData->lineStartOffset); |
1933 | shift(); |
1934 | shift(); |
1935 | break; |
1936 | } |
1937 | |
1938 | shift(); |
1939 | if (m_current == '=') { |
1940 | shift(); |
1941 | if (m_current == '=') { |
1942 | shift(); |
1943 | token = STREQ; |
1944 | break; |
1945 | } |
1946 | token = EQEQ; |
1947 | break; |
1948 | } |
1949 | token = EQUAL; |
1950 | break; |
1951 | } |
1952 | case CharacterLess: |
1953 | shift(); |
1954 | if (m_current == '!' && peek(1) == '-' && peek(2) == '-') { |
1955 | if (m_scriptMode == JSParserScriptMode::Classic) { |
1956 | // <!-- marks the beginning of a line comment (for www usage) |
1957 | goto inSingleLineComment; |
1958 | } |
1959 | } |
1960 | if (m_current == '<') { |
1961 | shift(); |
1962 | if (m_current == '=') { |
1963 | shift(); |
1964 | token = LSHIFTEQUAL; |
1965 | break; |
1966 | } |
1967 | token = LSHIFT; |
1968 | break; |
1969 | } |
1970 | if (m_current == '=') { |
1971 | shift(); |
1972 | token = LE; |
1973 | break; |
1974 | } |
1975 | token = LT; |
1976 | break; |
1977 | case CharacterExclamationMark: |
1978 | shift(); |
1979 | if (m_current == '=') { |
1980 | shift(); |
1981 | if (m_current == '=') { |
1982 | shift(); |
1983 | token = STRNEQ; |
1984 | break; |
1985 | } |
1986 | token = NE; |
1987 | break; |
1988 | } |
1989 | token = EXCLAMATION; |
1990 | break; |
1991 | case CharacterAdd: |
1992 | shift(); |
1993 | if (m_current == '+') { |
1994 | shift(); |
1995 | token = (!m_hasLineTerminatorBeforeToken) ? PLUSPLUS : AUTOPLUSPLUS; |
1996 | break; |
1997 | } |
1998 | if (m_current == '=') { |
1999 | shift(); |
2000 | token = PLUSEQUAL; |
2001 | break; |
2002 | } |
2003 | token = PLUS; |
2004 | break; |
2005 | case CharacterSub: |
2006 | shift(); |
2007 | if (m_current == '-') { |
2008 | shift(); |
2009 | if ((m_atLineStart || m_hasLineTerminatorBeforeToken) && m_current == '>') { |
2010 | if (m_scriptMode == JSParserScriptMode::Classic) { |
2011 | shift(); |
2012 | goto inSingleLineComment; |
2013 | } |
2014 | } |
2015 | token = (!m_hasLineTerminatorBeforeToken) ? MINUSMINUS : AUTOMINUSMINUS; |
2016 | break; |
2017 | } |
2018 | if (m_current == '=') { |
2019 | shift(); |
2020 | token = MINUSEQUAL; |
2021 | break; |
2022 | } |
2023 | token = MINUS; |
2024 | break; |
2025 | case CharacterMultiply: |
2026 | shift(); |
2027 | if (m_current == '=') { |
2028 | shift(); |
2029 | token = MULTEQUAL; |
2030 | break; |
2031 | } |
2032 | if (m_current == '*') { |
2033 | shift(); |
2034 | if (m_current == '=') { |
2035 | shift(); |
2036 | token = POWEQUAL; |
2037 | break; |
2038 | } |
2039 | token = POW; |
2040 | break; |
2041 | } |
2042 | token = TIMES; |
2043 | break; |
2044 | case CharacterSlash: |
2045 | shift(); |
2046 | if (m_current == '/') { |
2047 | shift(); |
2048 | goto inSingleLineCommentCheckForDirectives; |
2049 | } |
2050 | if (m_current == '*') { |
2051 | shift(); |
2052 | if (parseMultilineComment()) |
2053 | goto start; |
2054 | m_lexErrorMessage = "Multiline comment was not closed properly"_s ; |
2055 | token = UNTERMINATED_MULTILINE_COMMENT_ERRORTOK; |
2056 | goto returnError; |
2057 | } |
2058 | if (m_current == '=') { |
2059 | shift(); |
2060 | token = DIVEQUAL; |
2061 | break; |
2062 | } |
2063 | token = DIVIDE; |
2064 | break; |
2065 | case CharacterAnd: |
2066 | shift(); |
2067 | if (m_current == '&') { |
2068 | shift(); |
2069 | token = AND; |
2070 | break; |
2071 | } |
2072 | if (m_current == '=') { |
2073 | shift(); |
2074 | token = ANDEQUAL; |
2075 | break; |
2076 | } |
2077 | token = BITAND; |
2078 | break; |
2079 | case CharacterXor: |
2080 | shift(); |
2081 | if (m_current == '=') { |
2082 | shift(); |
2083 | token = XOREQUAL; |
2084 | break; |
2085 | } |
2086 | token = BITXOR; |
2087 | break; |
2088 | case CharacterModulo: |
2089 | shift(); |
2090 | if (m_current == '=') { |
2091 | shift(); |
2092 | token = MODEQUAL; |
2093 | break; |
2094 | } |
2095 | token = MOD; |
2096 | break; |
2097 | case CharacterOr: |
2098 | shift(); |
2099 | if (m_current == '=') { |
2100 | shift(); |
2101 | token = OREQUAL; |
2102 | break; |
2103 | } |
2104 | if (m_current == '|') { |
2105 | shift(); |
2106 | token = OR; |
2107 | break; |
2108 | } |
2109 | token = BITOR; |
2110 | break; |
2111 | case CharacterOpenParen: |
2112 | token = OPENPAREN; |
2113 | tokenData->line = lineNumber(); |
2114 | tokenData->offset = currentOffset(); |
2115 | tokenData->lineStartOffset = currentLineStartOffset(); |
2116 | shift(); |
2117 | break; |
2118 | case CharacterCloseParen: |
2119 | token = CLOSEPAREN; |
2120 | shift(); |
2121 | break; |
2122 | case CharacterOpenBracket: |
2123 | token = OPENBRACKET; |
2124 | shift(); |
2125 | break; |
2126 | case CharacterCloseBracket: |
2127 | token = CLOSEBRACKET; |
2128 | shift(); |
2129 | break; |
2130 | case CharacterComma: |
2131 | token = COMMA; |
2132 | shift(); |
2133 | break; |
2134 | case CharacterColon: |
2135 | token = COLON; |
2136 | shift(); |
2137 | break; |
2138 | case CharacterQuestion: |
2139 | shift(); |
2140 | if (m_current == '?') { |
2141 | shift(); |
2142 | token = COALESCE; |
2143 | break; |
2144 | } |
2145 | if (m_current == '.' && !isASCIIDigit(peek(1))) { |
2146 | shift(); |
2147 | token = QUESTIONDOT; |
2148 | break; |
2149 | } |
2150 | token = QUESTION; |
2151 | break; |
2152 | case CharacterTilde: |
2153 | token = TILDE; |
2154 | shift(); |
2155 | break; |
2156 | case CharacterSemicolon: |
2157 | shift(); |
2158 | token = SEMICOLON; |
2159 | break; |
2160 | case CharacterBackQuote: |
2161 | shift(); |
2162 | token = BACKQUOTE; |
2163 | break; |
2164 | case CharacterOpenBrace: |
2165 | tokenData->line = lineNumber(); |
2166 | tokenData->offset = currentOffset(); |
2167 | tokenData->lineStartOffset = currentLineStartOffset(); |
2168 | ASSERT(tokenData->offset >= tokenData->lineStartOffset); |
2169 | shift(); |
2170 | token = OPENBRACE; |
2171 | break; |
2172 | case CharacterCloseBrace: |
2173 | tokenData->line = lineNumber(); |
2174 | tokenData->offset = currentOffset(); |
2175 | tokenData->lineStartOffset = currentLineStartOffset(); |
2176 | ASSERT(tokenData->offset >= tokenData->lineStartOffset); |
2177 | shift(); |
2178 | token = CLOSEBRACE; |
2179 | break; |
2180 | case CharacterDot: |
2181 | shift(); |
2182 | if (!isASCIIDigit(m_current)) { |
2183 | if (UNLIKELY((m_current == '.') && (peek(1) == '.'))) { |
2184 | shift(); |
2185 | shift(); |
2186 | token = DOTDOTDOT; |
2187 | break; |
2188 | } |
2189 | token = DOT; |
2190 | break; |
2191 | } |
2192 | if (UNLIKELY(!parseNumberAfterDecimalPoint())) { |
2193 | m_lexErrorMessage = "Non-number found after decimal point"_s ; |
2194 | token = atEnd() ? UNTERMINATED_NUMERIC_LITERAL_ERRORTOK : INVALID_NUMERIC_LITERAL_ERRORTOK; |
2195 | goto returnError; |
2196 | } |
2197 | token = DOUBLE; |
2198 | if (UNLIKELY(isASCIIAlphaCaselessEqual(m_current, 'e') && !parseNumberAfterExponentIndicator())) { |
2199 | m_lexErrorMessage = "Non-number found after exponent indicator"_s ; |
2200 | token = atEnd() ? UNTERMINATED_NUMERIC_LITERAL_ERRORTOK : INVALID_NUMERIC_LITERAL_ERRORTOK; |
2201 | goto returnError; |
2202 | } |
2203 | size_t parsedLength; |
2204 | tokenData->doubleValue = parseDouble(m_buffer8.data(), m_buffer8.size(), parsedLength); |
2205 | if (token == INTEGER) |
2206 | token = tokenTypeForIntegerLikeToken(tokenData->doubleValue); |
2207 | |
2208 | if (UNLIKELY(isIdentStart(m_current))) { |
2209 | m_lexErrorMessage = "No identifiers allowed directly after numeric literal"_s ; |
2210 | token = atEnd() ? UNTERMINATED_NUMERIC_LITERAL_ERRORTOK : INVALID_NUMERIC_LITERAL_ERRORTOK; |
2211 | goto returnError; |
2212 | } |
2213 | m_buffer8.shrink(0); |
2214 | break; |
2215 | case CharacterZero: |
2216 | shift(); |
2217 | if (isASCIIAlphaCaselessEqual(m_current, 'x')) { |
2218 | if (UNLIKELY(!isASCIIHexDigit(peek(1)))) { |
2219 | m_lexErrorMessage = "No hexadecimal digits after '0x'"_s ; |
2220 | token = UNTERMINATED_HEX_NUMBER_ERRORTOK; |
2221 | goto returnError; |
2222 | } |
2223 | |
2224 | // Shift out the 'x' prefix. |
2225 | shift(); |
2226 | |
2227 | auto parseNumberResult = parseHex(); |
2228 | if (!parseNumberResult) |
2229 | tokenData->doubleValue = 0; |
2230 | else if (WTF::holds_alternative<double>(*parseNumberResult)) |
2231 | tokenData->doubleValue = WTF::get<double>(*parseNumberResult); |
2232 | else { |
2233 | token = BIGINT; |
2234 | shift(); |
2235 | tokenData->bigIntString = WTF::get<const Identifier*>(*parseNumberResult); |
2236 | tokenData->radix = 16; |
2237 | } |
2238 | |
2239 | if (UNLIKELY(isIdentStart(m_current))) { |
2240 | m_lexErrorMessage = "No space between hexadecimal literal and identifier"_s ; |
2241 | token = UNTERMINATED_HEX_NUMBER_ERRORTOK; |
2242 | goto returnError; |
2243 | } |
2244 | if (LIKELY(token != BIGINT)) |
2245 | token = tokenTypeForIntegerLikeToken(tokenData->doubleValue); |
2246 | m_buffer8.shrink(0); |
2247 | break; |
2248 | } |
2249 | if (isASCIIAlphaCaselessEqual(m_current, 'b')) { |
2250 | if (UNLIKELY(!isASCIIBinaryDigit(peek(1)))) { |
2251 | m_lexErrorMessage = "No binary digits after '0b'"_s ; |
2252 | token = UNTERMINATED_BINARY_NUMBER_ERRORTOK; |
2253 | goto returnError; |
2254 | } |
2255 | |
2256 | // Shift out the 'b' prefix. |
2257 | shift(); |
2258 | |
2259 | auto parseNumberResult = parseBinary(); |
2260 | if (!parseNumberResult) |
2261 | tokenData->doubleValue = 0; |
2262 | else if (WTF::holds_alternative<double>(*parseNumberResult)) |
2263 | tokenData->doubleValue = WTF::get<double>(*parseNumberResult); |
2264 | else { |
2265 | token = BIGINT; |
2266 | shift(); |
2267 | tokenData->bigIntString = WTF::get<const Identifier*>(*parseNumberResult); |
2268 | tokenData->radix = 2; |
2269 | } |
2270 | |
2271 | if (UNLIKELY(isIdentStart(m_current))) { |
2272 | m_lexErrorMessage = "No space between binary literal and identifier"_s ; |
2273 | token = UNTERMINATED_BINARY_NUMBER_ERRORTOK; |
2274 | goto returnError; |
2275 | } |
2276 | if (LIKELY(token != BIGINT)) |
2277 | token = tokenTypeForIntegerLikeToken(tokenData->doubleValue); |
2278 | m_buffer8.shrink(0); |
2279 | break; |
2280 | } |
2281 | |
2282 | if (isASCIIAlphaCaselessEqual(m_current, 'o')) { |
2283 | if (UNLIKELY(!isASCIIOctalDigit(peek(1)))) { |
2284 | m_lexErrorMessage = "No octal digits after '0o'"_s ; |
2285 | token = UNTERMINATED_OCTAL_NUMBER_ERRORTOK; |
2286 | goto returnError; |
2287 | } |
2288 | |
2289 | // Shift out the 'o' prefix. |
2290 | shift(); |
2291 | |
2292 | auto parseNumberResult = parseOctal(); |
2293 | if (!parseNumberResult) |
2294 | tokenData->doubleValue = 0; |
2295 | else if (WTF::holds_alternative<double>(*parseNumberResult)) |
2296 | tokenData->doubleValue = WTF::get<double>(*parseNumberResult); |
2297 | else { |
2298 | token = BIGINT; |
2299 | shift(); |
2300 | tokenData->bigIntString = WTF::get<const Identifier*>(*parseNumberResult); |
2301 | tokenData->radix = 8; |
2302 | } |
2303 | |
2304 | if (UNLIKELY(isIdentStart(m_current))) { |
2305 | m_lexErrorMessage = "No space between octal literal and identifier"_s ; |
2306 | token = UNTERMINATED_OCTAL_NUMBER_ERRORTOK; |
2307 | goto returnError; |
2308 | } |
2309 | if (LIKELY(token != BIGINT)) |
2310 | token = tokenTypeForIntegerLikeToken(tokenData->doubleValue); |
2311 | m_buffer8.shrink(0); |
2312 | break; |
2313 | } |
2314 | |
2315 | if (UNLIKELY(m_current == '_')) { |
2316 | m_lexErrorMessage = "Numeric literals may not begin with 0_"_s ; |
2317 | token = UNTERMINATED_OCTAL_NUMBER_ERRORTOK; |
2318 | goto returnError; |
2319 | } |
2320 | |
2321 | record8('0'); |
2322 | if (UNLIKELY(strictMode && isASCIIDigit(m_current))) { |
2323 | m_lexErrorMessage = "Decimal integer literals with a leading zero are forbidden in strict mode"_s ; |
2324 | token = UNTERMINATED_OCTAL_NUMBER_ERRORTOK; |
2325 | goto returnError; |
2326 | } |
2327 | if (isASCIIOctalDigit(m_current)) { |
2328 | auto parseNumberResult = parseOctal(); |
2329 | if (parseNumberResult && WTF::holds_alternative<double>(*parseNumberResult)) { |
2330 | tokenData->doubleValue = WTF::get<double>(*parseNumberResult); |
2331 | token = tokenTypeForIntegerLikeToken(tokenData->doubleValue); |
2332 | } |
2333 | } |
2334 | FALLTHROUGH; |
2335 | case CharacterNumber: |
2336 | if (LIKELY(token != INTEGER && token != DOUBLE)) { |
2337 | auto parseNumberResult = parseDecimal(); |
2338 | if (parseNumberResult) { |
2339 | if (WTF::holds_alternative<double>(*parseNumberResult)) { |
2340 | tokenData->doubleValue = WTF::get<double>(*parseNumberResult); |
2341 | token = tokenTypeForIntegerLikeToken(tokenData->doubleValue); |
2342 | } else { |
2343 | token = BIGINT; |
2344 | shift(); |
2345 | tokenData->bigIntString = WTF::get<const Identifier*>(*parseNumberResult); |
2346 | tokenData->radix = 10; |
2347 | } |
2348 | } else { |
2349 | token = INTEGER; |
2350 | if (m_current == '.') { |
2351 | shift(); |
2352 | if (UNLIKELY(isASCIIDigit(m_current) && !parseNumberAfterDecimalPoint())) { |
2353 | m_lexErrorMessage = "Non-number found after decimal point"_s ; |
2354 | token = atEnd() ? UNTERMINATED_NUMERIC_LITERAL_ERRORTOK : INVALID_NUMERIC_LITERAL_ERRORTOK; |
2355 | goto returnError; |
2356 | } |
2357 | token = DOUBLE; |
2358 | } |
2359 | if (UNLIKELY(isASCIIAlphaCaselessEqual(m_current, 'e') && !parseNumberAfterExponentIndicator())) { |
2360 | m_lexErrorMessage = "Non-number found after exponent indicator"_s ; |
2361 | token = atEnd() ? UNTERMINATED_NUMERIC_LITERAL_ERRORTOK : INVALID_NUMERIC_LITERAL_ERRORTOK; |
2362 | goto returnError; |
2363 | } |
2364 | size_t parsedLength; |
2365 | tokenData->doubleValue = parseDouble(m_buffer8.data(), m_buffer8.size(), parsedLength); |
2366 | if (token == INTEGER) |
2367 | token = tokenTypeForIntegerLikeToken(tokenData->doubleValue); |
2368 | } |
2369 | } |
2370 | |
2371 | if (UNLIKELY(isIdentStart(m_current))) { |
2372 | m_lexErrorMessage = "No identifiers allowed directly after numeric literal"_s ; |
2373 | token = atEnd() ? UNTERMINATED_NUMERIC_LITERAL_ERRORTOK : INVALID_NUMERIC_LITERAL_ERRORTOK; |
2374 | goto returnError; |
2375 | } |
2376 | m_buffer8.shrink(0); |
2377 | break; |
2378 | case CharacterQuote: { |
2379 | StringParseResult result = StringCannotBeParsed; |
2380 | if (lexerFlags.contains(LexerFlags::DontBuildStrings)) |
2381 | result = parseString<false>(tokenData, strictMode); |
2382 | else |
2383 | result = parseString<true>(tokenData, strictMode); |
2384 | |
2385 | if (UNLIKELY(result != StringParsedSuccessfully)) { |
2386 | token = result == StringUnterminated ? UNTERMINATED_STRING_LITERAL_ERRORTOK : INVALID_STRING_LITERAL_ERRORTOK; |
2387 | goto returnError; |
2388 | } |
2389 | shift(); |
2390 | token = STRING; |
2391 | break; |
2392 | } |
2393 | case CharacterIdentifierStart: |
2394 | ASSERT(isIdentStart(m_current)); |
2395 | FALLTHROUGH; |
2396 | case CharacterBackSlash: |
2397 | parseIdent: |
2398 | if (lexerFlags.contains(LexerFlags::DontBuildKeywords)) |
2399 | token = parseIdentifier<false>(tokenData, lexerFlags, strictMode); |
2400 | else |
2401 | token = parseIdentifier<true>(tokenData, lexerFlags, strictMode); |
2402 | break; |
2403 | case CharacterLineTerminator: |
2404 | ASSERT(isLineTerminator(m_current)); |
2405 | shiftLineTerminator(); |
2406 | m_atLineStart = true; |
2407 | m_hasLineTerminatorBeforeToken = true; |
2408 | m_lineStart = m_code; |
2409 | goto start; |
2410 | case CharacterHash: |
2411 | // Hashbang is only permitted at the start of the source text. |
2412 | if (peek(1) == '!' && !currentOffset()) { |
2413 | shift(); |
2414 | shift(); |
2415 | goto inSingleLineComment; |
2416 | } |
2417 | goto invalidCharacter; |
2418 | case CharacterPrivateIdentifierStart: |
2419 | if (m_parsingBuiltinFunction) |
2420 | goto parseIdent; |
2421 | goto invalidCharacter; |
2422 | case CharacterOtherIdentifierPart: |
2423 | case CharacterInvalid: |
2424 | goto invalidCharacter; |
2425 | default: |
2426 | RELEASE_ASSERT_NOT_REACHED(); |
2427 | m_lexErrorMessage = "Internal Error"_s ; |
2428 | token = ERRORTOK; |
2429 | goto returnError; |
2430 | } |
2431 | |
2432 | m_atLineStart = false; |
2433 | goto returnToken; |
2434 | |
2435 | : |
2436 | // Script comment directives like "//# sourceURL=test.js". |
2437 | if (UNLIKELY((m_current == '#' || m_current == '@') && isWhiteSpace(peek(1)))) { |
2438 | shift(); |
2439 | shift(); |
2440 | parseCommentDirective(); |
2441 | } |
2442 | // Fall through to complete single line comment parsing. |
2443 | |
2444 | : |
2445 | { |
2446 | auto lineNumber = m_lineNumber; |
2447 | auto endOffset = currentOffset(); |
2448 | auto lineStartOffset = currentLineStartOffset(); |
2449 | auto endPosition = currentPosition(); |
2450 | |
2451 | while (!isLineTerminator(m_current)) { |
2452 | if (atEnd()) { |
2453 | token = EOFTOK; |
2454 | fillTokenInfo(tokenRecord, token, lineNumber, endOffset, lineStartOffset, endPosition); |
2455 | return token; |
2456 | } |
2457 | shift(); |
2458 | } |
2459 | shiftLineTerminator(); |
2460 | m_atLineStart = true; |
2461 | m_hasLineTerminatorBeforeToken = true; |
2462 | m_lineStart = m_code; |
2463 | if (!lastTokenWasRestrKeyword()) |
2464 | goto start; |
2465 | |
2466 | token = SEMICOLON; |
2467 | fillTokenInfo(tokenRecord, token, lineNumber, endOffset, lineStartOffset, endPosition); |
2468 | return token; |
2469 | } |
2470 | |
2471 | returnToken: |
2472 | fillTokenInfo(tokenRecord, token, m_lineNumber, currentOffset(), currentLineStartOffset(), currentPosition()); |
2473 | return token; |
2474 | |
2475 | invalidCharacter: |
2476 | m_lexErrorMessage = invalidCharacterMessage(); |
2477 | token = ERRORTOK; |
2478 | // Falls through to return error. |
2479 | |
2480 | returnError: |
2481 | m_error = true; |
2482 | fillTokenInfo(tokenRecord, token, m_lineNumber, currentOffset(), currentLineStartOffset(), currentPosition()); |
2483 | RELEASE_ASSERT(token & ErrorTokenFlag); |
2484 | return token; |
2485 | } |
2486 | |
2487 | template <typename T> |
2488 | static inline void orCharacter(UChar&, UChar); |
2489 | |
2490 | template <> |
2491 | inline void orCharacter<LChar>(UChar&, UChar) { } |
2492 | |
2493 | template <> |
2494 | inline void orCharacter<UChar>(UChar& orAccumulator, UChar character) |
2495 | { |
2496 | orAccumulator |= character; |
2497 | } |
2498 | |
2499 | template <typename T> |
2500 | JSTokenType Lexer<T>::scanRegExp(JSToken* tokenRecord, UChar patternPrefix) |
2501 | { |
2502 | JSTokenData* tokenData = &tokenRecord->m_data; |
2503 | ASSERT(m_buffer16.isEmpty()); |
2504 | |
2505 | bool lastWasEscape = false; |
2506 | bool inBrackets = false; |
2507 | UChar charactersOredTogether = 0; |
2508 | |
2509 | if (patternPrefix) { |
2510 | ASSERT(!isLineTerminator(patternPrefix)); |
2511 | ASSERT(patternPrefix != '/'); |
2512 | ASSERT(patternPrefix != '['); |
2513 | record16(patternPrefix); |
2514 | } |
2515 | |
2516 | while (true) { |
2517 | if (isLineTerminator(m_current) || atEnd()) { |
2518 | m_buffer16.shrink(0); |
2519 | JSTokenType token = UNTERMINATED_REGEXP_LITERAL_ERRORTOK; |
2520 | fillTokenInfo(tokenRecord, token, m_lineNumber, currentOffset(), currentLineStartOffset(), currentPosition()); |
2521 | m_error = true; |
2522 | m_lexErrorMessage = makeString("Unterminated regular expression literal '" , getToken(*tokenRecord), "'" ); |
2523 | return token; |
2524 | } |
2525 | |
2526 | T prev = m_current; |
2527 | |
2528 | shift(); |
2529 | |
2530 | if (prev == '/' && !lastWasEscape && !inBrackets) |
2531 | break; |
2532 | |
2533 | record16(prev); |
2534 | orCharacter<T>(charactersOredTogether, prev); |
2535 | |
2536 | if (lastWasEscape) { |
2537 | lastWasEscape = false; |
2538 | continue; |
2539 | } |
2540 | |
2541 | switch (prev) { |
2542 | case '[': |
2543 | inBrackets = true; |
2544 | break; |
2545 | case ']': |
2546 | inBrackets = false; |
2547 | break; |
2548 | case '\\': |
2549 | lastWasEscape = true; |
2550 | break; |
2551 | } |
2552 | } |
2553 | |
2554 | tokenData->pattern = makeRightSizedIdentifier(m_buffer16.data(), m_buffer16.size(), charactersOredTogether); |
2555 | |
2556 | m_buffer16.shrink(0); |
2557 | charactersOredTogether = 0; |
2558 | |
2559 | while (isIdentPart(m_current)) { |
2560 | record16(m_current); |
2561 | orCharacter<T>(charactersOredTogether, m_current); |
2562 | shift(); |
2563 | } |
2564 | |
2565 | tokenData->flags = makeRightSizedIdentifier(m_buffer16.data(), m_buffer16.size(), charactersOredTogether); |
2566 | m_buffer16.shrink(0); |
2567 | |
2568 | // Since RegExp always ends with /, m_atLineStart always becomes false. |
2569 | m_atLineStart = false; |
2570 | |
2571 | JSTokenType token = REGEXP; |
2572 | fillTokenInfo(tokenRecord, token, m_lineNumber, currentOffset(), currentLineStartOffset(), currentPosition()); |
2573 | return token; |
2574 | } |
2575 | |
2576 | template <typename T> |
2577 | JSTokenType Lexer<T>::scanTemplateString(JSToken* tokenRecord, RawStringsBuildMode rawStringsBuildMode) |
2578 | { |
2579 | JSTokenData* tokenData = &tokenRecord->m_data; |
2580 | ASSERT(!m_error); |
2581 | ASSERT(m_buffer16.isEmpty()); |
2582 | |
2583 | // Leading backquote ` (for template head) or closing brace } (for template trailing) are already shifted in the previous token scan. |
2584 | // So in this re-scan phase, shift() is not needed here. |
2585 | StringParseResult result = parseTemplateLiteral(tokenData, rawStringsBuildMode); |
2586 | JSTokenType token = ERRORTOK; |
2587 | if (UNLIKELY(result != StringParsedSuccessfully)) { |
2588 | token = result == StringUnterminated ? UNTERMINATED_TEMPLATE_LITERAL_ERRORTOK : INVALID_TEMPLATE_LITERAL_ERRORTOK; |
2589 | m_error = true; |
2590 | } else |
2591 | token = TEMPLATE; |
2592 | |
2593 | // Since TemplateString always ends with ` or }, m_atLineStart always becomes false. |
2594 | m_atLineStart = false; |
2595 | fillTokenInfo(tokenRecord, token, m_lineNumber, currentOffset(), currentLineStartOffset(), currentPosition()); |
2596 | return token; |
2597 | } |
2598 | |
2599 | template <typename T> |
2600 | void Lexer<T>::clear() |
2601 | { |
2602 | m_arena = 0; |
2603 | |
2604 | Vector<LChar> newBuffer8; |
2605 | m_buffer8.swap(newBuffer8); |
2606 | |
2607 | Vector<UChar> newBuffer16; |
2608 | m_buffer16.swap(newBuffer16); |
2609 | |
2610 | Vector<UChar> newBufferForRawTemplateString16; |
2611 | m_bufferForRawTemplateString16.swap(newBufferForRawTemplateString16); |
2612 | |
2613 | m_isReparsingFunction = false; |
2614 | } |
2615 | |
2616 | // Instantiate the two flavors of Lexer we need instead of putting most of this file in Lexer.h |
2617 | template class Lexer<LChar>; |
2618 | template class Lexer<UChar>; |
2619 | |
2620 | } // namespace JSC |
2621 | |