]>
Commit | Line | Data |
---|---|---|
1 | /* | |
2 | * Copyright (C) 2009 Apple Inc. All rights reserved. | |
3 | * Copyright (C) 2012 Mathias Bynens (mathias@qiwi.be) | |
4 | * | |
5 | * Redistribution and use in source and binary forms, with or without | |
6 | * modification, are permitted provided that the following conditions | |
7 | * are met: | |
8 | * 1. Redistributions of source code must retain the above copyright | |
9 | * notice, this list of conditions and the following disclaimer. | |
10 | * 2. Redistributions in binary form must reproduce the above copyright | |
11 | * notice, this list of conditions and the following disclaimer in the | |
12 | * documentation and/or other materials provided with the distribution. | |
13 | * | |
14 | * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY | |
15 | * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
16 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR | |
17 | * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR | |
18 | * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, | |
19 | * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, | |
20 | * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR | |
21 | * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY | |
22 | * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | |
23 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
24 | * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
25 | */ | |
26 | ||
27 | #include "config.h" | |
28 | #include "LiteralParser.h" | |
29 | ||
30 | #include "ButterflyInlines.h" | |
31 | #include "CopiedSpaceInlines.h" | |
32 | #include "JSArray.h" | |
33 | #include "JSString.h" | |
34 | #include "Lexer.h" | |
35 | #include "ObjectConstructor.h" | |
36 | #include "Operations.h" | |
37 | #include "StrongInlines.h" | |
38 | #include <wtf/ASCIICType.h> | |
39 | #include <wtf/dtoa.h> | |
40 | #include <wtf/text/StringBuilder.h> | |
41 | ||
42 | namespace JSC { | |
43 | ||
44 | template <typename CharType> | |
45 | static inline bool isJSONWhiteSpace(const CharType& c) | |
46 | { | |
47 | // The JSON RFC 4627 defines a list of allowed characters to be considered | |
48 | // insignificant white space: http://www.ietf.org/rfc/rfc4627.txt (2. JSON Grammar). | |
49 | return c == ' ' || c == 0x9 || c == 0xA || c == 0xD; | |
50 | } | |
51 | ||
52 | template <typename CharType> | |
53 | bool LiteralParser<CharType>::tryJSONPParse(Vector<JSONPData>& results, bool needsFullSourceInfo) | |
54 | { | |
55 | if (m_lexer.next() != TokIdentifier) | |
56 | return false; | |
57 | do { | |
58 | Vector<JSONPPathEntry> path; | |
59 | // Unguarded next to start off the lexer | |
60 | Identifier name = Identifier(&m_exec->vm(), m_lexer.currentToken().start, m_lexer.currentToken().end - m_lexer.currentToken().start); | |
61 | JSONPPathEntry entry; | |
62 | if (name == m_exec->vm().propertyNames->varKeyword) { | |
63 | if (m_lexer.next() != TokIdentifier) | |
64 | return false; | |
65 | entry.m_type = JSONPPathEntryTypeDeclare; | |
66 | entry.m_pathEntryName = Identifier(&m_exec->vm(), m_lexer.currentToken().start, m_lexer.currentToken().end - m_lexer.currentToken().start); | |
67 | path.append(entry); | |
68 | } else { | |
69 | entry.m_type = JSONPPathEntryTypeDot; | |
70 | entry.m_pathEntryName = Identifier(&m_exec->vm(), m_lexer.currentToken().start, m_lexer.currentToken().end - m_lexer.currentToken().start); | |
71 | path.append(entry); | |
72 | } | |
73 | if (m_exec->vm().keywords->isKeyword(entry.m_pathEntryName)) | |
74 | return false; | |
75 | TokenType tokenType = m_lexer.next(); | |
76 | if (entry.m_type == JSONPPathEntryTypeDeclare && tokenType != TokAssign) | |
77 | return false; | |
78 | while (tokenType != TokAssign) { | |
79 | switch (tokenType) { | |
80 | case TokLBracket: { | |
81 | entry.m_type = JSONPPathEntryTypeLookup; | |
82 | if (m_lexer.next() != TokNumber) | |
83 | return false; | |
84 | double doubleIndex = m_lexer.currentToken().numberToken; | |
85 | int index = (int)doubleIndex; | |
86 | if (index != doubleIndex || index < 0) | |
87 | return false; | |
88 | entry.m_pathIndex = index; | |
89 | if (m_lexer.next() != TokRBracket) | |
90 | return false; | |
91 | break; | |
92 | } | |
93 | case TokDot: { | |
94 | entry.m_type = JSONPPathEntryTypeDot; | |
95 | if (m_lexer.next() != TokIdentifier) | |
96 | return false; | |
97 | entry.m_pathEntryName = Identifier(&m_exec->vm(), m_lexer.currentToken().start, m_lexer.currentToken().end - m_lexer.currentToken().start); | |
98 | break; | |
99 | } | |
100 | case TokLParen: { | |
101 | if (path.last().m_type != JSONPPathEntryTypeDot || needsFullSourceInfo) | |
102 | return false; | |
103 | path.last().m_type = JSONPPathEntryTypeCall; | |
104 | entry = path.last(); | |
105 | goto startJSON; | |
106 | } | |
107 | default: | |
108 | return false; | |
109 | } | |
110 | path.append(entry); | |
111 | tokenType = m_lexer.next(); | |
112 | } | |
113 | startJSON: | |
114 | m_lexer.next(); | |
115 | results.append(JSONPData()); | |
116 | results.last().m_value.set(m_exec->vm(), parse(StartParseExpression)); | |
117 | if (!results.last().m_value) | |
118 | return false; | |
119 | results.last().m_path.swap(path); | |
120 | if (entry.m_type == JSONPPathEntryTypeCall) { | |
121 | if (m_lexer.currentToken().type != TokRParen) | |
122 | return false; | |
123 | m_lexer.next(); | |
124 | } | |
125 | if (m_lexer.currentToken().type != TokSemi) | |
126 | break; | |
127 | m_lexer.next(); | |
128 | } while (m_lexer.currentToken().type == TokIdentifier); | |
129 | return m_lexer.currentToken().type == TokEnd; | |
130 | } | |
131 | ||
132 | template <typename CharType> | |
133 | ALWAYS_INLINE const Identifier LiteralParser<CharType>::makeIdentifier(const LChar* characters, size_t length) | |
134 | { | |
135 | if (!length) | |
136 | return m_exec->vm().propertyNames->emptyIdentifier; | |
137 | if (characters[0] >= MaximumCachableCharacter) | |
138 | return Identifier(&m_exec->vm(), characters, length); | |
139 | ||
140 | if (length == 1) { | |
141 | if (!m_shortIdentifiers[characters[0]].isNull()) | |
142 | return m_shortIdentifiers[characters[0]]; | |
143 | m_shortIdentifiers[characters[0]] = Identifier(&m_exec->vm(), characters, length); | |
144 | return m_shortIdentifiers[characters[0]]; | |
145 | } | |
146 | if (!m_recentIdentifiers[characters[0]].isNull() && Identifier::equal(m_recentIdentifiers[characters[0]].impl(), characters, length)) | |
147 | return m_recentIdentifiers[characters[0]]; | |
148 | m_recentIdentifiers[characters[0]] = Identifier(&m_exec->vm(), characters, length); | |
149 | return m_recentIdentifiers[characters[0]]; | |
150 | } | |
151 | ||
152 | template <typename CharType> | |
153 | ALWAYS_INLINE const Identifier LiteralParser<CharType>::makeIdentifier(const UChar* characters, size_t length) | |
154 | { | |
155 | if (!length) | |
156 | return m_exec->vm().propertyNames->emptyIdentifier; | |
157 | if (characters[0] >= MaximumCachableCharacter) | |
158 | return Identifier(&m_exec->vm(), characters, length); | |
159 | ||
160 | if (length == 1) { | |
161 | if (!m_shortIdentifiers[characters[0]].isNull()) | |
162 | return m_shortIdentifiers[characters[0]]; | |
163 | m_shortIdentifiers[characters[0]] = Identifier(&m_exec->vm(), characters, length); | |
164 | return m_shortIdentifiers[characters[0]]; | |
165 | } | |
166 | if (!m_recentIdentifiers[characters[0]].isNull() && Identifier::equal(m_recentIdentifiers[characters[0]].impl(), characters, length)) | |
167 | return m_recentIdentifiers[characters[0]]; | |
168 | m_recentIdentifiers[characters[0]] = Identifier(&m_exec->vm(), characters, length); | |
169 | return m_recentIdentifiers[characters[0]]; | |
170 | } | |
171 | ||
172 | template <typename CharType> | |
173 | template <ParserMode mode> TokenType LiteralParser<CharType>::Lexer::lex(LiteralParserToken<CharType>& token) | |
174 | { | |
175 | while (m_ptr < m_end && isJSONWhiteSpace(*m_ptr)) | |
176 | ++m_ptr; | |
177 | ||
178 | ASSERT(m_ptr <= m_end); | |
179 | if (m_ptr >= m_end) { | |
180 | token.type = TokEnd; | |
181 | token.start = token.end = m_ptr; | |
182 | return TokEnd; | |
183 | } | |
184 | token.type = TokError; | |
185 | token.start = m_ptr; | |
186 | switch (*m_ptr) { | |
187 | case '[': | |
188 | token.type = TokLBracket; | |
189 | token.end = ++m_ptr; | |
190 | return TokLBracket; | |
191 | case ']': | |
192 | token.type = TokRBracket; | |
193 | token.end = ++m_ptr; | |
194 | return TokRBracket; | |
195 | case '(': | |
196 | token.type = TokLParen; | |
197 | token.end = ++m_ptr; | |
198 | return TokLParen; | |
199 | case ')': | |
200 | token.type = TokRParen; | |
201 | token.end = ++m_ptr; | |
202 | return TokRParen; | |
203 | case '{': | |
204 | token.type = TokLBrace; | |
205 | token.end = ++m_ptr; | |
206 | return TokLBrace; | |
207 | case '}': | |
208 | token.type = TokRBrace; | |
209 | token.end = ++m_ptr; | |
210 | return TokRBrace; | |
211 | case ',': | |
212 | token.type = TokComma; | |
213 | token.end = ++m_ptr; | |
214 | return TokComma; | |
215 | case ':': | |
216 | token.type = TokColon; | |
217 | token.end = ++m_ptr; | |
218 | return TokColon; | |
219 | case '"': | |
220 | return lexString<mode, '"'>(token); | |
221 | case 't': | |
222 | if (m_end - m_ptr >= 4 && m_ptr[1] == 'r' && m_ptr[2] == 'u' && m_ptr[3] == 'e') { | |
223 | m_ptr += 4; | |
224 | token.type = TokTrue; | |
225 | token.end = m_ptr; | |
226 | return TokTrue; | |
227 | } | |
228 | break; | |
229 | case 'f': | |
230 | if (m_end - m_ptr >= 5 && m_ptr[1] == 'a' && m_ptr[2] == 'l' && m_ptr[3] == 's' && m_ptr[4] == 'e') { | |
231 | m_ptr += 5; | |
232 | token.type = TokFalse; | |
233 | token.end = m_ptr; | |
234 | return TokFalse; | |
235 | } | |
236 | break; | |
237 | case 'n': | |
238 | if (m_end - m_ptr >= 4 && m_ptr[1] == 'u' && m_ptr[2] == 'l' && m_ptr[3] == 'l') { | |
239 | m_ptr += 4; | |
240 | token.type = TokNull; | |
241 | token.end = m_ptr; | |
242 | return TokNull; | |
243 | } | |
244 | break; | |
245 | case '-': | |
246 | case '0': | |
247 | case '1': | |
248 | case '2': | |
249 | case '3': | |
250 | case '4': | |
251 | case '5': | |
252 | case '6': | |
253 | case '7': | |
254 | case '8': | |
255 | case '9': | |
256 | return lexNumber(token); | |
257 | } | |
258 | if (m_ptr < m_end) { | |
259 | if (*m_ptr == '.') { | |
260 | token.type = TokDot; | |
261 | token.end = ++m_ptr; | |
262 | return TokDot; | |
263 | } | |
264 | if (*m_ptr == '=') { | |
265 | token.type = TokAssign; | |
266 | token.end = ++m_ptr; | |
267 | return TokAssign; | |
268 | } | |
269 | if (*m_ptr == ';') { | |
270 | token.type = TokSemi; | |
271 | token.end = ++m_ptr; | |
272 | return TokAssign; | |
273 | } | |
274 | if (isASCIIAlpha(*m_ptr) || *m_ptr == '_' || *m_ptr == '$') | |
275 | return lexIdentifier(token); | |
276 | if (*m_ptr == '\'') { | |
277 | if (mode == StrictJSON) { | |
278 | m_lexErrorMessage = ASCIILiteral("Single quotes (\') are not allowed in JSON"); | |
279 | return TokError; | |
280 | } | |
281 | return lexString<mode, '\''>(token); | |
282 | } | |
283 | } | |
284 | m_lexErrorMessage = String::format("Unrecognized token '%c'", *m_ptr).impl(); | |
285 | return TokError; | |
286 | } | |
287 | ||
288 | template <> | |
289 | ALWAYS_INLINE TokenType LiteralParser<LChar>::Lexer::lexIdentifier(LiteralParserToken<LChar>& token) | |
290 | { | |
291 | while (m_ptr < m_end && (isASCIIAlphanumeric(*m_ptr) || *m_ptr == '_' || *m_ptr == '$')) | |
292 | m_ptr++; | |
293 | token.stringIs8Bit = 1; | |
294 | token.stringToken8 = token.start; | |
295 | token.stringLength = m_ptr - token.start; | |
296 | token.type = TokIdentifier; | |
297 | token.end = m_ptr; | |
298 | return TokIdentifier; | |
299 | } | |
300 | ||
301 | template <> | |
302 | ALWAYS_INLINE TokenType LiteralParser<UChar>::Lexer::lexIdentifier(LiteralParserToken<UChar>& token) | |
303 | { | |
304 | while (m_ptr < m_end && (isASCIIAlphanumeric(*m_ptr) || *m_ptr == '_' || *m_ptr == '$' || *m_ptr == 0x200C || *m_ptr == 0x200D)) | |
305 | m_ptr++; | |
306 | token.stringIs8Bit = 0; | |
307 | token.stringToken16 = token.start; | |
308 | token.stringLength = m_ptr - token.start; | |
309 | token.type = TokIdentifier; | |
310 | token.end = m_ptr; | |
311 | return TokIdentifier; | |
312 | } | |
313 | ||
314 | template <typename CharType> | |
315 | TokenType LiteralParser<CharType>::Lexer::next() | |
316 | { | |
317 | if (m_mode == NonStrictJSON) | |
318 | return lex<NonStrictJSON>(m_currentToken); | |
319 | if (m_mode == JSONP) | |
320 | return lex<JSONP>(m_currentToken); | |
321 | return lex<StrictJSON>(m_currentToken); | |
322 | } | |
323 | ||
324 | template <> | |
325 | ALWAYS_INLINE void setParserTokenString<LChar>(LiteralParserToken<LChar>& token, const LChar* string) | |
326 | { | |
327 | token.stringIs8Bit = 1; | |
328 | token.stringToken8 = string; | |
329 | } | |
330 | ||
331 | template <> | |
332 | ALWAYS_INLINE void setParserTokenString<UChar>(LiteralParserToken<UChar>& token, const UChar* string) | |
333 | { | |
334 | token.stringIs8Bit = 0; | |
335 | token.stringToken16 = string; | |
336 | } | |
337 | ||
338 | template <ParserMode mode, typename CharType, LChar terminator> static inline bool isSafeStringCharacter(LChar c) | |
339 | { | |
340 | return (c >= ' ' && c != '\\' && c != terminator) || (c == '\t' && mode != StrictJSON); | |
341 | } | |
342 | ||
343 | template <ParserMode mode, typename CharType, UChar terminator> static inline bool isSafeStringCharacter(UChar c) | |
344 | { | |
345 | return (c >= ' ' && (mode == StrictJSON || c <= 0xff) && c != '\\' && c != terminator) || (c == '\t' && mode != StrictJSON); | |
346 | } | |
347 | ||
348 | template <typename CharType> | |
349 | template <ParserMode mode, char terminator> ALWAYS_INLINE TokenType LiteralParser<CharType>::Lexer::lexString(LiteralParserToken<CharType>& token) | |
350 | { | |
351 | ++m_ptr; | |
352 | const CharType* runStart = m_ptr; | |
353 | StringBuilder builder; | |
354 | do { | |
355 | runStart = m_ptr; | |
356 | while (m_ptr < m_end && isSafeStringCharacter<mode, CharType, terminator>(*m_ptr)) | |
357 | ++m_ptr; | |
358 | if (builder.length()) | |
359 | builder.append(runStart, m_ptr - runStart); | |
360 | if ((mode != NonStrictJSON) && m_ptr < m_end && *m_ptr == '\\') { | |
361 | if (builder.isEmpty() && runStart < m_ptr) | |
362 | builder.append(runStart, m_ptr - runStart); | |
363 | ++m_ptr; | |
364 | if (m_ptr >= m_end) { | |
365 | m_lexErrorMessage = ASCIILiteral("Unterminated string"); | |
366 | return TokError; | |
367 | } | |
368 | switch (*m_ptr) { | |
369 | case '"': | |
370 | builder.append('"'); | |
371 | m_ptr++; | |
372 | break; | |
373 | case '\\': | |
374 | builder.append('\\'); | |
375 | m_ptr++; | |
376 | break; | |
377 | case '/': | |
378 | builder.append('/'); | |
379 | m_ptr++; | |
380 | break; | |
381 | case 'b': | |
382 | builder.append('\b'); | |
383 | m_ptr++; | |
384 | break; | |
385 | case 'f': | |
386 | builder.append('\f'); | |
387 | m_ptr++; | |
388 | break; | |
389 | case 'n': | |
390 | builder.append('\n'); | |
391 | m_ptr++; | |
392 | break; | |
393 | case 'r': | |
394 | builder.append('\r'); | |
395 | m_ptr++; | |
396 | break; | |
397 | case 't': | |
398 | builder.append('\t'); | |
399 | m_ptr++; | |
400 | break; | |
401 | ||
402 | case 'u': | |
403 | if ((m_end - m_ptr) < 5) { | |
404 | m_lexErrorMessage = ASCIILiteral("\\u must be followed by 4 hex digits"); | |
405 | return TokError; | |
406 | } // uNNNN == 5 characters | |
407 | for (int i = 1; i < 5; i++) { | |
408 | if (!isASCIIHexDigit(m_ptr[i])) { | |
409 | m_lexErrorMessage = String::format("\"\\%s\" is not a valid unicode escape", String(m_ptr, 5).ascii().data()).impl(); | |
410 | return TokError; | |
411 | } | |
412 | } | |
413 | builder.append(JSC::Lexer<CharType>::convertUnicode(m_ptr[1], m_ptr[2], m_ptr[3], m_ptr[4])); | |
414 | m_ptr += 5; | |
415 | break; | |
416 | ||
417 | default: | |
418 | if (*m_ptr == '\'' && mode != StrictJSON) { | |
419 | builder.append('\''); | |
420 | m_ptr++; | |
421 | break; | |
422 | } | |
423 | m_lexErrorMessage = String::format("Invalid escape character %c", *m_ptr).impl(); | |
424 | return TokError; | |
425 | } | |
426 | } | |
427 | } while ((mode != NonStrictJSON) && m_ptr != runStart && (m_ptr < m_end) && *m_ptr != terminator); | |
428 | ||
429 | if (m_ptr >= m_end || *m_ptr != terminator) { | |
430 | m_lexErrorMessage = ASCIILiteral("Unterminated string"); | |
431 | return TokError; | |
432 | } | |
433 | ||
434 | if (builder.isEmpty()) { | |
435 | token.stringBuffer = String(); | |
436 | setParserTokenString<CharType>(token, runStart); | |
437 | token.stringLength = m_ptr - runStart; | |
438 | } else { | |
439 | token.stringBuffer = builder.toString(); | |
440 | if (token.stringBuffer.is8Bit()) { | |
441 | token.stringIs8Bit = 1; | |
442 | token.stringToken8 = token.stringBuffer.characters8(); | |
443 | } else { | |
444 | token.stringIs8Bit = 0; | |
445 | token.stringToken16 = token.stringBuffer.characters16(); | |
446 | } | |
447 | token.stringLength = token.stringBuffer.length(); | |
448 | } | |
449 | token.type = TokString; | |
450 | token.end = ++m_ptr; | |
451 | return TokString; | |
452 | } | |
453 | ||
454 | template <typename CharType> | |
455 | TokenType LiteralParser<CharType>::Lexer::lexNumber(LiteralParserToken<CharType>& token) | |
456 | { | |
457 | // ES5 and json.org define numbers as | |
458 | // number | |
459 | // int | |
460 | // int frac? exp? | |
461 | // | |
462 | // int | |
463 | // -? 0 | |
464 | // -? digit1-9 digits? | |
465 | // | |
466 | // digits | |
467 | // digit digits? | |
468 | // | |
469 | // -?(0 | [1-9][0-9]*) ('.' [0-9]+)? ([eE][+-]? [0-9]+)? | |
470 | ||
471 | if (m_ptr < m_end && *m_ptr == '-') // -? | |
472 | ++m_ptr; | |
473 | ||
474 | // (0 | [1-9][0-9]*) | |
475 | if (m_ptr < m_end && *m_ptr == '0') // 0 | |
476 | ++m_ptr; | |
477 | else if (m_ptr < m_end && *m_ptr >= '1' && *m_ptr <= '9') { // [1-9] | |
478 | ++m_ptr; | |
479 | // [0-9]* | |
480 | while (m_ptr < m_end && isASCIIDigit(*m_ptr)) | |
481 | ++m_ptr; | |
482 | } else { | |
483 | m_lexErrorMessage = ASCIILiteral("Invalid number"); | |
484 | return TokError; | |
485 | } | |
486 | ||
487 | // ('.' [0-9]+)? | |
488 | if (m_ptr < m_end && *m_ptr == '.') { | |
489 | ++m_ptr; | |
490 | // [0-9]+ | |
491 | if (m_ptr >= m_end || !isASCIIDigit(*m_ptr)) { | |
492 | m_lexErrorMessage = ASCIILiteral("Invalid digits after decimal point"); | |
493 | return TokError; | |
494 | } | |
495 | ||
496 | ++m_ptr; | |
497 | while (m_ptr < m_end && isASCIIDigit(*m_ptr)) | |
498 | ++m_ptr; | |
499 | } else if (m_ptr < m_end && (*m_ptr != 'e' && *m_ptr != 'E') && (m_ptr - token.start) < 10) { | |
500 | int result = 0; | |
501 | token.type = TokNumber; | |
502 | token.end = m_ptr; | |
503 | const CharType* digit = token.start; | |
504 | int negative = 1; | |
505 | if (*digit == '-') { | |
506 | negative = -1; | |
507 | digit++; | |
508 | } | |
509 | ||
510 | while (digit < m_ptr) | |
511 | result = result * 10 + (*digit++) - '0'; | |
512 | result *= negative; | |
513 | token.numberToken = result; | |
514 | return TokNumber; | |
515 | } | |
516 | ||
517 | // ([eE][+-]? [0-9]+)? | |
518 | if (m_ptr < m_end && (*m_ptr == 'e' || *m_ptr == 'E')) { // [eE] | |
519 | ++m_ptr; | |
520 | ||
521 | // [-+]? | |
522 | if (m_ptr < m_end && (*m_ptr == '-' || *m_ptr == '+')) | |
523 | ++m_ptr; | |
524 | ||
525 | // [0-9]+ | |
526 | if (m_ptr >= m_end || !isASCIIDigit(*m_ptr)) { | |
527 | m_lexErrorMessage = ASCIILiteral("Exponent symbols should be followed by an optional '+' or '-' and then by at least one number"); | |
528 | return TokError; | |
529 | } | |
530 | ||
531 | ++m_ptr; | |
532 | while (m_ptr < m_end && isASCIIDigit(*m_ptr)) | |
533 | ++m_ptr; | |
534 | } | |
535 | ||
536 | token.type = TokNumber; | |
537 | token.end = m_ptr; | |
538 | size_t parsedLength; | |
539 | token.numberToken = parseDouble(token.start, token.end - token.start, parsedLength); | |
540 | return TokNumber; | |
541 | } | |
542 | ||
543 | template <typename CharType> | |
544 | JSValue LiteralParser<CharType>::parse(ParserState initialState) | |
545 | { | |
546 | ParserState state = initialState; | |
547 | MarkedArgumentBuffer objectStack; | |
548 | JSValue lastValue; | |
549 | Vector<ParserState, 16, UnsafeVectorOverflow> stateStack; | |
550 | Vector<Identifier, 16, UnsafeVectorOverflow> identifierStack; | |
551 | while (1) { | |
552 | switch(state) { | |
553 | startParseArray: | |
554 | case StartParseArray: { | |
555 | JSArray* array = constructEmptyArray(m_exec, 0); | |
556 | objectStack.append(array); | |
557 | // fallthrough | |
558 | } | |
559 | doParseArrayStartExpression: | |
560 | case DoParseArrayStartExpression: { | |
561 | TokenType lastToken = m_lexer.currentToken().type; | |
562 | if (m_lexer.next() == TokRBracket) { | |
563 | if (lastToken == TokComma) { | |
564 | m_parseErrorMessage = ASCIILiteral("Unexpected comma at the end of array expression"); | |
565 | return JSValue(); | |
566 | } | |
567 | m_lexer.next(); | |
568 | lastValue = objectStack.last(); | |
569 | objectStack.removeLast(); | |
570 | break; | |
571 | } | |
572 | ||
573 | stateStack.append(DoParseArrayEndExpression); | |
574 | goto startParseExpression; | |
575 | } | |
576 | case DoParseArrayEndExpression: { | |
577 | JSArray* array = asArray(objectStack.last()); | |
578 | array->putDirectIndex(m_exec, array->length(), lastValue); | |
579 | ||
580 | if (m_lexer.currentToken().type == TokComma) | |
581 | goto doParseArrayStartExpression; | |
582 | ||
583 | if (m_lexer.currentToken().type != TokRBracket) { | |
584 | m_parseErrorMessage = ASCIILiteral("Expected ']'"); | |
585 | return JSValue(); | |
586 | } | |
587 | ||
588 | m_lexer.next(); | |
589 | lastValue = objectStack.last(); | |
590 | objectStack.removeLast(); | |
591 | break; | |
592 | } | |
593 | startParseObject: | |
594 | case StartParseObject: { | |
595 | JSObject* object = constructEmptyObject(m_exec); | |
596 | objectStack.append(object); | |
597 | ||
598 | TokenType type = m_lexer.next(); | |
599 | if (type == TokString || (m_mode != StrictJSON && type == TokIdentifier)) { | |
600 | LiteralParserToken<CharType> identifierToken = m_lexer.currentToken(); | |
601 | ||
602 | // Check for colon | |
603 | if (m_lexer.next() != TokColon) { | |
604 | m_parseErrorMessage = ASCIILiteral("Expected ':' before value in object property definition"); | |
605 | return JSValue(); | |
606 | } | |
607 | ||
608 | m_lexer.next(); | |
609 | if (identifierToken.stringIs8Bit) | |
610 | identifierStack.append(makeIdentifier(identifierToken.stringToken8, identifierToken.stringLength)); | |
611 | else | |
612 | identifierStack.append(makeIdentifier(identifierToken.stringToken16, identifierToken.stringLength)); | |
613 | stateStack.append(DoParseObjectEndExpression); | |
614 | goto startParseExpression; | |
615 | } | |
616 | if (type != TokRBrace) { | |
617 | m_parseErrorMessage = ASCIILiteral("Expected '}'"); | |
618 | return JSValue(); | |
619 | } | |
620 | m_lexer.next(); | |
621 | lastValue = objectStack.last(); | |
622 | objectStack.removeLast(); | |
623 | break; | |
624 | } | |
625 | doParseObjectStartExpression: | |
626 | case DoParseObjectStartExpression: { | |
627 | TokenType type = m_lexer.next(); | |
628 | if (type != TokString && (m_mode == StrictJSON || type != TokIdentifier)) { | |
629 | m_parseErrorMessage = ASCIILiteral("Property name must be a string literal"); | |
630 | return JSValue(); | |
631 | } | |
632 | LiteralParserToken<CharType> identifierToken = m_lexer.currentToken(); | |
633 | ||
634 | // Check for colon | |
635 | if (m_lexer.next() != TokColon) { | |
636 | m_parseErrorMessage = ASCIILiteral("Expected ':'"); | |
637 | return JSValue(); | |
638 | } | |
639 | ||
640 | m_lexer.next(); | |
641 | if (identifierToken.stringIs8Bit) | |
642 | identifierStack.append(makeIdentifier(identifierToken.stringToken8, identifierToken.stringLength)); | |
643 | else | |
644 | identifierStack.append(makeIdentifier(identifierToken.stringToken16, identifierToken.stringLength)); | |
645 | stateStack.append(DoParseObjectEndExpression); | |
646 | goto startParseExpression; | |
647 | } | |
648 | case DoParseObjectEndExpression: | |
649 | { | |
650 | JSObject* object = asObject(objectStack.last()); | |
651 | PropertyName ident = identifierStack.last(); | |
652 | unsigned i = ident.asIndex(); | |
653 | if (i != PropertyName::NotAnIndex) | |
654 | object->putDirectIndex(m_exec, i, lastValue); | |
655 | else | |
656 | object->putDirect(m_exec->vm(), ident, lastValue); | |
657 | identifierStack.removeLast(); | |
658 | if (m_lexer.currentToken().type == TokComma) | |
659 | goto doParseObjectStartExpression; | |
660 | if (m_lexer.currentToken().type != TokRBrace) { | |
661 | m_parseErrorMessage = ASCIILiteral("Expected '}'"); | |
662 | return JSValue(); | |
663 | } | |
664 | m_lexer.next(); | |
665 | lastValue = objectStack.last(); | |
666 | objectStack.removeLast(); | |
667 | break; | |
668 | } | |
669 | startParseExpression: | |
670 | case StartParseExpression: { | |
671 | switch (m_lexer.currentToken().type) { | |
672 | case TokLBracket: | |
673 | goto startParseArray; | |
674 | case TokLBrace: | |
675 | goto startParseObject; | |
676 | case TokString: { | |
677 | LiteralParserToken<CharType> stringToken = m_lexer.currentToken(); | |
678 | m_lexer.next(); | |
679 | if (stringToken.stringIs8Bit) | |
680 | lastValue = jsString(m_exec, makeIdentifier(stringToken.stringToken8, stringToken.stringLength).string()); | |
681 | else | |
682 | lastValue = jsString(m_exec, makeIdentifier(stringToken.stringToken16, stringToken.stringLength).string()); | |
683 | break; | |
684 | } | |
685 | case TokNumber: { | |
686 | LiteralParserToken<CharType> numberToken = m_lexer.currentToken(); | |
687 | m_lexer.next(); | |
688 | lastValue = jsNumber(numberToken.numberToken); | |
689 | break; | |
690 | } | |
691 | case TokNull: | |
692 | m_lexer.next(); | |
693 | lastValue = jsNull(); | |
694 | break; | |
695 | ||
696 | case TokTrue: | |
697 | m_lexer.next(); | |
698 | lastValue = jsBoolean(true); | |
699 | break; | |
700 | ||
701 | case TokFalse: | |
702 | m_lexer.next(); | |
703 | lastValue = jsBoolean(false); | |
704 | break; | |
705 | case TokRBracket: | |
706 | m_parseErrorMessage = ASCIILiteral("Unexpected token ']'"); | |
707 | return JSValue(); | |
708 | case TokRBrace: | |
709 | m_parseErrorMessage = ASCIILiteral("Unexpected token '}'"); | |
710 | return JSValue(); | |
711 | case TokIdentifier: { | |
712 | const LiteralParserToken<CharType>& token = m_lexer.currentToken(); | |
713 | if (token.stringIs8Bit) | |
714 | m_parseErrorMessage = String::format("Unexpected identifier \"%s\"", String(m_lexer.currentToken().stringToken8, m_lexer.currentToken().stringLength).ascii().data()).impl(); | |
715 | else | |
716 | m_parseErrorMessage = String::format("Unexpected identifier \"%s\"", String(m_lexer.currentToken().stringToken16, m_lexer.currentToken().stringLength).ascii().data()).impl(); | |
717 | return JSValue(); | |
718 | } | |
719 | case TokColon: | |
720 | m_parseErrorMessage = ASCIILiteral("Unexpected token ':'"); | |
721 | return JSValue(); | |
722 | case TokLParen: | |
723 | m_parseErrorMessage = ASCIILiteral("Unexpected token '('"); | |
724 | return JSValue(); | |
725 | case TokRParen: | |
726 | m_parseErrorMessage = ASCIILiteral("Unexpected token ')'"); | |
727 | return JSValue(); | |
728 | case TokComma: | |
729 | m_parseErrorMessage = ASCIILiteral("Unexpected token ','"); | |
730 | return JSValue(); | |
731 | case TokDot: | |
732 | m_parseErrorMessage = ASCIILiteral("Unexpected token '.'"); | |
733 | return JSValue(); | |
734 | case TokAssign: | |
735 | m_parseErrorMessage = ASCIILiteral("Unexpected token '='"); | |
736 | return JSValue(); | |
737 | case TokSemi: | |
738 | m_parseErrorMessage = ASCIILiteral("Unexpected token ';'"); | |
739 | return JSValue(); | |
740 | case TokEnd: | |
741 | m_parseErrorMessage = ASCIILiteral("Unexpected EOF"); | |
742 | return JSValue(); | |
743 | case TokError: | |
744 | default: | |
745 | // Error | |
746 | m_parseErrorMessage = ASCIILiteral("Could not parse value expression"); | |
747 | return JSValue(); | |
748 | } | |
749 | break; | |
750 | } | |
751 | case StartParseStatement: { | |
752 | switch (m_lexer.currentToken().type) { | |
753 | case TokLBracket: | |
754 | case TokNumber: | |
755 | case TokString: | |
756 | goto startParseExpression; | |
757 | ||
758 | case TokLParen: { | |
759 | m_lexer.next(); | |
760 | stateStack.append(StartParseStatementEndStatement); | |
761 | goto startParseExpression; | |
762 | } | |
763 | case TokRBracket: | |
764 | m_parseErrorMessage = ASCIILiteral("Unexpected token ']'"); | |
765 | return JSValue(); | |
766 | case TokLBrace: | |
767 | m_parseErrorMessage = ASCIILiteral("Unexpected token '{'"); | |
768 | return JSValue(); | |
769 | case TokRBrace: | |
770 | m_parseErrorMessage = ASCIILiteral("Unexpected token '}'"); | |
771 | return JSValue(); | |
772 | case TokIdentifier: | |
773 | m_parseErrorMessage = ASCIILiteral("Unexpected identifier"); | |
774 | return JSValue(); | |
775 | case TokColon: | |
776 | m_parseErrorMessage = ASCIILiteral("Unexpected token ':'"); | |
777 | return JSValue(); | |
778 | case TokRParen: | |
779 | m_parseErrorMessage = ASCIILiteral("Unexpected token ')'"); | |
780 | return JSValue(); | |
781 | case TokComma: | |
782 | m_parseErrorMessage = ASCIILiteral("Unexpected token ','"); | |
783 | return JSValue(); | |
784 | case TokTrue: | |
785 | m_parseErrorMessage = ASCIILiteral("Unexpected token 'true'"); | |
786 | return JSValue(); | |
787 | case TokFalse: | |
788 | m_parseErrorMessage = ASCIILiteral("Unexpected token 'false'"); | |
789 | return JSValue(); | |
790 | case TokNull: | |
791 | m_parseErrorMessage = ASCIILiteral("Unexpected token 'null'"); | |
792 | return JSValue(); | |
793 | case TokEnd: | |
794 | m_parseErrorMessage = ASCIILiteral("Unexpected EOF"); | |
795 | return JSValue(); | |
796 | case TokDot: | |
797 | m_parseErrorMessage = ASCIILiteral("Unexpected token '.'"); | |
798 | return JSValue(); | |
799 | case TokAssign: | |
800 | m_parseErrorMessage = ASCIILiteral("Unexpected token '='"); | |
801 | return JSValue(); | |
802 | case TokSemi: | |
803 | m_parseErrorMessage = ASCIILiteral("Unexpected token ';'"); | |
804 | return JSValue(); | |
805 | case TokError: | |
806 | default: | |
807 | m_parseErrorMessage = ASCIILiteral("Could not parse statement"); | |
808 | return JSValue(); | |
809 | } | |
810 | } | |
811 | case StartParseStatementEndStatement: { | |
812 | ASSERT(stateStack.isEmpty()); | |
813 | if (m_lexer.currentToken().type != TokRParen) | |
814 | return JSValue(); | |
815 | if (m_lexer.next() == TokEnd) | |
816 | return lastValue; | |
817 | m_parseErrorMessage = ASCIILiteral("Unexpected content at end of JSON literal"); | |
818 | return JSValue(); | |
819 | } | |
820 | default: | |
821 | RELEASE_ASSERT_NOT_REACHED(); | |
822 | } | |
823 | if (stateStack.isEmpty()) | |
824 | return lastValue; | |
825 | state = stateStack.last(); | |
826 | stateStack.removeLast(); | |
827 | continue; | |
828 | } | |
829 | } | |
830 | ||
831 | // Instantiate the two flavors of LiteralParser we need instead of putting most of this file in LiteralParser.h | |
832 | template class LiteralParser<LChar>; | |
833 | template class LiteralParser<UChar>; | |
834 | ||
835 | } |