1 # |
|
2 # Jasy - Web Tooling Framework |
|
3 # Copyright 2010-2012 Zynga Inc. |
|
4 # Copyright 2013-2014 Sebastian Werner |
|
5 # |
|
6 |
|
7 # |
|
8 # License: MPL 1.1/GPL 2.0/LGPL 2.1 |
|
9 # Authors: |
|
10 # - Brendan Eich <brendan@mozilla.org> (Original JavaScript) (2004-2010) |
|
11 # - Sebastian Werner <info@sebastian-werner.net> (Python Port) (2010-2012) |
|
12 # |
|
13 |
|
14 from __future__ import unicode_literals |
|
15 |
|
16 import jasy.js.tokenize.Tokenizer |
|
17 import jasy.js.parse.VanillaBuilder |
|
18 import jasy.js.tokenize.Lang |
|
19 |
|
20 __all__ = [ "parse", "parseExpression" ] |
|
21 |
|
22 def parseExpression(source, fileId=None, line=1, builder=None): |
|
23 if builder == None: |
|
24 builder = jasy.js.parse.VanillaBuilder.VanillaBuilder() |
|
25 |
|
26 # Convert source into expression statement to be friendly to the Tokenizer |
|
27 if not source.endswith(";"): |
|
28 source = source + ";" |
|
29 |
|
30 tokenizer = jasy.js.tokenize.Tokenizer.Tokenizer(source, fileId, line) |
|
31 staticContext = StaticContext(False, builder) |
|
32 |
|
33 return Expression(tokenizer, staticContext) |
|
34 |
|
35 |
|
36 |
|
37 def parse(source, fileId=None, line=1, builder=None): |
|
38 if builder == None: |
|
39 builder = jasy.js.parse.VanillaBuilder.VanillaBuilder() |
|
40 |
|
41 tokenizer = jasy.js.tokenize.Tokenizer.Tokenizer(source, fileId, line) |
|
42 staticContext = StaticContext(False, builder) |
|
43 node = Script(tokenizer, staticContext) |
|
44 |
|
45 # store fileId on top-level node |
|
46 node.fileId = tokenizer.fileId |
|
47 |
|
48 # add missing comments e.g. empty file with only a comment etc. |
|
49 # if there is something non-attached by an inner node it is attached to |
|
50 # the top level node, which is not correct, but might be better than |
|
51 # just ignoring the comment after all. |
|
52 if len(node) > 0: |
|
53 builder.COMMENTS_add(node[-1], None, tokenizer.getComments()) |
|
54 else: |
|
55 builder.COMMENTS_add(node, None, tokenizer.getComments()) |
|
56 |
|
57 if not tokenizer.done(): |
|
58 raise SyntaxError("Unexpected end of file", tokenizer) |
|
59 |
|
60 return node |
|
61 |
|
62 |
|
63 |
|
64 class SyntaxError(Exception): |
|
65 def __init__(self, message, tokenizer): |
|
66 Exception.__init__(self, "Syntax error: %s\n%s:%s" % (message, tokenizer.fileId, tokenizer.line)) |
|
67 |
|
68 |
|
69 # Used as a status container during tree-building for every def body and the global body |
|
70 class StaticContext(object): |
|
71 # inFunction is used to check if a return stm appears in a valid context. |
|
72 def __init__(self, inFunction, builder): |
|
73 # Whether this is inside a function, mostly True, only for top-level scope it's False |
|
74 self.inFunction = inFunction |
|
75 |
|
76 self.hasEmptyReturn = False |
|
77 self.hasReturnWithValue = False |
|
78 self.isGenerator = False |
|
79 self.blockId = 0 |
|
80 self.builder = builder |
|
81 self.statementStack = [] |
|
82 |
|
83 # Sets to store variable uses |
|
84 # self.functions = set() |
|
85 # self.variables = set() |
|
86 |
|
87 # Status |
|
88 # self.needsHoisting = False |
|
89 self.bracketLevel = 0 |
|
90 self.curlyLevel = 0 |
|
91 self.parenLevel = 0 |
|
92 self.hookLevel = 0 |
|
93 |
|
94 # Configure strict ecmascript 3 mode |
|
95 self.ecma3OnlyMode = False |
|
96 |
|
97 # Status flag during parsing |
|
98 self.inForLoopInit = False |
|
99 |
|
100 |
|
101 def Script(tokenizer, staticContext): |
|
102 """Parses the toplevel and def bodies.""" |
|
103 node = Statements(tokenizer, staticContext) |
|
104 |
|
105 # change type from "block" to "script" for script root |
|
106 node.type = "script" |
|
107 |
|
108 # copy over data from compiler context |
|
109 # node.functions = staticContext.functions |
|
110 # node.variables = staticContext.variables |
|
111 |
|
112 return node |
|
113 |
|
114 |
|
115 def nest(tokenizer, staticContext, node, func, end=None): |
|
116 """Statement stack and nested statement handler.""" |
|
117 staticContext.statementStack.append(node) |
|
118 node = func(tokenizer, staticContext) |
|
119 staticContext.statementStack.pop() |
|
120 end and tokenizer.mustMatch(end) |
|
121 |
|
122 return node |
|
123 |
|
124 |
|
125 def Statements(tokenizer, staticContext): |
|
126 """Parses a list of Statements.""" |
|
127 |
|
128 builder = staticContext.builder |
|
129 node = builder.BLOCK_build(tokenizer, staticContext.blockId) |
|
130 staticContext.blockId += 1 |
|
131 |
|
132 builder.BLOCK_hoistLets(node) |
|
133 staticContext.statementStack.append(node) |
|
134 |
|
135 prevNode = None |
|
136 while not tokenizer.done() and tokenizer.peek(True) != "right_curly": |
|
137 comments = tokenizer.getComments() |
|
138 childNode = Statement(tokenizer, staticContext) |
|
139 builder.COMMENTS_add(childNode, prevNode, comments) |
|
140 builder.BLOCK_addStatement(node, childNode) |
|
141 prevNode = childNode |
|
142 |
|
143 staticContext.statementStack.pop() |
|
144 builder.BLOCK_finish(node) |
|
145 |
|
146 # if getattr(node, "needsHoisting", False): |
|
147 # # TODO |
|
148 # raise Exception("Needs hoisting went true!!!") |
|
149 # builder.setHoists(node.id, node.variables) |
|
150 # # Propagate up to the function. |
|
151 # staticContext.needsHoisting = True |
|
152 |
|
153 return node |
|
154 |
|
155 |
|
156 def Block(tokenizer, staticContext): |
|
157 tokenizer.mustMatch("left_curly") |
|
158 node = Statements(tokenizer, staticContext) |
|
159 tokenizer.mustMatch("right_curly") |
|
160 |
|
161 return node |
|
162 |
|
163 |
|
164 def Statement(tokenizer, staticContext): |
|
165 """Parses a Statement.""" |
|
166 |
|
167 tokenType = tokenizer.get(True) |
|
168 builder = staticContext.builder |
|
169 |
|
170 # Cases for statements ending in a right curly return early, avoiding the |
|
171 # common semicolon insertion magic after this switch. |
|
172 |
|
173 if tokenType == "function": |
|
174 # "declared_form" extends functions of staticContext, "statement_form" doesn'tokenizer. |
|
175 if len(staticContext.statementStack) > 1: |
|
176 kind = "statement_form" |
|
177 else: |
|
178 kind = "declared_form" |
|
179 |
|
180 return FunctionDefinition(tokenizer, staticContext, True, kind) |
|
181 |
|
182 |
|
183 elif tokenType == "left_curly": |
|
184 node = Statements(tokenizer, staticContext) |
|
185 tokenizer.mustMatch("right_curly") |
|
186 |
|
187 return node |
|
188 |
|
189 |
|
190 elif tokenType == "if": |
|
191 node = builder.IF_build(tokenizer) |
|
192 builder.IF_setCondition(node, ParenExpression(tokenizer, staticContext)) |
|
193 staticContext.statementStack.append(node) |
|
194 builder.IF_setThenPart(node, Statement(tokenizer, staticContext)) |
|
195 |
|
196 if tokenizer.match("else"): |
|
197 comments = tokenizer.getComments() |
|
198 elsePart = Statement(tokenizer, staticContext) |
|
199 builder.COMMENTS_add(elsePart, node, comments) |
|
200 builder.IF_setElsePart(node, elsePart) |
|
201 |
|
202 staticContext.statementStack.pop() |
|
203 builder.IF_finish(node) |
|
204 |
|
205 return node |
|
206 |
|
207 |
|
208 elif tokenType == "switch": |
|
209 # This allows CASEs after a "default", which is in the standard. |
|
210 node = builder.SWITCH_build(tokenizer) |
|
211 builder.SWITCH_setDiscriminant(node, ParenExpression(tokenizer, staticContext)) |
|
212 staticContext.statementStack.append(node) |
|
213 |
|
214 tokenizer.mustMatch("left_curly") |
|
215 tokenType = tokenizer.get() |
|
216 |
|
217 while tokenType != "right_curly": |
|
218 if tokenType == "default": |
|
219 if node.defaultIndex >= 0: |
|
220 raise SyntaxError("More than one switch default", tokenizer) |
|
221 |
|
222 childNode = builder.DEFAULT_build(tokenizer) |
|
223 builder.SWITCH_setDefaultIndex(node, len(node)-1) |
|
224 tokenizer.mustMatch("colon") |
|
225 builder.DEFAULT_initializeStatements(childNode, tokenizer) |
|
226 |
|
227 while True: |
|
228 tokenType=tokenizer.peek(True) |
|
229 if tokenType == "case" or tokenType == "default" or tokenType == "right_curly": |
|
230 break |
|
231 builder.DEFAULT_addStatement(childNode, Statement(tokenizer, staticContext)) |
|
232 |
|
233 builder.DEFAULT_finish(childNode) |
|
234 |
|
235 elif tokenType == "case": |
|
236 childNode = builder.CASE_build(tokenizer) |
|
237 builder.CASE_setLabel(childNode, Expression(tokenizer, staticContext)) |
|
238 tokenizer.mustMatch("colon") |
|
239 builder.CASE_initializeStatements(childNode, tokenizer) |
|
240 |
|
241 while True: |
|
242 tokenType=tokenizer.peek(True) |
|
243 if tokenType == "case" or tokenType == "default" or tokenType == "right_curly": |
|
244 break |
|
245 builder.CASE_addStatement(childNode, Statement(tokenizer, staticContext)) |
|
246 |
|
247 builder.CASE_finish(childNode) |
|
248 |
|
249 else: |
|
250 raise SyntaxError("Invalid switch case", tokenizer) |
|
251 |
|
252 builder.SWITCH_addCase(node, childNode) |
|
253 tokenType = tokenizer.get() |
|
254 |
|
255 staticContext.statementStack.pop() |
|
256 builder.SWITCH_finish(node) |
|
257 |
|
258 return node |
|
259 |
|
260 |
|
261 elif tokenType == "for": |
|
262 node = builder.FOR_build(tokenizer) |
|
263 forBlock = None |
|
264 |
|
265 if tokenizer.match("identifier") and tokenizer.token.value == "each": |
|
266 builder.FOR_rebuildForEach(node) |
|
267 |
|
268 tokenizer.mustMatch("left_paren") |
|
269 tokenType = tokenizer.peek() |
|
270 childNode = None |
|
271 |
|
272 if tokenType != "semicolon": |
|
273 staticContext.inForLoopInit = True |
|
274 |
|
275 if tokenType == "var" or tokenType == "const": |
|
276 tokenizer.get() |
|
277 childNode = Variables(tokenizer, staticContext) |
|
278 |
|
279 elif tokenType == "let": |
|
280 tokenizer.get() |
|
281 |
|
282 if tokenizer.peek() == "left_paren": |
|
283 childNode = LetBlock(tokenizer, staticContext, False) |
|
284 |
|
285 else: |
|
286 # Let in for head, we need to add an implicit block |
|
287 # around the rest of the for. |
|
288 forBlock = builder.BLOCK_build(tokenizer, staticContext.blockId) |
|
289 staticContext.blockId += 1 |
|
290 staticContext.statementStack.append(forBlock) |
|
291 childNode = Variables(tokenizer, staticContext, forBlock) |
|
292 |
|
293 else: |
|
294 childNode = Expression(tokenizer, staticContext) |
|
295 |
|
296 staticContext.inForLoopInit = False |
|
297 |
|
298 if childNode and tokenizer.match("in"): |
|
299 builder.FOR_rebuildForIn(node) |
|
300 builder.FOR_setObject(node, Expression(tokenizer, staticContext), forBlock) |
|
301 |
|
302 if childNode.type == "var" or childNode.type == "let": |
|
303 if len(childNode) != 1: |
|
304 raise SyntaxError("Invalid for..in left-hand side", tokenizer) |
|
305 |
|
306 builder.FOR_setIterator(node, childNode, forBlock) |
|
307 |
|
308 else: |
|
309 builder.FOR_setIterator(node, childNode, forBlock) |
|
310 |
|
311 else: |
|
312 builder.FOR_setSetup(node, childNode) |
|
313 tokenizer.mustMatch("semicolon") |
|
314 |
|
315 if node.isEach: |
|
316 raise SyntaxError("Invalid for each..in loop", tokenizer) |
|
317 |
|
318 if tokenizer.peek() == "semicolon": |
|
319 builder.FOR_setCondition(node, None) |
|
320 else: |
|
321 builder.FOR_setCondition(node, Expression(tokenizer, staticContext)) |
|
322 |
|
323 tokenizer.mustMatch("semicolon") |
|
324 |
|
325 if tokenizer.peek() == "right_paren": |
|
326 builder.FOR_setUpdate(node, None) |
|
327 else: |
|
328 builder.FOR_setUpdate(node, Expression(tokenizer, staticContext)) |
|
329 |
|
330 tokenizer.mustMatch("right_paren") |
|
331 builder.FOR_setBody(node, nest(tokenizer, staticContext, node, Statement)) |
|
332 |
|
333 if forBlock: |
|
334 builder.BLOCK_finish(forBlock) |
|
335 staticContext.statementStack.pop() |
|
336 |
|
337 builder.FOR_finish(node) |
|
338 return node |
|
339 |
|
340 |
|
341 elif tokenType == "while": |
|
342 node = builder.WHILE_build(tokenizer) |
|
343 |
|
344 builder.WHILE_setCondition(node, ParenExpression(tokenizer, staticContext)) |
|
345 builder.WHILE_setBody(node, nest(tokenizer, staticContext, node, Statement)) |
|
346 builder.WHILE_finish(node) |
|
347 |
|
348 return node |
|
349 |
|
350 |
|
351 elif tokenType == "do": |
|
352 node = builder.DO_build(tokenizer) |
|
353 |
|
354 builder.DO_setBody(node, nest(tokenizer, staticContext, node, Statement, "while")) |
|
355 builder.DO_setCondition(node, ParenExpression(tokenizer, staticContext)) |
|
356 builder.DO_finish(node) |
|
357 |
|
358 if not staticContext.ecma3OnlyMode: |
|
359 # <script language="JavaScript"> (without version hints) may need |
|
360 # automatic semicolon insertion without a newline after do-while. |
|
361 # See http://bugzilla.mozilla.org/show_bug.cgi?id=238945. |
|
362 tokenizer.match("semicolon") |
|
363 return node |
|
364 |
|
365 # NO RETURN |
|
366 |
|
367 |
|
368 elif tokenType == "break" or tokenType == "continue": |
|
369 if tokenType == "break": |
|
370 node = builder.BREAK_build(tokenizer) |
|
371 else: |
|
372 node = builder.CONTINUE_build(tokenizer) |
|
373 |
|
374 if tokenizer.peekOnSameLine() == "identifier": |
|
375 tokenizer.get() |
|
376 |
|
377 if tokenType == "break": |
|
378 builder.BREAK_setLabel(node, tokenizer.token.value) |
|
379 else: |
|
380 builder.CONTINUE_setLabel(node, tokenizer.token.value) |
|
381 |
|
382 statementStack = staticContext.statementStack |
|
383 i = len(statementStack) |
|
384 label = node.label if hasattr(node, "label") else None |
|
385 |
|
386 if label: |
|
387 while True: |
|
388 i -= 1 |
|
389 if i < 0: |
|
390 raise SyntaxError("Label not found", tokenizer) |
|
391 if getattr(statementStack[i], "label", None) == label: |
|
392 break |
|
393 |
|
394 # |
|
395 # Both break and continue to label need to be handled specially |
|
396 # within a labeled loop, so that they target that loop. If not in |
|
397 # a loop, then break targets its labeled statement. Labels can be |
|
398 # nested so we skip all labels immediately enclosing the nearest |
|
399 # non-label statement. |
|
400 # |
|
401 while i < len(statementStack) - 1 and statementStack[i+1].type == "label": |
|
402 i += 1 |
|
403 |
|
404 if i < len(statementStack) - 1 and getattr(statementStack[i+1], "isLoop", False): |
|
405 i += 1 |
|
406 elif tokenType == "continue": |
|
407 raise SyntaxError("Invalid continue", tokenizer) |
|
408 |
|
409 else: |
|
410 while True: |
|
411 i -= 1 |
|
412 if i < 0: |
|
413 if tokenType == "break": |
|
414 raise SyntaxError("Invalid break", tokenizer) |
|
415 else: |
|
416 raise SyntaxError("Invalid continue", tokenizer) |
|
417 |
|
418 if getattr(statementStack[i], "isLoop", False) or (tokenType == "break" and statementStack[i].type == "switch"): |
|
419 break |
|
420 |
|
421 if tokenType == "break": |
|
422 builder.BREAK_finish(node) |
|
423 else: |
|
424 builder.CONTINUE_finish(node) |
|
425 |
|
426 # NO RETURN |
|
427 |
|
428 |
|
429 elif tokenType == "try": |
|
430 node = builder.TRY_build(tokenizer) |
|
431 builder.TRY_setTryBlock(node, Block(tokenizer, staticContext)) |
|
432 |
|
433 while tokenizer.match("catch"): |
|
434 childNode = builder.CATCH_build(tokenizer) |
|
435 tokenizer.mustMatch("left_paren") |
|
436 nextTokenType = tokenizer.get() |
|
437 |
|
438 if nextTokenType == "left_bracket" or nextTokenType == "left_curly": |
|
439 # Destructured catch identifiers. |
|
440 tokenizer.unget() |
|
441 exception = DestructuringExpression(tokenizer, staticContext, True) |
|
442 |
|
443 elif nextTokenType == "identifier": |
|
444 exception = builder.CATCH_wrapException(tokenizer) |
|
445 |
|
446 else: |
|
447 raise SyntaxError("Missing identifier in catch", tokenizer) |
|
448 |
|
449 builder.CATCH_setException(childNode, exception) |
|
450 |
|
451 if tokenizer.match("if"): |
|
452 if staticContext.ecma3OnlyMode: |
|
453 raise SyntaxError("Illegal catch guard", tokenizer) |
|
454 |
|
455 if node.getChildrenLength() > 0 and not node.getUnrelatedChildren()[0].guard: |
|
456 raise SyntaxError("Guarded catch after unguarded", tokenizer) |
|
457 |
|
458 builder.CATCH_setGuard(childNode, Expression(tokenizer, staticContext)) |
|
459 |
|
460 else: |
|
461 builder.CATCH_setGuard(childNode, None) |
|
462 |
|
463 tokenizer.mustMatch("right_paren") |
|
464 |
|
465 builder.CATCH_setBlock(childNode, Block(tokenizer, staticContext)) |
|
466 builder.CATCH_finish(childNode) |
|
467 |
|
468 builder.TRY_addCatch(node, childNode) |
|
469 |
|
470 builder.TRY_finishCatches(node) |
|
471 |
|
472 if tokenizer.match("finally"): |
|
473 builder.TRY_setFinallyBlock(node, Block(tokenizer, staticContext)) |
|
474 |
|
475 if node.getChildrenLength() == 0 and not hasattr(node, "finallyBlock"): |
|
476 raise SyntaxError("Invalid try statement", tokenizer) |
|
477 |
|
478 builder.TRY_finish(node) |
|
479 return node |
|
480 |
|
481 |
|
482 elif tokenType == "catch" or tokenType == "finally": |
|
483 raise SyntaxError(tokens[tokenType] + " without preceding try", tokenizer) |
|
484 |
|
485 |
|
486 elif tokenType == "throw": |
|
487 node = builder.THROW_build(tokenizer) |
|
488 |
|
489 builder.THROW_setException(node, Expression(tokenizer, staticContext)) |
|
490 builder.THROW_finish(node) |
|
491 |
|
492 # NO RETURN |
|
493 |
|
494 |
|
495 elif tokenType == "return": |
|
496 node = returnOrYield(tokenizer, staticContext) |
|
497 |
|
498 # NO RETURN |
|
499 |
|
500 |
|
501 elif tokenType == "with": |
|
502 node = builder.WITH_build(tokenizer) |
|
503 |
|
504 builder.WITH_setObject(node, ParenExpression(tokenizer, staticContext)) |
|
505 builder.WITH_setBody(node, nest(tokenizer, staticContext, node, Statement)) |
|
506 builder.WITH_finish(node) |
|
507 |
|
508 return node |
|
509 |
|
510 |
|
511 elif tokenType == "var" or tokenType == "const": |
|
512 node = Variables(tokenizer, staticContext) |
|
513 |
|
514 # NO RETURN |
|
515 |
|
516 |
|
517 elif tokenType == "let": |
|
518 if tokenizer.peek() == "left_paren": |
|
519 node = LetBlock(tokenizer, staticContext, True) |
|
520 else: |
|
521 node = Variables(tokenizer, staticContext) |
|
522 |
|
523 # NO RETURN |
|
524 |
|
525 |
|
526 elif tokenType == "debugger": |
|
527 node = builder.DEBUGGER_build(tokenizer) |
|
528 |
|
529 # NO RETURN |
|
530 |
|
531 |
|
532 elif tokenType == "newline" or tokenType == "semicolon": |
|
533 node = builder.SEMICOLON_build(tokenizer) |
|
534 |
|
535 builder.SEMICOLON_setExpression(node, None) |
|
536 builder.SEMICOLON_finish(tokenizer) |
|
537 |
|
538 return node |
|
539 |
|
540 |
|
541 else: |
|
542 if tokenType == "identifier": |
|
543 tokenType = tokenizer.peek() |
|
544 |
|
545 # Labeled statement. |
|
546 if tokenType == "colon": |
|
547 label = tokenizer.token.value |
|
548 statementStack = staticContext.statementStack |
|
549 |
|
550 i = len(statementStack)-1 |
|
551 while i >= 0: |
|
552 if getattr(statementStack[i], "label", None) == label: |
|
553 raise SyntaxError("Duplicate label", tokenizer) |
|
554 |
|
555 i -= 1 |
|
556 |
|
557 tokenizer.get() |
|
558 node = builder.LABEL_build(tokenizer) |
|
559 |
|
560 builder.LABEL_setLabel(node, label) |
|
561 builder.LABEL_setStatement(node, nest(tokenizer, staticContext, node, Statement)) |
|
562 builder.LABEL_finish(node) |
|
563 |
|
564 return node |
|
565 |
|
566 # Expression statement. |
|
567 # We unget the current token to parse the expression as a whole. |
|
568 node = builder.SEMICOLON_build(tokenizer) |
|
569 tokenizer.unget() |
|
570 builder.SEMICOLON_setExpression(node, Expression(tokenizer, staticContext)) |
|
571 node.end = node.expression.end |
|
572 builder.SEMICOLON_finish(node) |
|
573 |
|
574 # NO RETURN |
|
575 |
|
576 |
|
577 MagicalSemicolon(tokenizer) |
|
578 return node |
|
579 |
|
580 |
|
581 |
|
582 def MagicalSemicolon(tokenizer): |
|
583 if tokenizer.line == tokenizer.token.line: |
|
584 tokenType = tokenizer.peekOnSameLine() |
|
585 |
|
586 if tokenType != "end" and tokenType != "newline" and tokenType != "semicolon" and tokenType != "right_curly": |
|
587 raise SyntaxError("Missing ; before statement", tokenizer) |
|
588 |
|
589 tokenizer.match("semicolon") |
|
590 |
|
591 |
|
592 |
|
593 def returnOrYield(tokenizer, staticContext): |
|
594 builder = staticContext.builder |
|
595 tokenType = tokenizer.token.type |
|
596 |
|
597 if tokenType == "return": |
|
598 if not staticContext.inFunction: |
|
599 raise SyntaxError("Return not in function", tokenizer) |
|
600 |
|
601 node = builder.RETURN_build(tokenizer) |
|
602 |
|
603 else: |
|
604 if not staticContext.inFunction: |
|
605 raise SyntaxError("Yield not in function", tokenizer) |
|
606 |
|
607 staticContext.isGenerator = True |
|
608 node = builder.YIELD_build(tokenizer) |
|
609 |
|
610 nextTokenType = tokenizer.peek(True) |
|
611 if nextTokenType != "end" and nextTokenType != "newline" and nextTokenType != "semicolon" and nextTokenType != "right_curly" and (tokenType != "yield" or (nextTokenType != tokenType and nextTokenType != "right_bracket" and nextTokenType != "right_paren" and nextTokenType != "colon" and nextTokenType != "comma")): |
|
612 if tokenType == "return": |
|
613 builder.RETURN_setValue(node, Expression(tokenizer, staticContext)) |
|
614 staticContext.hasReturnWithValue = True |
|
615 else: |
|
616 builder.YIELD_setValue(node, AssignExpression(tokenizer, staticContext)) |
|
617 |
|
618 elif tokenType == "return": |
|
619 staticContext.hasEmptyReturn = True |
|
620 |
|
621 # Disallow return v; in generator. |
|
622 if staticContext.hasReturnWithValue and staticContext.isGenerator: |
|
623 raise SyntaxError("Generator returns a value", tokenizer) |
|
624 |
|
625 if tokenType == "return": |
|
626 builder.RETURN_finish(node) |
|
627 else: |
|
628 builder.YIELD_finish(node) |
|
629 |
|
630 return node |
|
631 |
|
632 |
|
633 |
|
634 def FunctionDefinition(tokenizer, staticContext, requireName, functionForm): |
|
635 builder = staticContext.builder |
|
636 functionNode = builder.FUNCTION_build(tokenizer) |
|
637 |
|
638 if tokenizer.match("identifier"): |
|
639 builder.FUNCTION_setName(functionNode, tokenizer.token.value) |
|
640 elif requireName: |
|
641 raise SyntaxError("Missing def identifier", tokenizer) |
|
642 |
|
643 tokenizer.mustMatch("left_paren") |
|
644 |
|
645 if not tokenizer.match("right_paren"): |
|
646 builder.FUNCTION_initParams(functionNode, tokenizer) |
|
647 prevParamNode = None |
|
648 while True: |
|
649 tokenType = tokenizer.get() |
|
650 if tokenType == "left_bracket" or tokenType == "left_curly": |
|
651 # Destructured formal parameters. |
|
652 tokenizer.unget() |
|
653 paramNode = DestructuringExpression(tokenizer, staticContext) |
|
654 |
|
655 elif tokenType == "identifier": |
|
656 paramNode = builder.FUNCTION_wrapParam(tokenizer) |
|
657 |
|
658 else: |
|
659 raise SyntaxError("Missing formal parameter", tokenizer) |
|
660 |
|
661 builder.FUNCTION_addParam(functionNode, tokenizer, paramNode) |
|
662 builder.COMMENTS_add(paramNode, prevParamNode, tokenizer.getComments()) |
|
663 |
|
664 if not tokenizer.match("comma"): |
|
665 break |
|
666 |
|
667 prevParamNode = paramNode |
|
668 |
|
669 tokenizer.mustMatch("right_paren") |
|
670 |
|
671 # Do we have an expression closure or a normal body? |
|
672 tokenType = tokenizer.get() |
|
673 if tokenType != "left_curly": |
|
674 builder.FUNCTION_setExpressionClosure(functionNode, True) |
|
675 tokenizer.unget() |
|
676 |
|
677 childContext = StaticContext(True, builder) |
|
678 |
|
679 if staticContext.inFunction: |
|
680 # Inner functions don't reset block numbering, only functions at |
|
681 # the top level of the program do. |
|
682 childContext.blockId = staticContext.blockId |
|
683 |
|
684 if tokenType != "left_curly": |
|
685 builder.FUNCTION_setBody(functionNode, AssignExpression(tokenizer, staticContext)) |
|
686 if staticContext.isGenerator: |
|
687 raise SyntaxError("Generator returns a value", tokenizer) |
|
688 |
|
689 else: |
|
690 builder.FUNCTION_hoistVars(childContext.blockId) |
|
691 builder.FUNCTION_setBody(functionNode, Script(tokenizer, childContext)) |
|
692 |
|
693 if tokenType == "left_curly": |
|
694 tokenizer.mustMatch("right_curly") |
|
695 |
|
696 functionNode.end = tokenizer.token.end |
|
697 functionNode.functionForm = functionForm |
|
698 |
|
699 builder.COMMENTS_add(functionNode.body, functionNode.body, tokenizer.getComments()) |
|
700 builder.FUNCTION_finish(functionNode, staticContext) |
|
701 |
|
702 return functionNode |
|
703 |
|
704 |
|
705 |
|
706 def Variables(tokenizer, staticContext, letBlock=None): |
|
707 """Parses a comma-separated list of var declarations (and maybe initializations).""" |
|
708 |
|
709 builder = staticContext.builder |
|
710 if tokenizer.token.type == "var": |
|
711 build = builder.VAR_build |
|
712 addDecl = builder.VAR_addDecl |
|
713 finish = builder.VAR_finish |
|
714 childContext = staticContext |
|
715 |
|
716 elif tokenizer.token.type == "const": |
|
717 build = builder.CONST_build |
|
718 addDecl = builder.CONST_addDecl |
|
719 finish = builder.CONST_finish |
|
720 childContext = staticContext |
|
721 |
|
722 elif tokenizer.token.type == "let" or tokenizer.token.type == "left_paren": |
|
723 build = builder.LET_build |
|
724 addDecl = builder.LET_addDecl |
|
725 finish = builder.LET_finish |
|
726 |
|
727 if not letBlock: |
|
728 statementStack = staticContext.statementStack |
|
729 i = len(statementStack) - 1 |
|
730 |
|
731 # a BLOCK *must* be found. |
|
732 while statementStack[i].type != "block": |
|
733 i -= 1 |
|
734 |
|
735 # Lets at the def toplevel are just vars, at least in SpiderMonkey. |
|
736 if i == 0: |
|
737 build = builder.VAR_build |
|
738 addDecl = builder.VAR_addDecl |
|
739 finish = builder.VAR_finish |
|
740 childContext = staticContext |
|
741 |
|
742 else: |
|
743 childContext = statementStack[i] |
|
744 |
|
745 else: |
|
746 childContext = letBlock |
|
747 |
|
748 node = build(tokenizer) |
|
749 |
|
750 while True: |
|
751 tokenType = tokenizer.get() |
|
752 |
|
753 # Done in Python port! |
|
754 # FIXME Should have a special DECLARATION node instead of overloading |
|
755 # IDENTIFIER to mean both identifier declarations and destructured |
|
756 # declarations. |
|
757 childNode = builder.DECL_build(tokenizer) |
|
758 |
|
759 if tokenType == "left_bracket" or tokenType == "left_curly": |
|
760 # Pass in childContext if we need to add each pattern matched into |
|
761 # its variables, else pass in staticContext. |
|
762 # Need to unget to parse the full destructured expression. |
|
763 tokenizer.unget() |
|
764 builder.DECL_setNames(childNode, DestructuringExpression(tokenizer, staticContext, True, childContext)) |
|
765 |
|
766 if staticContext.inForLoopInit and tokenizer.peek() == "in": |
|
767 addDecl(node, childNode, childContext) |
|
768 if tokenizer.match("comma"): |
|
769 continue |
|
770 else: |
|
771 break |
|
772 |
|
773 tokenizer.mustMatch("assign") |
|
774 if tokenizer.token.assignOp: |
|
775 raise SyntaxError("Invalid variable initialization", tokenizer) |
|
776 |
|
777 # Parse the init as a normal assignment. |
|
778 builder.DECL_setInitializer(childNode, AssignExpression(tokenizer, staticContext)) |
|
779 builder.DECL_finish(childNode) |
|
780 addDecl(node, childNode, childContext) |
|
781 |
|
782 # Copy over names for variable list |
|
783 # for nameNode in childNode.names: |
|
784 # childContext.variables.add(nameNode.value) |
|
785 |
|
786 if tokenizer.match("comma"): |
|
787 continue |
|
788 else: |
|
789 break |
|
790 |
|
791 if tokenType != "identifier": |
|
792 raise SyntaxError("Missing variable name", tokenizer) |
|
793 |
|
794 builder.DECL_setName(childNode, tokenizer.token.value) |
|
795 builder.DECL_setReadOnly(childNode, node.type == "const") |
|
796 addDecl(node, childNode, childContext) |
|
797 |
|
798 if tokenizer.match("assign"): |
|
799 if tokenizer.token.assignOp: |
|
800 raise SyntaxError("Invalid variable initialization", tokenizer) |
|
801 |
|
802 initializerNode = AssignExpression(tokenizer, staticContext) |
|
803 builder.DECL_setInitializer(childNode, initializerNode) |
|
804 |
|
805 builder.DECL_finish(childNode) |
|
806 |
|
807 # If we directly use the node in "let" constructs |
|
808 # if not hasattr(childContext, "variables"): |
|
809 # childContext.variables = set() |
|
810 |
|
811 # childContext.variables.add(childNode.name) |
|
812 |
|
813 if not tokenizer.match("comma"): |
|
814 break |
|
815 |
|
816 finish(node) |
|
817 return node |
|
818 |
|
819 |
|
820 |
|
821 def LetBlock(tokenizer, staticContext, isStatement): |
|
822 """Does not handle let inside of for loop init.""" |
|
823 builder = staticContext.builder |
|
824 |
|
825 # tokenizer.token.type must be "let" |
|
826 node = builder.LETBLOCK_build(tokenizer) |
|
827 tokenizer.mustMatch("left_paren") |
|
828 builder.LETBLOCK_setVariables(node, Variables(tokenizer, staticContext, node)) |
|
829 tokenizer.mustMatch("right_paren") |
|
830 |
|
831 if isStatement and tokenizer.peek() != "left_curly": |
|
832 # If this is really an expression in let statement guise, then we |
|
833 # need to wrap the "let_block" node in a "semicolon" node so that we pop |
|
834 # the return value of the expression. |
|
835 childNode = builder.SEMICOLON_build(tokenizer) |
|
836 builder.SEMICOLON_setExpression(childNode, node) |
|
837 builder.SEMICOLON_finish(childNode) |
|
838 isStatement = False |
|
839 |
|
840 if isStatement: |
|
841 childNode = Block(tokenizer, staticContext) |
|
842 builder.LETBLOCK_setBlock(node, childNode) |
|
843 |
|
844 else: |
|
845 childNode = AssignExpression(tokenizer, staticContext) |
|
846 builder.LETBLOCK_setExpression(node, childNode) |
|
847 |
|
848 builder.LETBLOCK_finish(node) |
|
849 return node |
|
850 |
|
851 |
|
852 def checkDestructuring(tokenizer, staticContext, node, simpleNamesOnly=None, data=None): |
|
853 if node.type == "array_comp": |
|
854 raise SyntaxError("Invalid array comprehension left-hand side", tokenizer) |
|
855 |
|
856 if node.type != "array_init" and node.type != "object_init": |
|
857 return |
|
858 |
|
859 builder = staticContext.builder |
|
860 |
|
861 for child in node: |
|
862 if child == None: |
|
863 continue |
|
864 |
|
865 if child.type == "property_init": |
|
866 lhs = child[0] |
|
867 rhs = child[1] |
|
868 else: |
|
869 lhs = None |
|
870 rhs = None |
|
871 |
|
872 |
|
873 if rhs and (rhs.type == "array_init" or rhs.type == "object_init"): |
|
874 checkDestructuring(tokenizer, staticContext, rhs, simpleNamesOnly, data) |
|
875 |
|
876 if lhs and simpleNamesOnly: |
|
877 # In declarations, lhs must be simple names |
|
878 if lhs.type != "identifier": |
|
879 raise SyntaxError("Missing name in pattern", tokenizer) |
|
880 |
|
881 elif data: |
|
882 childNode = builder.DECL_build(tokenizer) |
|
883 builder.DECL_setName(childNode, lhs.value) |
|
884 |
|
885 # Don't need to set initializer because it's just for |
|
886 # hoisting anyways. |
|
887 builder.DECL_finish(childNode) |
|
888 |
|
889 # Each pattern needs to be added to variables. |
|
890 # data.variables.add(childNode.name) |
|
891 |
|
892 |
|
893 # JavaScript 1.7 |
|
894 def DestructuringExpression(tokenizer, staticContext, simpleNamesOnly=None, data=None): |
|
895 node = PrimaryExpression(tokenizer, staticContext) |
|
896 checkDestructuring(tokenizer, staticContext, node, simpleNamesOnly, data) |
|
897 |
|
898 return node |
|
899 |
|
900 |
|
901 # JavsScript 1.7 |
|
902 def GeneratorExpression(tokenizer, staticContext, expression): |
|
903 builder = staticContext.builder |
|
904 node = builder.GENERATOR_build(tokenizer) |
|
905 |
|
906 builder.GENERATOR_setExpression(node, expression) |
|
907 builder.GENERATOR_setTail(node, comprehensionTail(tokenizer, staticContext)) |
|
908 builder.GENERATOR_finish(node) |
|
909 |
|
910 return node |
|
911 |
|
912 |
|
913 # JavaScript 1.7 Comprehensions Tails (Generators / Arrays) |
|
914 def comprehensionTail(tokenizer, staticContext): |
|
915 builder = staticContext.builder |
|
916 |
|
917 # tokenizer.token.type must be "for" |
|
918 body = builder.COMPTAIL_build(tokenizer) |
|
919 |
|
920 while True: |
|
921 node = builder.FOR_build(tokenizer) |
|
922 |
|
923 # Comprehension tails are always for..in loops. |
|
924 builder.FOR_rebuildForIn(node) |
|
925 if tokenizer.match("identifier"): |
|
926 # But sometimes they're for each..in. |
|
927 if tokenizer.token.value == "each": |
|
928 builder.FOR_rebuildForEach(node) |
|
929 else: |
|
930 tokenizer.unget() |
|
931 |
|
932 tokenizer.mustMatch("left_paren") |
|
933 |
|
934 tokenType = tokenizer.get() |
|
935 if tokenType == "left_bracket" or tokenType == "left_curly": |
|
936 tokenizer.unget() |
|
937 # Destructured left side of for in comprehension tails. |
|
938 builder.FOR_setIterator(node, DestructuringExpression(tokenizer, staticContext)) |
|
939 |
|
940 elif tokenType == "identifier": |
|
941 # Removed variable/declaration substructure in Python port. |
|
942 # Variable declarations are not allowed here. So why process them in such a way? |
|
943 |
|
944 # declaration = builder.DECL_build(tokenizer) |
|
945 # builder.DECL_setName(declaration, tokenizer.token.value) |
|
946 # builder.DECL_finish(declaration) |
|
947 # childNode = builder.VAR_build(tokenizer) |
|
948 # builder.VAR_addDecl(childNode, declaration) |
|
949 # builder.VAR_finish(childNode) |
|
950 # builder.FOR_setIterator(node, declaration) |
|
951 |
|
952 # Don't add to variables since the semantics of comprehensions is |
|
953 # such that the variables are in their own def when desugared. |
|
954 |
|
955 identifier = builder.PRIMARY_build(tokenizer, "identifier") |
|
956 builder.FOR_setIterator(node, identifier) |
|
957 |
|
958 else: |
|
959 raise SyntaxError("Missing identifier", tokenizer) |
|
960 |
|
961 tokenizer.mustMatch("in") |
|
962 builder.FOR_setObject(node, Expression(tokenizer, staticContext)) |
|
963 tokenizer.mustMatch("right_paren") |
|
964 builder.COMPTAIL_addFor(body, node) |
|
965 |
|
966 if not tokenizer.match("for"): |
|
967 break |
|
968 |
|
969 # Optional guard. |
|
970 if tokenizer.match("if"): |
|
971 builder.COMPTAIL_setGuard(body, ParenExpression(tokenizer, staticContext)) |
|
972 |
|
973 builder.COMPTAIL_finish(body) |
|
974 |
|
975 return body |
|
976 |
|
977 |
|
978 def ParenExpression(tokenizer, staticContext): |
|
979 tokenizer.mustMatch("left_paren") |
|
980 |
|
981 # Always accept the 'in' operator in a parenthesized expression, |
|
982 # where it's unambiguous, even if we might be parsing the init of a |
|
983 # for statement. |
|
984 oldLoopInit = staticContext.inForLoopInit |
|
985 staticContext.inForLoopInit = False |
|
986 node = Expression(tokenizer, staticContext) |
|
987 staticContext.inForLoopInit = oldLoopInit |
|
988 |
|
989 err = "expression must be parenthesized" |
|
990 if tokenizer.match("for"): |
|
991 if node.type == "yield" and not node.parenthesized: |
|
992 raise SyntaxError("Yield " + err, tokenizer) |
|
993 |
|
994 if node.type == "comma" and not node.parenthesized: |
|
995 raise SyntaxError("Generator " + err, tokenizer) |
|
996 |
|
997 node = GeneratorExpression(tokenizer, staticContext, node) |
|
998 |
|
999 tokenizer.mustMatch("right_paren") |
|
1000 |
|
1001 return node |
|
1002 |
|
1003 |
|
1004 def Expression(tokenizer, staticContext): |
|
1005 """Top-down expression parser matched against SpiderMonkey.""" |
|
1006 builder = staticContext.builder |
|
1007 node = AssignExpression(tokenizer, staticContext) |
|
1008 |
|
1009 if tokenizer.match("comma"): |
|
1010 childNode = builder.COMMA_build(tokenizer) |
|
1011 builder.COMMA_addOperand(childNode, node) |
|
1012 node = childNode |
|
1013 while True: |
|
1014 childNode = node[len(node)-1] |
|
1015 if childNode.type == "yield" and not childNode.parenthesized: |
|
1016 raise SyntaxError("Yield expression must be parenthesized", tokenizer) |
|
1017 builder.COMMA_addOperand(node, AssignExpression(tokenizer, staticContext)) |
|
1018 |
|
1019 if not tokenizer.match("comma"): |
|
1020 break |
|
1021 |
|
1022 builder.COMMA_finish(node) |
|
1023 |
|
1024 return node |
|
1025 |
|
1026 |
|
1027 def AssignExpression(tokenizer, staticContext): |
|
1028 builder = staticContext.builder |
|
1029 |
|
1030 # Have to treat yield like an operand because it could be the leftmost |
|
1031 # operand of the expression. |
|
1032 if tokenizer.match("yield", True): |
|
1033 return returnOrYield(tokenizer, staticContext) |
|
1034 |
|
1035 comments = tokenizer.getComments() |
|
1036 node = builder.ASSIGN_build(tokenizer) |
|
1037 lhs = ConditionalExpression(tokenizer, staticContext) |
|
1038 builder.COMMENTS_add(lhs, None, comments) |
|
1039 |
|
1040 if not tokenizer.match("assign"): |
|
1041 builder.ASSIGN_finish(node) |
|
1042 return lhs |
|
1043 |
|
1044 if lhs.type == "object_init" or lhs.type == "array_init": |
|
1045 checkDestructuring(tokenizer, staticContext, lhs) |
|
1046 elif lhs.type == "identifier" or lhs.type == "dot" or lhs.type == "index" or lhs.type == "call": |
|
1047 pass |
|
1048 else: |
|
1049 raise SyntaxError("Bad left-hand side of assignment", tokenizer) |
|
1050 |
|
1051 builder.ASSIGN_setAssignOp(node, tokenizer.token.assignOp) |
|
1052 builder.ASSIGN_addOperand(node, lhs) |
|
1053 builder.ASSIGN_addOperand(node, AssignExpression(tokenizer, staticContext)) |
|
1054 builder.ASSIGN_finish(node) |
|
1055 |
|
1056 return node |
|
1057 |
|
1058 |
|
1059 def ConditionalExpression(tokenizer, staticContext): |
|
1060 builder = staticContext.builder |
|
1061 node = OrExpression(tokenizer, staticContext) |
|
1062 |
|
1063 if tokenizer.match("hook"): |
|
1064 childNode = node |
|
1065 node = builder.HOOK_build(tokenizer) |
|
1066 builder.HOOK_setCondition(node, childNode) |
|
1067 |
|
1068 # Always accept the 'in' operator in the middle clause of a ternary, |
|
1069 # where it's unambiguous, even if we might be parsing the init of a |
|
1070 # for statement. |
|
1071 oldLoopInit = staticContext.inForLoopInit |
|
1072 staticContext.inForLoopInit = False |
|
1073 builder.HOOK_setThenPart(node, AssignExpression(tokenizer, staticContext)) |
|
1074 staticContext.inForLoopInit = oldLoopInit |
|
1075 |
|
1076 if not tokenizer.match("colon"): |
|
1077 raise SyntaxError("Missing : after ?", tokenizer) |
|
1078 |
|
1079 builder.HOOK_setElsePart(node, AssignExpression(tokenizer, staticContext)) |
|
1080 builder.HOOK_finish(node) |
|
1081 |
|
1082 return node |
|
1083 |
|
1084 |
|
1085 def OrExpression(tokenizer, staticContext): |
|
1086 builder = staticContext.builder |
|
1087 node = AndExpression(tokenizer, staticContext) |
|
1088 |
|
1089 while tokenizer.match("or"): |
|
1090 childNode = builder.OR_build(tokenizer) |
|
1091 builder.OR_addOperand(childNode, node) |
|
1092 builder.OR_addOperand(childNode, AndExpression(tokenizer, staticContext)) |
|
1093 builder.OR_finish(childNode) |
|
1094 node = childNode |
|
1095 |
|
1096 return node |
|
1097 |
|
1098 |
|
1099 def AndExpression(tokenizer, staticContext): |
|
1100 builder = staticContext.builder |
|
1101 node = BitwiseOrExpression(tokenizer, staticContext) |
|
1102 |
|
1103 while tokenizer.match("and"): |
|
1104 childNode = builder.AND_build(tokenizer) |
|
1105 builder.AND_addOperand(childNode, node) |
|
1106 builder.AND_addOperand(childNode, BitwiseOrExpression(tokenizer, staticContext)) |
|
1107 builder.AND_finish(childNode) |
|
1108 node = childNode |
|
1109 |
|
1110 return node |
|
1111 |
|
1112 |
|
1113 def BitwiseOrExpression(tokenizer, staticContext): |
|
1114 builder = staticContext.builder |
|
1115 node = BitwiseXorExpression(tokenizer, staticContext) |
|
1116 |
|
1117 while tokenizer.match("bitwise_or"): |
|
1118 childNode = builder.BITWISEOR_build(tokenizer) |
|
1119 builder.BITWISEOR_addOperand(childNode, node) |
|
1120 builder.BITWISEOR_addOperand(childNode, BitwiseXorExpression(tokenizer, staticContext)) |
|
1121 builder.BITWISEOR_finish(childNode) |
|
1122 node = childNode |
|
1123 |
|
1124 return node |
|
1125 |
|
1126 |
|
1127 def BitwiseXorExpression(tokenizer, staticContext): |
|
1128 builder = staticContext.builder |
|
1129 node = BitwiseAndExpression(tokenizer, staticContext) |
|
1130 |
|
1131 while tokenizer.match("bitwise_xor"): |
|
1132 childNode = builder.BITWISEXOR_build(tokenizer) |
|
1133 builder.BITWISEXOR_addOperand(childNode, node) |
|
1134 builder.BITWISEXOR_addOperand(childNode, BitwiseAndExpression(tokenizer, staticContext)) |
|
1135 builder.BITWISEXOR_finish(childNode) |
|
1136 node = childNode |
|
1137 |
|
1138 return node |
|
1139 |
|
1140 |
|
1141 def BitwiseAndExpression(tokenizer, staticContext): |
|
1142 builder = staticContext.builder |
|
1143 node = EqualityExpression(tokenizer, staticContext) |
|
1144 |
|
1145 while tokenizer.match("bitwise_and"): |
|
1146 childNode = builder.BITWISEAND_build(tokenizer) |
|
1147 builder.BITWISEAND_addOperand(childNode, node) |
|
1148 builder.BITWISEAND_addOperand(childNode, EqualityExpression(tokenizer, staticContext)) |
|
1149 builder.BITWISEAND_finish(childNode) |
|
1150 node = childNode |
|
1151 |
|
1152 return node |
|
1153 |
|
1154 |
|
1155 def EqualityExpression(tokenizer, staticContext): |
|
1156 builder = staticContext.builder |
|
1157 node = RelationalExpression(tokenizer, staticContext) |
|
1158 |
|
1159 while tokenizer.match("eq") or tokenizer.match("ne") or tokenizer.match("strict_eq") or tokenizer.match("strict_ne"): |
|
1160 childNode = builder.EQUALITY_build(tokenizer) |
|
1161 builder.EQUALITY_addOperand(childNode, node) |
|
1162 builder.EQUALITY_addOperand(childNode, RelationalExpression(tokenizer, staticContext)) |
|
1163 builder.EQUALITY_finish(childNode) |
|
1164 node = childNode |
|
1165 |
|
1166 return node |
|
1167 |
|
1168 |
|
1169 def RelationalExpression(tokenizer, staticContext): |
|
1170 builder = staticContext.builder |
|
1171 oldLoopInit = staticContext.inForLoopInit |
|
1172 |
|
1173 # Uses of the in operator in shiftExprs are always unambiguous, |
|
1174 # so unset the flag that prohibits recognizing it. |
|
1175 staticContext.inForLoopInit = False |
|
1176 node = ShiftExpression(tokenizer, staticContext) |
|
1177 |
|
1178 while tokenizer.match("lt") or tokenizer.match("le") or tokenizer.match("ge") or tokenizer.match("gt") or (oldLoopInit == False and tokenizer.match("in")) or tokenizer.match("instanceof"): |
|
1179 childNode = builder.RELATIONAL_build(tokenizer) |
|
1180 builder.RELATIONAL_addOperand(childNode, node) |
|
1181 builder.RELATIONAL_addOperand(childNode, ShiftExpression(tokenizer, staticContext)) |
|
1182 builder.RELATIONAL_finish(childNode) |
|
1183 node = childNode |
|
1184 |
|
1185 staticContext.inForLoopInit = oldLoopInit |
|
1186 |
|
1187 return node |
|
1188 |
|
1189 |
|
1190 def ShiftExpression(tokenizer, staticContext): |
|
1191 builder = staticContext.builder |
|
1192 node = AddExpression(tokenizer, staticContext) |
|
1193 |
|
1194 while tokenizer.match("lsh") or tokenizer.match("rsh") or tokenizer.match("ursh"): |
|
1195 childNode = builder.SHIFT_build(tokenizer) |
|
1196 builder.SHIFT_addOperand(childNode, node) |
|
1197 builder.SHIFT_addOperand(childNode, AddExpression(tokenizer, staticContext)) |
|
1198 builder.SHIFT_finish(childNode) |
|
1199 node = childNode |
|
1200 |
|
1201 return node |
|
1202 |
|
1203 |
|
1204 def AddExpression(tokenizer, staticContext): |
|
1205 builder = staticContext.builder |
|
1206 node = MultiplyExpression(tokenizer, staticContext) |
|
1207 |
|
1208 while tokenizer.match("plus") or tokenizer.match("minus"): |
|
1209 childNode = builder.ADD_build(tokenizer) |
|
1210 builder.ADD_addOperand(childNode, node) |
|
1211 builder.ADD_addOperand(childNode, MultiplyExpression(tokenizer, staticContext)) |
|
1212 builder.ADD_finish(childNode) |
|
1213 node = childNode |
|
1214 |
|
1215 return node |
|
1216 |
|
1217 |
|
1218 def MultiplyExpression(tokenizer, staticContext): |
|
1219 builder = staticContext.builder |
|
1220 node = UnaryExpression(tokenizer, staticContext) |
|
1221 |
|
1222 while tokenizer.match("mul") or tokenizer.match("div") or tokenizer.match("mod"): |
|
1223 childNode = builder.MULTIPLY_build(tokenizer) |
|
1224 builder.MULTIPLY_addOperand(childNode, node) |
|
1225 builder.MULTIPLY_addOperand(childNode, UnaryExpression(tokenizer, staticContext)) |
|
1226 builder.MULTIPLY_finish(childNode) |
|
1227 node = childNode |
|
1228 |
|
1229 return node |
|
1230 |
|
1231 |
|
1232 def UnaryExpression(tokenizer, staticContext): |
|
1233 builder = staticContext.builder |
|
1234 tokenType = tokenizer.get(True) |
|
1235 |
|
1236 if tokenType in ["delete", "void", "typeof", "not", "bitwise_not", "plus", "minus"]: |
|
1237 node = builder.UNARY_build(tokenizer) |
|
1238 builder.UNARY_addOperand(node, UnaryExpression(tokenizer, staticContext)) |
|
1239 |
|
1240 elif tokenType == "increment" or tokenType == "decrement": |
|
1241 # Prefix increment/decrement. |
|
1242 node = builder.UNARY_build(tokenizer) |
|
1243 builder.UNARY_addOperand(node, MemberExpression(tokenizer, staticContext, True)) |
|
1244 |
|
1245 else: |
|
1246 tokenizer.unget() |
|
1247 node = MemberExpression(tokenizer, staticContext, True) |
|
1248 |
|
1249 # Don't look across a newline boundary for a postfix {in,de}crement. |
|
1250 if tokenizer.tokens[(tokenizer.tokenIndex + tokenizer.lookahead - 1) & 3].line == tokenizer.line: |
|
1251 if tokenizer.match("increment") or tokenizer.match("decrement"): |
|
1252 childNode = builder.UNARY_build(tokenizer) |
|
1253 builder.UNARY_setPostfix(childNode) |
|
1254 builder.UNARY_finish(node) |
|
1255 builder.UNARY_addOperand(childNode, node) |
|
1256 node = childNode |
|
1257 |
|
1258 builder.UNARY_finish(node) |
|
1259 return node |
|
1260 |
|
1261 |
|
1262 def MemberExpression(tokenizer, staticContext, allowCallSyntax): |
|
1263 builder = staticContext.builder |
|
1264 |
|
1265 if tokenizer.match("new"): |
|
1266 node = builder.MEMBER_build(tokenizer) |
|
1267 builder.MEMBER_addOperand(node, MemberExpression(tokenizer, staticContext, False)) |
|
1268 |
|
1269 if tokenizer.match("left_paren"): |
|
1270 builder.MEMBER_rebuildNewWithArgs(node) |
|
1271 builder.MEMBER_addOperand(node, ArgumentList(tokenizer, staticContext)) |
|
1272 |
|
1273 builder.MEMBER_finish(node) |
|
1274 |
|
1275 else: |
|
1276 node = PrimaryExpression(tokenizer, staticContext) |
|
1277 |
|
1278 while True: |
|
1279 tokenType = tokenizer.get() |
|
1280 if tokenType == "end": |
|
1281 break |
|
1282 |
|
1283 if tokenType == "dot": |
|
1284 childNode = builder.MEMBER_build(tokenizer) |
|
1285 builder.MEMBER_addOperand(childNode, node) |
|
1286 tokenizer.mustMatch("identifier") |
|
1287 builder.MEMBER_addOperand(childNode, builder.MEMBER_build(tokenizer)) |
|
1288 |
|
1289 elif tokenType == "left_bracket": |
|
1290 childNode = builder.MEMBER_build(tokenizer, "index") |
|
1291 builder.MEMBER_addOperand(childNode, node) |
|
1292 builder.MEMBER_addOperand(childNode, Expression(tokenizer, staticContext)) |
|
1293 tokenizer.mustMatch("right_bracket") |
|
1294 |
|
1295 elif tokenType == "left_paren" and allowCallSyntax: |
|
1296 childNode = builder.MEMBER_build(tokenizer, "call") |
|
1297 builder.MEMBER_addOperand(childNode, node) |
|
1298 builder.MEMBER_addOperand(childNode, ArgumentList(tokenizer, staticContext)) |
|
1299 |
|
1300 else: |
|
1301 tokenizer.unget() |
|
1302 return node |
|
1303 |
|
1304 builder.MEMBER_finish(childNode) |
|
1305 node = childNode |
|
1306 |
|
1307 return node |
|
1308 |
|
1309 |
|
1310 def ArgumentList(tokenizer, staticContext): |
|
1311 builder = staticContext.builder |
|
1312 node = builder.LIST_build(tokenizer) |
|
1313 |
|
1314 if tokenizer.match("right_paren", True): |
|
1315 return node |
|
1316 |
|
1317 while True: |
|
1318 childNode = AssignExpression(tokenizer, staticContext) |
|
1319 if childNode.type == "yield" and not childNode.parenthesized and tokenizer.peek() == "comma": |
|
1320 raise SyntaxError("Yield expression must be parenthesized", tokenizer) |
|
1321 |
|
1322 if tokenizer.match("for"): |
|
1323 childNode = GeneratorExpression(tokenizer, staticContext, childNode) |
|
1324 if len(node) > 1 or tokenizer.peek(True) == "comma": |
|
1325 raise SyntaxError("Generator expression must be parenthesized", tokenizer) |
|
1326 |
|
1327 builder.LIST_addOperand(node, childNode) |
|
1328 if not tokenizer.match("comma"): |
|
1329 break |
|
1330 |
|
1331 tokenizer.mustMatch("right_paren") |
|
1332 builder.LIST_finish(node) |
|
1333 |
|
1334 return node |
|
1335 |
|
1336 |
|
1337 def PrimaryExpression(tokenizer, staticContext): |
|
1338 builder = staticContext.builder |
|
1339 tokenType = tokenizer.get(True) |
|
1340 |
|
1341 if tokenType == "function": |
|
1342 node = FunctionDefinition(tokenizer, staticContext, False, "expressed_form") |
|
1343 |
|
1344 elif tokenType == "left_bracket": |
|
1345 node = builder.ARRAYINIT_build(tokenizer) |
|
1346 while True: |
|
1347 tokenType = tokenizer.peek(True) |
|
1348 if tokenType == "right_bracket": |
|
1349 break |
|
1350 |
|
1351 if tokenType == "comma": |
|
1352 tokenizer.get() |
|
1353 builder.ARRAYINIT_addElement(node, None) |
|
1354 continue |
|
1355 |
|
1356 builder.ARRAYINIT_addElement(node, AssignExpression(tokenizer, staticContext)) |
|
1357 |
|
1358 if tokenType != "comma" and not tokenizer.match("comma"): |
|
1359 break |
|
1360 |
|
1361 # If we matched exactly one element and got a "for", we have an |
|
1362 # array comprehension. |
|
1363 if len(node) == 1 and tokenizer.match("for"): |
|
1364 childNode = builder.ARRAYCOMP_build(tokenizer) |
|
1365 builder.ARRAYCOMP_setExpression(childNode, node[0]) |
|
1366 builder.ARRAYCOMP_setTail(childNode, comprehensionTail(tokenizer, staticContext)) |
|
1367 node = childNode |
|
1368 |
|
1369 builder.COMMENTS_add(node, node, tokenizer.getComments()) |
|
1370 tokenizer.mustMatch("right_bracket") |
|
1371 builder.PRIMARY_finish(node) |
|
1372 |
|
1373 elif tokenType == "left_curly": |
|
1374 node = builder.OBJECTINIT_build(tokenizer) |
|
1375 |
|
1376 if not tokenizer.match("right_curly"): |
|
1377 while True: |
|
1378 tokenType = tokenizer.get() |
|
1379 tokenValue = getattr(tokenizer.token, "value", None) |
|
1380 comments = tokenizer.getComments() |
|
1381 |
|
1382 if tokenValue in ("get", "set") and tokenizer.peek() == "identifier": |
|
1383 if staticContext.ecma3OnlyMode: |
|
1384 raise SyntaxError("Illegal property accessor", tokenizer) |
|
1385 |
|
1386 fd = FunctionDefinition(tokenizer, staticContext, True, "expressed_form") |
|
1387 builder.OBJECTINIT_addProperty(node, fd) |
|
1388 |
|
1389 else: |
|
1390 if tokenType == "identifier" or tokenType == "number" or tokenType == "string": |
|
1391 id = builder.PRIMARY_build(tokenizer, "identifier") |
|
1392 builder.PRIMARY_finish(id) |
|
1393 |
|
1394 elif tokenType == "right_curly": |
|
1395 if staticContext.ecma3OnlyMode: |
|
1396 raise SyntaxError("Illegal trailing ,", tokenizer) |
|
1397 |
|
1398 tokenizer.unget() |
|
1399 break |
|
1400 |
|
1401 else: |
|
1402 if tokenValue in jasy.js.tokenize.Lang.keywords: |
|
1403 id = builder.PRIMARY_build(tokenizer, "identifier") |
|
1404 builder.PRIMARY_finish(id) |
|
1405 else: |
|
1406 print("Value is '%s'" % tokenValue) |
|
1407 raise SyntaxError("Invalid property name", tokenizer) |
|
1408 |
|
1409 if tokenizer.match("colon"): |
|
1410 childNode = builder.PROPERTYINIT_build(tokenizer) |
|
1411 builder.COMMENTS_add(childNode, node, comments) |
|
1412 builder.PROPERTYINIT_addOperand(childNode, id) |
|
1413 builder.PROPERTYINIT_addOperand(childNode, AssignExpression(tokenizer, staticContext)) |
|
1414 builder.PROPERTYINIT_finish(childNode) |
|
1415 builder.OBJECTINIT_addProperty(node, childNode) |
|
1416 |
|
1417 else: |
|
1418 # Support, e.g., |var {staticContext, y} = o| as destructuring shorthand |
|
1419 # for |var {staticContext: staticContext, y: y} = o|, per proposed JS2/ES4 for JS1.8. |
|
1420 if tokenizer.peek() != "comma" and tokenizer.peek() != "right_curly": |
|
1421 raise SyntaxError("Missing : after property", tokenizer) |
|
1422 builder.OBJECTINIT_addProperty(node, id) |
|
1423 |
|
1424 if not tokenizer.match("comma"): |
|
1425 break |
|
1426 |
|
1427 builder.COMMENTS_add(node, node, tokenizer.getComments()) |
|
1428 tokenizer.mustMatch("right_curly") |
|
1429 |
|
1430 builder.OBJECTINIT_finish(node) |
|
1431 |
|
1432 elif tokenType == "left_paren": |
|
1433 # ParenExpression does its own matching on parentheses, so we need to unget. |
|
1434 tokenizer.unget() |
|
1435 node = ParenExpression(tokenizer, staticContext) |
|
1436 node.parenthesized = True |
|
1437 |
|
1438 elif tokenType == "let": |
|
1439 node = LetBlock(tokenizer, staticContext, False) |
|
1440 |
|
1441 elif tokenType in ["null", "this", "true", "false", "identifier", "number", "string", "regexp"]: |
|
1442 node = builder.PRIMARY_build(tokenizer, tokenType) |
|
1443 builder.PRIMARY_finish(node) |
|
1444 |
|
1445 else: |
|
1446 raise SyntaxError("Missing operand. Found type: %s" % tokenType, tokenizer) |
|
1447 |
|
1448 return node |
|