blob: dfd909a00490769e7c8e7161cd9157ee1c0a9f6b [file] [log] [blame]
Ben Murdochf87a2032010-10-22 12:50:53 +01001// Copyright 2010 the V8 project authors. All rights reserved.
Steve Blocka7e24c12009-10-30 11:49:00 +00002// Redistribution and use in source and binary forms, with or without
3// modification, are permitted provided that the following conditions are
4// met:
5//
6// * Redistributions of source code must retain the above copyright
7// notice, this list of conditions and the following disclaimer.
8// * Redistributions in binary form must reproduce the above
9// copyright notice, this list of conditions and the following
10// disclaimer in the documentation and/or other materials provided
11// with the distribution.
12// * Neither the name of Google Inc. nor the names of its
13// contributors may be used to endorse or promote products derived
14// from this software without specific prior written permission.
15//
16// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
17// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
18// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
19// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
20// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
21// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
22// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
26// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27
28#ifndef V8_PARSER_H_
29#define V8_PARSER_H_
30
Steve Blocka7e24c12009-10-30 11:49:00 +000031#include "allocation.h"
Ben Murdochf87a2032010-10-22 12:50:53 +010032#include "ast.h"
33#include "scanner.h"
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -080034#include "scopes.h"
Shimeng (Simon) Wang8a31eba2010-12-06 19:01:33 -080035#include "preparse-data.h"
Steve Blocka7e24c12009-10-30 11:49:00 +000036
37namespace v8 {
38namespace internal {
39
Ben Murdochf87a2032010-10-22 12:50:53 +010040class CompilationInfo;
41class FuncNameInferrer;
Ben Murdochf87a2032010-10-22 12:50:53 +010042class ParserLog;
43class PositionStack;
44class Target;
45class TemporaryScope;
46
47template <typename T> class ZoneListWrapper;
48
Steve Blocka7e24c12009-10-30 11:49:00 +000049
50class ParserMessage : public Malloced {
51 public:
52 ParserMessage(Scanner::Location loc, const char* message,
53 Vector<const char*> args)
54 : loc_(loc),
55 message_(message),
56 args_(args) { }
57 ~ParserMessage();
58 Scanner::Location location() { return loc_; }
59 const char* message() { return message_; }
60 Vector<const char*> args() { return args_; }
61 private:
62 Scanner::Location loc_;
63 const char* message_;
64 Vector<const char*> args_;
65};
66
67
68class FunctionEntry BASE_EMBEDDED {
69 public:
70 explicit FunctionEntry(Vector<unsigned> backing) : backing_(backing) { }
71 FunctionEntry() : backing_(Vector<unsigned>::empty()) { }
72
73 int start_pos() { return backing_[kStartPosOffset]; }
74 void set_start_pos(int value) { backing_[kStartPosOffset] = value; }
75
76 int end_pos() { return backing_[kEndPosOffset]; }
77 void set_end_pos(int value) { backing_[kEndPosOffset] = value; }
78
79 int literal_count() { return backing_[kLiteralCountOffset]; }
80 void set_literal_count(int value) { backing_[kLiteralCountOffset] = value; }
81
82 int property_count() { return backing_[kPropertyCountOffset]; }
Kristian Monsen80d68ea2010-09-08 11:05:35 +010083 void set_property_count(int value) {
84 backing_[kPropertyCountOffset] = value;
85 }
86
Steve Blocka7e24c12009-10-30 11:49:00 +000087 bool is_valid() { return backing_.length() > 0; }
88
Ben Murdochf87a2032010-10-22 12:50:53 +010089 static const int kSize = 4;
Steve Blocka7e24c12009-10-30 11:49:00 +000090
91 private:
92 Vector<unsigned> backing_;
93 static const int kStartPosOffset = 0;
94 static const int kEndPosOffset = 1;
95 static const int kLiteralCountOffset = 2;
96 static const int kPropertyCountOffset = 3;
Steve Blocka7e24c12009-10-30 11:49:00 +000097};
98
99
100class ScriptDataImpl : public ScriptData {
101 public:
102 explicit ScriptDataImpl(Vector<unsigned> store)
103 : store_(store),
Kristian Monsen0d5e1162010-09-30 15:31:59 +0100104 owns_store_(true) { }
Iain Merrick9ac36c92010-09-13 15:29:50 +0100105
106 // Create an empty ScriptDataImpl that is guaranteed to not satisfy
107 // a SanityCheck.
108 ScriptDataImpl() : store_(Vector<unsigned>()), owns_store_(false) { }
109
Steve Blocka7e24c12009-10-30 11:49:00 +0000110 virtual ~ScriptDataImpl();
111 virtual int Length();
Leon Clarkef7060e22010-06-03 12:02:55 +0100112 virtual const char* Data();
Leon Clarkee46be812010-01-19 14:06:41 +0000113 virtual bool HasError();
Iain Merrick9ac36c92010-09-13 15:29:50 +0100114
Kristian Monsen0d5e1162010-09-30 15:31:59 +0100115 void Initialize();
116 void ReadNextSymbolPosition();
117
Kristian Monsen80d68ea2010-09-08 11:05:35 +0100118 FunctionEntry GetFunctionEntry(int start);
Kristian Monsen0d5e1162010-09-30 15:31:59 +0100119 int GetSymbolIdentifier();
Steve Blocka7e24c12009-10-30 11:49:00 +0000120 bool SanityCheck();
121
122 Scanner::Location MessageLocation();
123 const char* BuildMessage();
124 Vector<const char*> BuildArgs();
125
Iain Merrick9ac36c92010-09-13 15:29:50 +0100126 int symbol_count() {
Shimeng (Simon) Wang8a31eba2010-12-06 19:01:33 -0800127 return (store_.length() > PreparseDataConstants::kHeaderSize)
128 ? store_[PreparseDataConstants::kSymbolCountOffset]
129 : 0;
Iain Merrick9ac36c92010-09-13 15:29:50 +0100130 }
131 // The following functions should only be called if SanityCheck has
132 // returned true.
Shimeng (Simon) Wang8a31eba2010-12-06 19:01:33 -0800133 bool has_error() { return store_[PreparseDataConstants::kHasErrorOffset]; }
134 unsigned magic() { return store_[PreparseDataConstants::kMagicOffset]; }
135 unsigned version() { return store_[PreparseDataConstants::kVersionOffset]; }
Kristian Monsen0d5e1162010-09-30 15:31:59 +0100136
Steve Blocka7e24c12009-10-30 11:49:00 +0000137 private:
Kristian Monsen80d68ea2010-09-08 11:05:35 +0100138 Vector<unsigned> store_;
Kristian Monsen0d5e1162010-09-30 15:31:59 +0100139 unsigned char* symbol_data_;
140 unsigned char* symbol_data_end_;
Iain Merrick9ac36c92010-09-13 15:29:50 +0100141 int function_index_;
Iain Merrick9ac36c92010-09-13 15:29:50 +0100142 bool owns_store_;
Kristian Monsen80d68ea2010-09-08 11:05:35 +0100143
Steve Blocka7e24c12009-10-30 11:49:00 +0000144 unsigned Read(int position);
145 unsigned* ReadAddress(int position);
Kristian Monsen0d5e1162010-09-30 15:31:59 +0100146 // Reads a number from the current symbols
147 int ReadNumber(byte** source);
Steve Blocka7e24c12009-10-30 11:49:00 +0000148
Iain Merrick9ac36c92010-09-13 15:29:50 +0100149 ScriptDataImpl(const char* backing_store, int length)
150 : store_(reinterpret_cast<unsigned*>(const_cast<char*>(backing_store)),
Ben Murdochf87a2032010-10-22 12:50:53 +0100151 length / static_cast<int>(sizeof(unsigned))),
Iain Merrick9ac36c92010-09-13 15:29:50 +0100152 owns_store_(false) {
Ben Murdochf87a2032010-10-22 12:50:53 +0100153 ASSERT_EQ(0, static_cast<int>(
154 reinterpret_cast<intptr_t>(backing_store) % sizeof(unsigned)));
Iain Merrick9ac36c92010-09-13 15:29:50 +0100155 }
156
Kristian Monsen80d68ea2010-09-08 11:05:35 +0100157 // Read strings written by ParserRecorder::WriteString.
158 static const char* ReadString(unsigned* start, int* chars);
Iain Merrick9ac36c92010-09-13 15:29:50 +0100159
160 friend class ScriptData;
Steve Blocka7e24c12009-10-30 11:49:00 +0000161};
162
163
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800164class ParserApi {
165 public:
Ben Murdochf87a2032010-10-22 12:50:53 +0100166 // Parses the source code represented by the compilation info and sets its
167 // function literal. Returns false (and deallocates any allocated AST
168 // nodes) if parsing failed.
169 static bool Parse(CompilationInfo* info);
Steve Blocka7e24c12009-10-30 11:49:00 +0000170
Ben Murdochf87a2032010-10-22 12:50:53 +0100171 // Generic preparser generating full preparse data.
Ben Murdochb0fe1622011-05-05 13:52:32 +0100172 static ScriptDataImpl* PreParse(UC16CharacterStream* source,
Ben Murdochf87a2032010-10-22 12:50:53 +0100173 v8::Extension* extension);
174
175 // Preparser that only does preprocessing that makes sense if only used
176 // immediately after.
Ben Murdochb0fe1622011-05-05 13:52:32 +0100177 static ScriptDataImpl* PartialPreParse(UC16CharacterStream* source,
Ben Murdochf87a2032010-10-22 12:50:53 +0100178 v8::Extension* extension);
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800179};
180
181// ----------------------------------------------------------------------------
182// REGEXP PARSING
183
184// A BuffferedZoneList is an automatically growing list, just like (and backed
185// by) a ZoneList, that is optimized for the case of adding and removing
186// a single element. The last element added is stored outside the backing list,
187// and if no more than one element is ever added, the ZoneList isn't even
188// allocated.
189// Elements must not be NULL pointers.
190template <typename T, int initial_size>
191class BufferedZoneList {
192 public:
193 BufferedZoneList() : list_(NULL), last_(NULL) {}
194
195 // Adds element at end of list. This element is buffered and can
196 // be read using last() or removed using RemoveLast until a new Add or until
197 // RemoveLast or GetList has been called.
198 void Add(T* value) {
199 if (last_ != NULL) {
200 if (list_ == NULL) {
201 list_ = new ZoneList<T*>(initial_size);
202 }
203 list_->Add(last_);
204 }
205 last_ = value;
206 }
207
208 T* last() {
209 ASSERT(last_ != NULL);
210 return last_;
211 }
212
213 T* RemoveLast() {
214 ASSERT(last_ != NULL);
215 T* result = last_;
216 if ((list_ != NULL) && (list_->length() > 0))
217 last_ = list_->RemoveLast();
218 else
219 last_ = NULL;
220 return result;
221 }
222
223 T* Get(int i) {
224 ASSERT((0 <= i) && (i < length()));
225 if (list_ == NULL) {
226 ASSERT_EQ(0, i);
227 return last_;
228 } else {
229 if (i == list_->length()) {
230 ASSERT(last_ != NULL);
231 return last_;
232 } else {
233 return list_->at(i);
234 }
235 }
236 }
237
238 void Clear() {
239 list_ = NULL;
240 last_ = NULL;
241 }
242
243 int length() {
244 int length = (list_ == NULL) ? 0 : list_->length();
245 return length + ((last_ == NULL) ? 0 : 1);
246 }
247
248 ZoneList<T*>* GetList() {
249 if (list_ == NULL) {
250 list_ = new ZoneList<T*>(initial_size);
251 }
252 if (last_ != NULL) {
253 list_->Add(last_);
254 last_ = NULL;
255 }
256 return list_;
257 }
258
259 private:
260 ZoneList<T*>* list_;
261 T* last_;
262};
263
264
265// Accumulates RegExp atoms and assertions into lists of terms and alternatives.
266class RegExpBuilder: public ZoneObject {
267 public:
268 RegExpBuilder();
269 void AddCharacter(uc16 character);
270 // "Adds" an empty expression. Does nothing except consume a
271 // following quantifier
272 void AddEmpty();
273 void AddAtom(RegExpTree* tree);
274 void AddAssertion(RegExpTree* tree);
275 void NewAlternative(); // '|'
276 void AddQuantifierToAtom(int min, int max, RegExpQuantifier::Type type);
277 RegExpTree* ToRegExp();
278
279 private:
280 void FlushCharacters();
281 void FlushText();
282 void FlushTerms();
283 bool pending_empty_;
284 ZoneList<uc16>* characters_;
285 BufferedZoneList<RegExpTree, 2> terms_;
286 BufferedZoneList<RegExpTree, 2> text_;
287 BufferedZoneList<RegExpTree, 2> alternatives_;
288#ifdef DEBUG
289 enum {ADD_NONE, ADD_CHAR, ADD_TERM, ADD_ASSERT, ADD_ATOM} last_added_;
290#define LAST(x) last_added_ = x;
291#else
292#define LAST(x)
293#endif
294};
295
296
297class RegExpParser {
298 public:
299 RegExpParser(FlatStringReader* in,
300 Handle<String>* error,
301 bool multiline_mode);
Ben Murdochf87a2032010-10-22 12:50:53 +0100302
303 static bool ParseRegExp(FlatStringReader* input,
304 bool multiline,
305 RegExpCompileData* result);
306
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800307 RegExpTree* ParsePattern();
308 RegExpTree* ParseDisjunction();
309 RegExpTree* ParseGroup();
310 RegExpTree* ParseCharacterClass();
Ben Murdochf87a2032010-10-22 12:50:53 +0100311
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800312 // Parses a {...,...} quantifier and stores the range in the given
313 // out parameters.
314 bool ParseIntervalQuantifier(int* min_out, int* max_out);
Steve Block59151502010-09-22 15:07:15 +0100315
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800316 // Parses and returns a single escaped character. The character
317 // must not be 'b' or 'B' since they are usually handle specially.
318 uc32 ParseClassCharacterEscape();
319
320 // Checks whether the following is a length-digit hexadecimal number,
321 // and sets the value if it is.
322 bool ParseHexEscape(int length, uc32* value);
323
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800324 uc32 ParseOctalLiteral();
325
326 // Tries to parse the input as a back reference. If successful it
327 // stores the result in the output parameter and returns true. If
328 // it fails it will push back the characters read so the same characters
329 // can be reparsed.
330 bool ParseBackReferenceIndex(int* index_out);
331
332 CharacterRange ParseClassAtom(uc16* char_class);
333 RegExpTree* ReportError(Vector<const char> message);
334 void Advance();
335 void Advance(int dist);
336 void Reset(int pos);
337
338 // Reports whether the pattern might be used as a literal search string.
339 // Only use if the result of the parse is a single atom node.
340 bool simple();
341 bool contains_anchor() { return contains_anchor_; }
342 void set_contains_anchor() { contains_anchor_ = true; }
343 int captures_started() { return captures_ == NULL ? 0 : captures_->length(); }
344 int position() { return next_pos_ - 1; }
345 bool failed() { return failed_; }
346
347 static const int kMaxCaptures = 1 << 16;
348 static const uc32 kEndMarker = (1 << 21);
349
350 private:
351 enum SubexpressionType {
352 INITIAL,
353 CAPTURE, // All positive values represent captures.
354 POSITIVE_LOOKAHEAD,
355 NEGATIVE_LOOKAHEAD,
356 GROUPING
357 };
358
359 class RegExpParserState : public ZoneObject {
360 public:
361 RegExpParserState(RegExpParserState* previous_state,
362 SubexpressionType group_type,
363 int disjunction_capture_index)
364 : previous_state_(previous_state),
365 builder_(new RegExpBuilder()),
366 group_type_(group_type),
367 disjunction_capture_index_(disjunction_capture_index) {}
368 // Parser state of containing expression, if any.
369 RegExpParserState* previous_state() { return previous_state_; }
370 bool IsSubexpression() { return previous_state_ != NULL; }
371 // RegExpBuilder building this regexp's AST.
372 RegExpBuilder* builder() { return builder_; }
373 // Type of regexp being parsed (parenthesized group or entire regexp).
374 SubexpressionType group_type() { return group_type_; }
375 // Index in captures array of first capture in this sub-expression, if any.
376 // Also the capture index of this sub-expression itself, if group_type
377 // is CAPTURE.
378 int capture_index() { return disjunction_capture_index_; }
379
380 private:
381 // Linked list implementation of stack of states.
382 RegExpParserState* previous_state_;
383 // Builder for the stored disjunction.
384 RegExpBuilder* builder_;
385 // Stored disjunction type (capture, look-ahead or grouping), if any.
386 SubexpressionType group_type_;
387 // Stored disjunction's capture index (if any).
388 int disjunction_capture_index_;
389 };
390
391 uc32 current() { return current_; }
392 bool has_more() { return has_more_; }
393 bool has_next() { return next_pos_ < in()->length(); }
394 uc32 Next();
395 FlatStringReader* in() { return in_; }
396 void ScanForCaptures();
397
398 Handle<String>* error_;
399 ZoneList<RegExpCapture*>* captures_;
400 FlatStringReader* in_;
401 uc32 current_;
402 int next_pos_;
403 // The capture count is only valid after we have scanned for captures.
404 int capture_count_;
405 bool has_more_;
406 bool multiline_;
407 bool simple_;
408 bool contains_anchor_;
409 bool is_scanned_for_captures_;
410 bool failed_;
411};
412
413// ----------------------------------------------------------------------------
414// JAVASCRIPT PARSING
415
416class Parser {
417 public:
418 Parser(Handle<Script> script,
419 bool allow_natives_syntax,
420 v8::Extension* extension,
421 ScriptDataImpl* pre_data);
422 virtual ~Parser() { }
Steve Blocka7e24c12009-10-30 11:49:00 +0000423
Ben Murdochf87a2032010-10-22 12:50:53 +0100424 // Returns NULL if parsing failed.
425 FunctionLiteral* ParseProgram(Handle<String> source,
Steve Block1e0659c2011-05-24 12:43:12 +0100426 bool in_global_context,
427 StrictModeFlag strict_mode);
Ben Murdochf87a2032010-10-22 12:50:53 +0100428
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800429 FunctionLiteral* ParseLazy(Handle<SharedFunctionInfo> info);
430
431 void ReportMessageAt(Scanner::Location loc,
432 const char* message,
433 Vector<const char*> args);
Ben Murdochb8e0da22011-05-16 14:20:40 +0100434 void ReportMessageAt(Scanner::Location loc,
435 const char* message,
436 Vector<Handle<String> > args);
Ben Murdochf87a2032010-10-22 12:50:53 +0100437
438 protected:
Steve Block1e0659c2011-05-24 12:43:12 +0100439 // Limit on number of function parameters is chosen arbitrarily.
440 // Code::Flags uses only the low 17 bits of num-parameters to
441 // construct a hashable id, so if more than 2^17 are allowed, this
442 // should be checked.
443 static const int kMaxNumFunctionParameters = 32766;
Ben Murdochb0fe1622011-05-05 13:52:32 +0100444 FunctionLiteral* ParseLazy(Handle<SharedFunctionInfo> info,
445 UC16CharacterStream* source,
446 ZoneScope* zone_scope);
Ben Murdochf87a2032010-10-22 12:50:53 +0100447 enum Mode {
448 PARSE_LAZILY,
449 PARSE_EAGERLY
450 };
451
Ben Murdochb0fe1622011-05-05 13:52:32 +0100452 // Called by ParseProgram after setting up the scanner.
453 FunctionLiteral* DoParseProgram(Handle<String> source,
454 bool in_global_context,
Steve Block1e0659c2011-05-24 12:43:12 +0100455 StrictModeFlag strict_mode,
Ben Murdochb0fe1622011-05-05 13:52:32 +0100456 ZoneScope* zone_scope);
457
Ben Murdochf87a2032010-10-22 12:50:53 +0100458 // Report syntax error
459 void ReportUnexpectedToken(Token::Value token);
460 void ReportInvalidPreparseData(Handle<String> name, bool* ok);
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800461 void ReportMessage(const char* message, Vector<const char*> args);
Ben Murdochf87a2032010-10-22 12:50:53 +0100462
463 bool inside_with() const { return with_nesting_level_ > 0; }
Ben Murdochb0fe1622011-05-05 13:52:32 +0100464 V8JavaScriptScanner& scanner() { return scanner_; }
Ben Murdochf87a2032010-10-22 12:50:53 +0100465 Mode mode() const { return mode_; }
466 ScriptDataImpl* pre_data() const { return pre_data_; }
467
468 // All ParseXXX functions take as the last argument an *ok parameter
469 // which is set to false if parsing failed; it is unchanged otherwise.
470 // By making the 'exception handling' explicit, we are forced to check
471 // for failure at the call sites.
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800472 void* ParseSourceElements(ZoneList<Statement*>* processor,
Ben Murdochf87a2032010-10-22 12:50:53 +0100473 int end_token, bool* ok);
474 Statement* ParseStatement(ZoneStringList* labels, bool* ok);
475 Statement* ParseFunctionDeclaration(bool* ok);
476 Statement* ParseNativeDeclaration(bool* ok);
477 Block* ParseBlock(ZoneStringList* labels, bool* ok);
478 Block* ParseVariableStatement(bool* ok);
479 Block* ParseVariableDeclarations(bool accept_IN, Expression** var, bool* ok);
480 Statement* ParseExpressionOrLabelledStatement(ZoneStringList* labels,
481 bool* ok);
482 IfStatement* ParseIfStatement(ZoneStringList* labels, bool* ok);
483 Statement* ParseContinueStatement(bool* ok);
484 Statement* ParseBreakStatement(ZoneStringList* labels, bool* ok);
485 Statement* ParseReturnStatement(bool* ok);
486 Block* WithHelper(Expression* obj,
487 ZoneStringList* labels,
488 bool is_catch_block,
489 bool* ok);
490 Statement* ParseWithStatement(ZoneStringList* labels, bool* ok);
491 CaseClause* ParseCaseClause(bool* default_seen_ptr, bool* ok);
492 SwitchStatement* ParseSwitchStatement(ZoneStringList* labels, bool* ok);
493 DoWhileStatement* ParseDoWhileStatement(ZoneStringList* labels, bool* ok);
494 WhileStatement* ParseWhileStatement(ZoneStringList* labels, bool* ok);
495 Statement* ParseForStatement(ZoneStringList* labels, bool* ok);
496 Statement* ParseThrowStatement(bool* ok);
497 Expression* MakeCatchContext(Handle<String> id, VariableProxy* value);
498 TryStatement* ParseTryStatement(bool* ok);
499 DebuggerStatement* ParseDebuggerStatement(bool* ok);
500
501 Expression* ParseExpression(bool accept_IN, bool* ok);
502 Expression* ParseAssignmentExpression(bool accept_IN, bool* ok);
503 Expression* ParseConditionalExpression(bool accept_IN, bool* ok);
504 Expression* ParseBinaryExpression(int prec, bool accept_IN, bool* ok);
505 Expression* ParseUnaryExpression(bool* ok);
506 Expression* ParsePostfixExpression(bool* ok);
507 Expression* ParseLeftHandSideExpression(bool* ok);
508 Expression* ParseNewExpression(bool* ok);
509 Expression* ParseMemberExpression(bool* ok);
510 Expression* ParseNewPrefix(PositionStack* stack, bool* ok);
511 Expression* ParseMemberWithNewPrefixesExpression(PositionStack* stack,
512 bool* ok);
513 Expression* ParsePrimaryExpression(bool* ok);
514 Expression* ParseArrayLiteral(bool* ok);
515 Expression* ParseObjectLiteral(bool* ok);
516 ObjectLiteral::Property* ParseObjectLiteralGetSet(bool is_getter, bool* ok);
517 Expression* ParseRegExpLiteral(bool seen_equal, bool* ok);
518
519 Expression* NewCompareNode(Token::Value op,
520 Expression* x,
521 Expression* y,
522 int position);
523
524 // Populate the constant properties fixed array for a materialized object
525 // literal.
526 void BuildObjectLiteralConstantProperties(
527 ZoneList<ObjectLiteral::Property*>* properties,
528 Handle<FixedArray> constants,
529 bool* is_simple,
530 bool* fast_elements,
531 int* depth);
532
533 // Populate the literals fixed array for a materialized array literal.
534 void BuildArrayLiteralBoilerplateLiterals(ZoneList<Expression*>* properties,
535 Handle<FixedArray> constants,
536 bool* is_simple,
537 int* depth);
538
539 // Decide if a property should be in the object boilerplate.
540 bool IsBoilerplateProperty(ObjectLiteral::Property* property);
541 // If the expression is a literal, return the literal value;
542 // if the expression is a materialized literal and is simple return a
543 // compile time value as encoded by CompileTimeValue::GetValue().
544 // Otherwise, return undefined literal as the placeholder
545 // in the object literal boilerplate.
546 Handle<Object> GetBoilerplateValue(Expression* expression);
547
548 enum FunctionLiteralType {
549 EXPRESSION,
550 DECLARATION,
551 NESTED
552 };
553
554 ZoneList<Expression*>* ParseArguments(bool* ok);
555 FunctionLiteral* ParseFunctionLiteral(Handle<String> var_name,
Steve Block1e0659c2011-05-24 12:43:12 +0100556 bool name_is_reserved,
Ben Murdochf87a2032010-10-22 12:50:53 +0100557 int function_token_position,
558 FunctionLiteralType type,
559 bool* ok);
Steve Blocka7e24c12009-10-30 11:49:00 +0000560
561
Ben Murdochf87a2032010-10-22 12:50:53 +0100562 // Magical syntax support.
563 Expression* ParseV8Intrinsic(bool* ok);
564
Ben Murdochb0fe1622011-05-05 13:52:32 +0100565 INLINE(Token::Value peek()) {
566 if (stack_overflow_) return Token::ILLEGAL;
567 return scanner().peek();
568 }
569
570 INLINE(Token::Value Next()) {
571 // BUG 1215673: Find a thread safe way to set a stack limit in
572 // pre-parse mode. Otherwise, we cannot safely pre-parse from other
573 // threads.
574 if (stack_overflow_) {
575 return Token::ILLEGAL;
576 }
577 if (StackLimitCheck().HasOverflowed()) {
578 // Any further calls to Next or peek will return the illegal token.
579 // The current call must return the next token, which might already
580 // have been peek'ed.
581 stack_overflow_ = true;
582 }
583 return scanner().Next();
584 }
585
Steve Block1e0659c2011-05-24 12:43:12 +0100586 bool peek_any_identifier();
587
Ben Murdochf87a2032010-10-22 12:50:53 +0100588 INLINE(void Consume(Token::Value token));
589 void Expect(Token::Value token, bool* ok);
590 bool Check(Token::Value token);
591 void ExpectSemicolon(bool* ok);
592
Steve Block9fac8402011-05-12 15:51:54 +0100593 Handle<String> LiteralString(PretenureFlag tenured) {
594 if (scanner().is_literal_ascii()) {
595 return Factory::NewStringFromAscii(scanner().literal_ascii_string(),
596 tenured);
597 } else {
598 return Factory::NewStringFromTwoByte(scanner().literal_uc16_string(),
599 tenured);
600 }
601 }
602
603 Handle<String> NextLiteralString(PretenureFlag tenured) {
604 if (scanner().is_next_literal_ascii()) {
605 return Factory::NewStringFromAscii(scanner().next_literal_ascii_string(),
606 tenured);
607 } else {
608 return Factory::NewStringFromTwoByte(scanner().next_literal_uc16_string(),
609 tenured);
610 }
611 }
612
Ben Murdochf87a2032010-10-22 12:50:53 +0100613 Handle<String> GetSymbol(bool* ok);
614
615 // Get odd-ball literals.
616 Literal* GetLiteralUndefined();
617 Literal* GetLiteralTheHole();
618 Literal* GetLiteralNumber(double value);
619
620 Handle<String> ParseIdentifier(bool* ok);
Steve Block1e0659c2011-05-24 12:43:12 +0100621 Handle<String> ParseIdentifierOrReservedWord(bool* is_reserved, bool* ok);
Ben Murdochf87a2032010-10-22 12:50:53 +0100622 Handle<String> ParseIdentifierName(bool* ok);
623 Handle<String> ParseIdentifierOrGetOrSet(bool* is_get,
624 bool* is_set,
625 bool* ok);
626
Steve Block1e0659c2011-05-24 12:43:12 +0100627 // Strict mode validation of LValue expressions
628 void CheckStrictModeLValue(Expression* expression,
629 const char* error,
630 bool* ok);
631
632 // Strict mode octal literal validation.
633 void CheckOctalLiteral(int beg_pos, int end_pos, bool* ok);
634
Ben Murdochf87a2032010-10-22 12:50:53 +0100635 // Parser support
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800636 VariableProxy* Declare(Handle<String> name, Variable::Mode mode,
637 FunctionLiteral* fun,
638 bool resolve,
639 bool* ok);
Ben Murdochf87a2032010-10-22 12:50:53 +0100640
641 bool TargetStackContainsLabel(Handle<String> label);
642 BreakableStatement* LookupBreakTarget(Handle<String> label, bool* ok);
643 IterationStatement* LookupContinueTarget(Handle<String> label, bool* ok);
644
645 void RegisterTargetUse(BreakTarget* target, Target* stop);
646
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800647 // Factory methods.
648
649 Statement* EmptyStatement() {
650 static v8::internal::EmptyStatement empty;
651 return &empty;
652 }
653
654 Scope* NewScope(Scope* parent, Scope::Type type, bool inside_with);
655
Steve Block9fac8402011-05-12 15:51:54 +0100656 Handle<String> LookupSymbol(int symbol_id);
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800657
Steve Block9fac8402011-05-12 15:51:54 +0100658 Handle<String> LookupCachedSymbol(int symbol_id);
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800659
660 Expression* NewCall(Expression* expression,
661 ZoneList<Expression*>* arguments,
662 int pos) {
663 return new Call(expression, arguments, pos);
664 }
665
666
Ben Murdochf87a2032010-10-22 12:50:53 +0100667 // Create a number literal.
668 Literal* NewNumberLiteral(double value);
669
670 // Generate AST node that throw a ReferenceError with the given type.
671 Expression* NewThrowReferenceError(Handle<String> type);
672
673 // Generate AST node that throw a SyntaxError with the given
674 // type. The first argument may be null (in the handle sense) in
675 // which case no arguments are passed to the constructor.
676 Expression* NewThrowSyntaxError(Handle<String> type, Handle<Object> first);
677
678 // Generate AST node that throw a TypeError with the given
679 // type. Both arguments must be non-null (in the handle sense).
680 Expression* NewThrowTypeError(Handle<String> type,
681 Handle<Object> first,
682 Handle<Object> second);
683
684 // Generic AST generator for throwing errors from compiled code.
685 Expression* NewThrowError(Handle<String> constructor,
686 Handle<String> type,
687 Vector< Handle<Object> > arguments);
688
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800689 ZoneList<Handle<String> > symbol_cache_;
Ben Murdochf87a2032010-10-22 12:50:53 +0100690
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800691 Handle<Script> script_;
Shimeng (Simon) Wang8a31eba2010-12-06 19:01:33 -0800692 V8JavaScriptScanner scanner_;
Ben Murdochf87a2032010-10-22 12:50:53 +0100693
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800694 Scope* top_scope_;
695 int with_nesting_level_;
Ben Murdochf87a2032010-10-22 12:50:53 +0100696
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800697 TemporaryScope* temp_scope_;
698 Mode mode_;
699
700 Target* target_stack_; // for break, continue statements
701 bool allow_natives_syntax_;
702 v8::Extension* extension_;
703 bool is_pre_parsing_;
704 ScriptDataImpl* pre_data_;
705 FuncNameInferrer* fni_;
Ben Murdochb0fe1622011-05-05 13:52:32 +0100706 bool stack_overflow_;
Ben Murdochb8e0da22011-05-16 14:20:40 +0100707 // If true, the next (and immediately following) function literal is
708 // preceded by a parenthesis.
709 // Heuristically that means that the function will be called immediately,
710 // so never lazily compile it.
711 bool parenthesized_function_;
Ben Murdochf87a2032010-10-22 12:50:53 +0100712};
Steve Blocka7e24c12009-10-30 11:49:00 +0000713
714
715// Support for handling complex values (array and object literals) that
716// can be fully handled at compile time.
717class CompileTimeValue: public AllStatic {
718 public:
719 enum Type {
Steve Block6ded16b2010-05-10 14:33:55 +0100720 OBJECT_LITERAL_FAST_ELEMENTS,
721 OBJECT_LITERAL_SLOW_ELEMENTS,
Steve Blocka7e24c12009-10-30 11:49:00 +0000722 ARRAY_LITERAL
723 };
724
725 static bool IsCompileTimeValue(Expression* expression);
726
Iain Merrick75681382010-08-19 15:07:18 +0100727 static bool ArrayLiteralElementNeedsInitialization(Expression* value);
728
Steve Blocka7e24c12009-10-30 11:49:00 +0000729 // Get the value as a compile time value.
730 static Handle<FixedArray> GetValue(Expression* expression);
731
732 // Get the type of a compile time value returned by GetValue().
733 static Type GetType(Handle<FixedArray> value);
734
735 // Get the elements array of a compile time value returned by GetValue().
736 static Handle<FixedArray> GetElements(Handle<FixedArray> value);
737
738 private:
739 static const int kTypeSlot = 0;
740 static const int kElementsSlot = 1;
741
742 DISALLOW_IMPLICIT_CONSTRUCTORS(CompileTimeValue);
743};
744
745
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800746// ----------------------------------------------------------------------------
747// JSON PARSING
748
749// JSON is a subset of JavaScript, as specified in, e.g., the ECMAScript 5
750// specification section 15.12.1 (and appendix A.8).
751// The grammar is given section 15.12.1.2 (and appendix A.8.2).
752class JsonParser BASE_EMBEDDED {
753 public:
754 // Parse JSON input as a single JSON value.
755 // Returns null handle and sets exception if parsing failed.
756 static Handle<Object> Parse(Handle<String> source) {
Ben Murdochb0fe1622011-05-05 13:52:32 +0100757 if (source->IsExternalTwoByteString()) {
758 ExternalTwoByteStringUC16CharacterStream stream(
759 Handle<ExternalTwoByteString>::cast(source), 0, source->length());
760 return JsonParser().ParseJson(source, &stream);
761 } else {
762 GenericStringUC16CharacterStream stream(source, 0, source->length());
763 return JsonParser().ParseJson(source, &stream);
764 }
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800765 }
766
767 private:
768 JsonParser() { }
769 ~JsonParser() { }
770
771 // Parse a string containing a single JSON value.
Ben Murdochb0fe1622011-05-05 13:52:32 +0100772 Handle<Object> ParseJson(Handle<String> script, UC16CharacterStream* source);
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800773 // Parse a single JSON value from input (grammar production JSONValue).
774 // A JSON value is either a (double-quoted) string literal, a number literal,
775 // one of "true", "false", or "null", or an object or array literal.
776 Handle<Object> ParseJsonValue();
777 // Parse a JSON object literal (grammar production JSONObject).
778 // An object literal is a squiggly-braced and comma separated sequence
779 // (possibly empty) of key/value pairs, where the key is a JSON string
780 // literal, the value is a JSON value, and the two are separated by a colon.
781 // A JSON array dosn't allow numbers and identifiers as keys, like a
782 // JavaScript array.
783 Handle<Object> ParseJsonObject();
784 // Parses a JSON array literal (grammar production JSONArray). An array
785 // literal is a square-bracketed and comma separated sequence (possibly empty)
786 // of JSON values.
787 // A JSON array doesn't allow leaving out values from the sequence, nor does
788 // it allow a terminal comma, like a JavaScript array does.
789 Handle<Object> ParseJsonArray();
790
791 // Mark that a parsing error has happened at the current token, and
792 // return a null handle. Primarily for readability.
793 Handle<Object> ReportUnexpectedToken() { return Handle<Object>::null(); }
794 // Converts the currently parsed literal to a JavaScript String.
795 Handle<String> GetString();
796
Shimeng (Simon) Wang8a31eba2010-12-06 19:01:33 -0800797 JsonScanner scanner_;
Ben Murdochb0fe1622011-05-05 13:52:32 +0100798 bool stack_overflow_;
Teng-Hui Zhu3e5fa292010-11-09 16:16:48 -0800799};
Steve Blocka7e24c12009-10-30 11:49:00 +0000800} } // namespace v8::internal
801
802#endif // V8_PARSER_H_