1 from __future__ import unicode_literals
2
3 import itertools
4 import json
5 import math
6 import operator
7 import re
8
9 from .utils import (
10 error_to_compat_str,
11 ExtractorError,
12 js_to_json,
13 remove_quotes,
14 unified_timestamp,
15 )
16 from .compat import (
17 compat_basestring,
18 compat_collections_chain_map as ChainMap,
19 compat_itertools_zip_longest as zip_longest,
20 compat_str,
21 )
22
23
24 def _js_bit_op(op):
25
26 def zeroise(x):
27 return 0 if x in (None, JS_Undefined) else x
28
29 def wrapped(a, b):
30 return op(zeroise(a), zeroise(b)) & 0xffffffff
31
32 return wrapped
33
34
35 def _js_arith_op(op):
36
37 def wrapped(a, b):
38 if JS_Undefined in (a, b):
39 return float('nan')
40 return op(a or 0, b or 0)
41
42 return wrapped
43
44
45 def _js_div(a, b):
46 if JS_Undefined in (a, b) or not (a and b):
47 return float('nan')
48 return operator.truediv(a or 0, b) if b else float('inf')
49
50
51 def _js_mod(a, b):
52 if JS_Undefined in (a, b) or not b:
53 return float('nan')
54 return (a or 0) % b
55
56
57 def _js_exp(a, b):
58 if not b:
59 return 1 # even 0 ** 0 !!
60 elif JS_Undefined in (a, b):
61 return float('nan')
62 return (a or 0) ** b
63
64
65 def _js_eq_op(op):
66
67 def wrapped(a, b):
68 if set((a, b)) <= set((None, JS_Undefined)):
69 return op(a, a)
70 return op(a, b)
71
72 return wrapped
73
74
75 def _js_comp_op(op):
76
77 def wrapped(a, b):
78 if JS_Undefined in (a, b):
79 return False
80 if isinstance(a, compat_basestring):
81 b = compat_str(b or 0)
82 elif isinstance(b, compat_basestring):
83 a = compat_str(a or 0)
84 return op(a or 0, b or 0)
85
86 return wrapped
87
88
89 def _js_ternary(cndn, if_true=True, if_false=False):
90 """Simulate JS's ternary operator (cndn?if_true:if_false)"""
91 if cndn in (False, None, 0, '', JS_Undefined):
92 return if_false
93 try:
94 if math.isnan(cndn): # NB: NaN cannot be checked by membership
95 return if_false
96 except TypeError:
97 pass
98 return if_true
99
100
101 # (op, definition) in order of binding priority, tightest first
102 # avoid dict to maintain order
103 # definition None => Defined in JSInterpreter._operator
104 _OPERATORS = (
105 ('>>', _js_bit_op(operator.rshift)),
106 ('<<', _js_bit_op(operator.lshift)),
107 ('+', _js_arith_op(operator.add)),
108 ('-', _js_arith_op(operator.sub)),
109 ('*', _js_arith_op(operator.mul)),
110 ('/', _js_div),
111 ('%', _js_mod),
112 ('**', _js_exp),
113 )
114
115 _COMP_OPERATORS = (
116 ('===', operator.is_),
117 ('!==', operator.is_not),
118 ('==', _js_eq_op(operator.eq)),
119 ('!=', _js_eq_op(operator.ne)),
120 ('<=', _js_comp_op(operator.le)),
121 ('>=', _js_comp_op(operator.ge)),
122 ('<', _js_comp_op(operator.lt)),
123 ('>', _js_comp_op(operator.gt)),
124 )
125
126 _LOG_OPERATORS = (
127 ('|', _js_bit_op(operator.or_)),
128 ('^', _js_bit_op(operator.xor)),
129 ('&', _js_bit_op(operator.and_)),
130 )
131
132 _SC_OPERATORS = (
133 ('?', None),
134 ('??', None),
135 ('||', None),
136 ('&&', None),
137 )
138
139 _OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS))
140
141 _NAME_RE = r'[a-zA-Z_$][\w$]*'
142 _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]')))
143 _QUOTES = '\'"/'
144
145
146 class JS_Undefined(object):
147 pass
148
149
150 class JS_Break(ExtractorError):
151 def __init__(self):
152 ExtractorError.__init__(self, 'Invalid break')
153
154
155 class JS_Continue(ExtractorError):
156 def __init__(self):
157 ExtractorError.__init__(self, 'Invalid continue')
158
159
160 class JS_Throw(ExtractorError):
161 def __init__(self, e):
162 self.error = e
163 ExtractorError.__init__(self, 'Uncaught exception ' + error_to_compat_str(e))
164
165
166 class LocalNameSpace(ChainMap):
167 def __getitem__(self, key):
168 try:
169 return super(LocalNameSpace, self).__getitem__(key)
170 except KeyError:
171 return JS_Undefined
172
173 def __setitem__(self, key, value):
174 for scope in self.maps:
175 if key in scope:
176 scope[key] = value
177 return
178 self.maps[0][key] = value
179
180 def __delitem__(self, key):
181 raise NotImplementedError('Deleting is not supported')
182
183 def __repr__(self):
184 return 'LocalNameSpace%s' % (self.maps, )
185
186
187 class JSInterpreter(object):
188 __named_object_counter = 0
189
190 _RE_FLAGS = {
191 # special knowledge: Python's re flags are bitmask values, current max 128
192 # invent new bitmask values well above that for literal parsing
193 # TODO: new pattern class to execute matches with these flags
194 'd': 1024, # Generate indices for substring matches
195 'g': 2048, # Global search
196 'i': re.I, # Case-insensitive search
197 'm': re.M, # Multi-line search
198 's': re.S, # Allows . to match newline characters
199 'u': re.U, # Treat a pattern as a sequence of unicode code points
200 'y': 4096, # Perform a "sticky" search that matches starting at the current position in the target string
201 }
202
203 _OBJ_NAME = '__youtube_dl_jsinterp_obj'
204
205 OP_CHARS = None
206
207 def __init__(self, code, objects=None):
208 self.code, self._functions = code, {}
209 self._objects = {} if objects is None else objects
210 if type(self).OP_CHARS is None:
211 type(self).OP_CHARS = self.OP_CHARS = self.__op_chars()
212
213 class Exception(ExtractorError):
214 def __init__(self, msg, *args, **kwargs):
215 expr = kwargs.pop('expr', None)
216 if expr is not None:
217 msg = '{0} in: {1!r}'.format(msg.rstrip(), expr[:100])
218 super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs)
219
220 @classmethod
221 def __op_chars(cls):
222 op_chars = set(';,')
223 for op in cls._all_operators():
224 for c in op[0]:
225 op_chars.add(c)
226 return op_chars
227
228 def _named_object(self, namespace, obj):
229 self.__named_object_counter += 1
230 name = '%s%d' % (self._OBJ_NAME, self.__named_object_counter)
231 namespace[name] = obj
232 return name
233
234 @classmethod
235 def _regex_flags(cls, expr):
236 flags = 0
237 if not expr:
238 return flags, expr
239 for idx, ch in enumerate(expr):
240 if ch not in cls._RE_FLAGS:
241 break
242 flags |= cls._RE_FLAGS[ch]
243 return flags, expr[idx + 1:]
244
245 @classmethod
246 def _separate(cls, expr, delim=',', max_split=None, skip_delims=None):
247 if not expr:
248 return
249 # collections.Counter() is ~10% slower in both 2.7 and 3.9
250 counters = {k: 0 for k in _MATCHING_PARENS.values()}
251 start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
252 in_quote, escaping, skipping = None, False, 0
253 after_op, in_regex_char_group, skip_re = True, False, 0
254
255 for idx, char in enumerate(expr):
256 if skip_re > 0:
257 skip_re -= 1
258 continue
259 if not in_quote:
260 if char in _MATCHING_PARENS:
261 counters[_MATCHING_PARENS[char]] += 1
262 elif char in counters:
263 counters[char] -= 1
264 if not escaping:
265 if char in _QUOTES and in_quote in (char, None):
266 if in_quote or after_op or char != '/':
267 in_quote = None if in_quote and not in_regex_char_group else char
268 elif in_quote == '/' and char in '[]':
269 in_regex_char_group = char == '['
270 escaping = not escaping and in_quote and char == '\\'
271 after_op = not in_quote and (char in cls.OP_CHARS or (char.isspace() and after_op))
272
273 if char != delim[pos] or any(counters.values()) or in_quote:
274 pos = skipping = 0
275 continue
276 elif skipping > 0:
277 skipping -= 1
278 continue
279 elif pos == 0 and skip_delims:
280 here = expr[idx:]
281 for s in skip_delims if isinstance(skip_delims, (list, tuple)) else [skip_delims]:
282 if here.startswith(s) and s:
283 skipping = len(s) - 1
284 break
285 if skipping > 0:
286 continue
287 if pos < delim_len:
288 pos += 1
289 continue
290 yield expr[start: idx - delim_len]
291 start, pos = idx + 1, 0
292 splits += 1
293 if max_split and splits >= max_split:
294 break
295 yield expr[start:]
296
297 @classmethod
298 def _separate_at_paren(cls, expr, delim=None):
299 if delim is None:
300 delim = expr and _MATCHING_PARENS[expr[0]]
301 separated = list(cls._separate(expr, delim, 1))
302
303 if len(separated) < 2:
304 raise cls.Exception('No terminating paren {delim} in {expr}'.format(**locals()))
305 return separated[0][1:].strip(), separated[1].strip()
306
307 @staticmethod
308 def _all_operators():
309 return itertools.chain(
310 # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
311 _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
312
313 def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
314 if op in ('||', '&&'):
315 if (op == '&&') ^ _js_ternary(left_val):
316 return left_val # short circuiting
317 elif op == '??':
318 if left_val not in (None, JS_Undefined):
319 return left_val
320 elif op == '?':
321 right_expr = _js_ternary(left_val, *self._separate(right_expr, ':', 1))
322
323 right_val = self.interpret_expression(right_expr, local_vars, allow_recursion)
324 opfunc = op and next((v for k, v in self._all_operators() if k == op), None)
325 if not opfunc:
326 return right_val
327
328 try:
329 return opfunc(left_val, right_val)
330 except Exception as e:
331 raise self.Exception('Failed to evaluate {left_val!r} {op} {right_val!r}'.format(**locals()), expr, cause=e)
332
333 def _index(self, obj, idx, allow_undefined=False):
334 if idx == 'length':
335 return len(obj)
336 try:
337 return obj[int(idx)] if isinstance(obj, list) else obj[idx]
338 except Exception as e:
339 if allow_undefined:
340 return JS_Undefined
341 raise self.Exception('Cannot get index {idx}'.format(**locals()), expr=repr(obj), cause=e)
342
343 def _dump(self, obj, namespace):
344 try:
345 return json.dumps(obj)
346 except TypeError:
347 return self._named_object(namespace, obj)
348
349 def interpret_statement(self, stmt, local_vars, allow_recursion=100):
350 if allow_recursion < 0:
351 raise self.Exception('Recursion limit reached')
352 allow_recursion -= 1
353
354 should_return = False
355 sub_statements = list(self._separate(stmt, ';')) or ['']
356 expr = stmt = sub_statements.pop().strip()
357 for sub_stmt in sub_statements:
358 ret, should_return = self.interpret_statement(sub_stmt, local_vars, allow_recursion)
359 if should_return:
360 return ret, should_return
361
362 m = re.match(r'(?P<var>(?:var|const|let)\s)|return(?:\s+|(?=["\'])|$)|(?P<throw>throw\s+)', stmt)
363 if m:
364 expr = stmt[len(m.group(0)):].strip()
365 if m.group('throw'):
366 raise JS_Throw(self.interpret_expression(expr, local_vars, allow_recursion))
367 should_return = not m.group('var')
368 if not expr:
369 return None, should_return
370
371 if expr[0] in _QUOTES:
372 inner, outer = self._separate(expr, expr[0], 1)
373 if expr[0] == '/':
374 flags, outer = self._regex_flags(outer)
375 inner = re.compile(inner[1:], flags=flags) # , strict=True))
376 else:
377 inner = json.loads(js_to_json(inner + expr[0])) # , strict=True))
378 if not outer:
379 return inner, should_return
380 expr = self._named_object(local_vars, inner) + outer
381
382 if expr.startswith('new '):
383 obj = expr[4:]
384 if obj.startswith('Date('):
385 left, right = self._separate_at_paren(obj[4:])
386 expr = unified_timestamp(
387 self.interpret_expression(left, local_vars, allow_recursion), False)
388 if not expr:
389 raise self.Exception('Failed to parse date {left!r}'.format(**locals()), expr=expr)
390 expr = self._dump(int(expr * 1000), local_vars) + right
391 else:
392 raise self.Exception('Unsupported object {obj}'.format(**locals()), expr=expr)
393
394 if expr.startswith('void '):
395 left = self.interpret_expression(expr[5:], local_vars, allow_recursion)
396 return None, should_return
397
398 if expr.startswith('{'):
399 inner, outer = self._separate_at_paren(expr)
400 # try for object expression (Map)
401 sub_expressions = [list(self._separate(sub_expr.strip(), ':', 1)) for sub_expr in self._separate(inner)]
402 if all(len(sub_expr) == 2 for sub_expr in sub_expressions):
403 return dict(
404 (key_expr if re.match(_NAME_RE, key_expr) else key_expr,
405 self.interpret_expression(val_expr, local_vars, allow_recursion))
406 for key_expr, val_expr in sub_expressions), should_return
407 # or statement list
408 inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion)
409 if not outer or should_abort:
410 return inner, should_abort or should_return
411 else:
412 expr = self._dump(inner, local_vars) + outer
413
414 if expr.startswith('('):
415 inner, outer = self._separate_at_paren(expr)
416 inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion)
417 if not outer or should_abort:
418 return inner, should_abort or should_return
419 else:
420 expr = self._dump(inner, local_vars) + outer
421
422 if expr.startswith('['):
423 inner, outer = self._separate_at_paren(expr)
424 name = self._named_object(local_vars, [
425 self.interpret_expression(item, local_vars, allow_recursion)
426 for item in self._separate(inner)])
427 expr = name + outer
428
429 m = re.match(r'''(?x)
430 (?P<try>try)\s*\{|
431 (?P<switch>switch)\s*\(|
432 (?P<for>for)\s*\(
433 ''', expr)
434 md = m.groupdict() if m else {}
435 if md.get('try'):
436 try_expr, expr = self._separate_at_paren(expr[m.end() - 1:])
437 err = None
438 try:
439 ret, should_abort = self.interpret_statement(try_expr, local_vars, allow_recursion)
440 if should_abort:
441 return ret, True
442 except Exception as e:
443 # XXX: This works for now, but makes debugging future issues very hard
444 err = e
445
446 pending = (None, False)
447 m = re.match(r'catch\s*(?P<err>\(\s*{_NAME_RE}\s*\))?\{{'.format(**globals()), expr)
448 if m:
449 sub_expr, expr = self._separate_at_paren(expr[m.end() - 1:])
450 if err:
451 catch_vars = {}
452 if m.group('err'):
453 catch_vars[m.group('err')] = err.error if isinstance(err, JS_Throw) else err
454 catch_vars = local_vars.new_child(m=catch_vars)
455 err = None
456 pending = self.interpret_statement(sub_expr, catch_vars, allow_recursion)
457
458 m = re.match(r'finally\s*\{', expr)
459 if m:
460 sub_expr, expr = self._separate_at_paren(expr[m.end() - 1:])
461 ret, should_abort = self.interpret_statement(sub_expr, local_vars, allow_recursion)
462 if should_abort:
463 return ret, True
464
465 ret, should_abort = pending
466 if should_abort:
467 return ret, True
468
469 if err:
470 raise err
471
472 elif md.get('for'):
473 constructor, remaining = self._separate_at_paren(expr[m.end() - 1:])
474 if remaining.startswith('{'):
475 body, expr = self._separate_at_paren(remaining)
476 else:
477 switch_m = re.match(r'switch\s*\(', remaining) # FIXME
478 if switch_m:
479 switch_val, remaining = self._separate_at_paren(remaining[switch_m.end() - 1:])
480 body, expr = self._separate_at_paren(remaining, '}')
481 body = 'switch(%s){%s}' % (switch_val, body)
482 else:
483 body, expr = remaining, ''
484 start, cndn, increment = self._separate(constructor, ';')
485 self.interpret_expression(start, local_vars, allow_recursion)
486 while True:
487 if not _js_ternary(self.interpret_expression(cndn, local_vars, allow_recursion)):
488 break
489 try:
490 ret, should_abort = self.interpret_statement(body, local_vars, allow_recursion)
491 if should_abort:
492 return ret, True
493 except JS_Break:
494 break
495 except JS_Continue:
496 pass
497 self.interpret_expression(increment, local_vars, allow_recursion)
498
499 elif md.get('switch'):
500 switch_val, remaining = self._separate_at_paren(expr[m.end() - 1:])
501 switch_val = self.interpret_expression(switch_val, local_vars, allow_recursion)
502 body, expr = self._separate_at_paren(remaining, '}')
503 items = body.replace('default:', 'case default:').split('case ')[1:]
504 for default in (False, True):
505 matched = False
506 for item in items:
507 case, stmt = (i.strip() for i in self._separate(item, ':', 1))
508 if default:
509 matched = matched or case == 'default'
510 elif not matched:
511 matched = (case != 'default'
512 and switch_val == self.interpret_expression(case, local_vars, allow_recursion))
513 if not matched:
514 continue
515 try:
516 ret, should_abort = self.interpret_statement(stmt, local_vars, allow_recursion)
517 if should_abort:
518 return ret
519 except JS_Break:
520 break
521 if matched:
522 break
523
524 if md:
525 ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion)
526 return ret, should_abort or should_return
527
528 # Comma separated statements
529 sub_expressions = list(self._separate(expr))
530 if len(sub_expressions) > 1:
531 for sub_expr in sub_expressions:
532 ret, should_abort = self.interpret_statement(sub_expr, local_vars, allow_recursion)
533 if should_abort:
534 return ret, True
535 return ret, False
536
537 for m in re.finditer(r'''(?x)
538 (?P<pre_sign>\+\+|--)(?P<var1>{_NAME_RE})|
539 (?P<var2>{_NAME_RE})(?P<post_sign>\+\+|--)'''.format(**globals()), expr):
540 var = m.group('var1') or m.group('var2')
541 start, end = m.span()
542 sign = m.group('pre_sign') or m.group('post_sign')
543 ret = local_vars[var]
544 local_vars[var] += 1 if sign[0] == '+' else -1
545 if m.group('pre_sign'):
546 ret = local_vars[var]
547 expr = expr[:start] + self._dump(ret, local_vars) + expr[end:]
548
549 if not expr:
550 return None, should_return
551
552 m = re.match(r'''(?x)
553 (?P<assign>
554 (?P<out>{_NAME_RE})(?:\[(?P<index>[^\]]+?)\])?\s*
555 (?P<op>{_OPERATOR_RE})?
556 =(?!=)(?P<expr>.*)$
557 )|(?P<return>
558 (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
559 )|(?P<indexing>
560 (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
561 )|(?P<attribute>
562 (?P<var>{_NAME_RE})(?:(?P<nullish>\?)?\.(?P<member>[^(]+)|\[(?P<member2>[^\]]+)\])\s*
563 )|(?P<function>
564 (?P<fname>{_NAME_RE})\((?P<args>.*)\)$
565 )'''.format(**globals()), expr)
566 md = m.groupdict() if m else {}
567 if md.get('assign'):
568 left_val = local_vars.get(m.group('out'))
569
570 if not m.group('index'):
571 local_vars[m.group('out')] = self._operator(
572 m.group('op'), left_val, m.group('expr'), expr, local_vars, allow_recursion)
573 return local_vars[m.group('out')], should_return
574 elif left_val in (None, JS_Undefined):
575 raise self.Exception('Cannot index undefined variable ' + m.group('out'), expr=expr)
576
577 idx = self.interpret_expression(m.group('index'), local_vars, allow_recursion)
578 if not isinstance(idx, (int, float)):
579 raise self.Exception('List index %s must be integer' % (idx, ), expr=expr)
580 idx = int(idx)
581 left_val[idx] = self._operator(
582 m.group('op'), self._index(left_val, idx), m.group('expr'), expr, local_vars, allow_recursion)
583 return left_val[idx], should_return
584
585 elif expr.isdigit():
586 return int(expr), should_return
587
588 elif expr == 'break':
589 raise JS_Break()
590 elif expr == 'continue':
591 raise JS_Continue()
592
593 elif expr == 'undefined':
594 return JS_Undefined, should_return
595 elif expr == 'NaN':
596 return float('NaN'), should_return
597
598 elif md.get('return'):
599 return local_vars[m.group('name')], should_return
600
601 try:
602 ret = json.loads(js_to_json(expr)) # strict=True)
603 if not md.get('attribute'):
604 return ret, should_return
605 except ValueError:
606 pass
607
608 if md.get('indexing'):
609 val = local_vars[m.group('in')]
610 idx = self.interpret_expression(m.group('idx'), local_vars, allow_recursion)
611 return self._index(val, idx), should_return
612
613 for op, _ in self._all_operators():
614 # hackety: </> have higher priority than <</>>, but don't confuse them
615 skip_delim = (op + op) if op in '<>*?' else None
616 if op == '?':
617 skip_delim = (skip_delim, '?.')
618 separated = list(self._separate(expr, op, skip_delims=skip_delim))
619 if len(separated) < 2:
620 continue
621
622 right_expr = separated.pop()
623 while op == '-' and len(separated) > 1 and not separated[-1].strip():
624 right_expr = '-' + right_expr
625 separated.pop()
626 left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion)
627 return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), should_return
628
629 if md.get('attribute'):
630 variable, member, nullish = m.group('var', 'member', 'nullish')
631 if not member:
632 member = self.interpret_expression(m.group('member2'), local_vars, allow_recursion)
633 arg_str = expr[m.end():]
634 if arg_str.startswith('('):
635 arg_str, remaining = self._separate_at_paren(arg_str)
636 else:
637 arg_str, remaining = None, arg_str
638
639 def assertion(cndn, msg):
640 """ assert, but without risk of getting optimized out """
641 if not cndn:
642 memb = member
643 raise self.Exception('{member} {msg}'.format(**locals()), expr=expr)
644
645 def eval_method():
646 if (variable, member) == ('console', 'debug'):
647 return
648 types = {
649 'String': compat_str,
650 'Math': float,
651 }
652 obj = local_vars.get(variable)
653 if obj in (JS_Undefined, None):
654 obj = types.get(variable, JS_Undefined)
655 if obj is JS_Undefined:
656 try:
657 if variable not in self._objects:
658 self._objects[variable] = self.extract_object(variable)
659 obj = self._objects[variable]
660 except self.Exception:
661 if not nullish:
662 raise
663
664 if nullish and obj is JS_Undefined:
665 return JS_Undefined
666
667 # Member access
668 if arg_str is None:
669 return self._index(obj, member, nullish)
670
671 # Function call
672 argvals = [
673 self.interpret_expression(v, local_vars, allow_recursion)
674 for v in self._separate(arg_str)]
675
676 if obj == compat_str:
677 if member == 'fromCharCode':
678 assertion(argvals, 'takes one or more arguments')
679 return ''.join(map(chr, argvals))
680 raise self.Exception('Unsupported string method ' + member, expr=expr)
681 elif obj == float:
682 if member == 'pow':
683 assertion(len(argvals) == 2, 'takes two arguments')
684 return argvals[0] ** argvals[1]
685 raise self.Exception('Unsupported Math method ' + member, expr=expr)
686
687 if member == 'split':
688 assertion(argvals, 'takes one or more arguments')
689 assertion(len(argvals) == 1, 'with limit argument is not implemented')
690 return obj.split(argvals[0]) if argvals[0] else list(obj)
691 elif member == 'join':
692 assertion(isinstance(obj, list), 'must be applied on a list')
693 assertion(len(argvals) == 1, 'takes exactly one argument')
694 return argvals[0].join(obj)
695 elif member == 'reverse':
696 assertion(not argvals, 'does not take any arguments')
697 obj.reverse()
698 return obj
699 elif member == 'slice':
700 assertion(isinstance(obj, list), 'must be applied on a list')
701 assertion(len(argvals) == 1, 'takes exactly one argument')
702 return obj[argvals[0]:]
703 elif member == 'splice':
704 assertion(isinstance(obj, list), 'must be applied on a list')
705 assertion(argvals, 'takes one or more arguments')
706 index, howMany = map(int, (argvals + [len(obj)])[:2])
707 if index < 0:
708 index += len(obj)
709 add_items = argvals[2:]
710 res = []
711 for i in range(index, min(index + howMany, len(obj))):
712 res.append(obj.pop(index))
713 for i, item in enumerate(add_items):
714 obj.insert(index + i, item)
715 return res
716 elif member == 'unshift':
717 assertion(isinstance(obj, list), 'must be applied on a list')
718 assertion(argvals, 'takes one or more arguments')
719 for item in reversed(argvals):
720 obj.insert(0, item)
721 return obj
722 elif member == 'pop':
723 assertion(isinstance(obj, list), 'must be applied on a list')
724 assertion(not argvals, 'does not take any arguments')
725 if not obj:
726 return
727 return obj.pop()
728 elif member == 'push':
729 assertion(argvals, 'takes one or more arguments')
730 obj.extend(argvals)
731 return obj
732 elif member == 'forEach':
733 assertion(argvals, 'takes one or more arguments')
734 assertion(len(argvals) <= 2, 'takes at-most 2 arguments')
735 f, this = (argvals + [''])[:2]
736 return [f((item, idx, obj), {'this': this}, allow_recursion) for idx, item in enumerate(obj)]
737 elif member == 'indexOf':
738 assertion(argvals, 'takes one or more arguments')
739 assertion(len(argvals) <= 2, 'takes at-most 2 arguments')
740 idx, start = (argvals + [0])[:2]
741 try:
742 return obj.index(idx, start)
743 except ValueError:
744 return -1
745 elif member == 'charCodeAt':
746 assertion(isinstance(obj, compat_str), 'must be applied on a string')
747 # assertion(len(argvals) == 1, 'takes exactly one argument') # but not enforced
748 idx = argvals[0] if isinstance(argvals[0], int) else 0
749 if idx >= len(obj):
750 return None
751 return ord(obj[idx])
752
753 idx = int(member) if isinstance(obj, list) else member
754 return obj[idx](argvals, allow_recursion=allow_recursion)
755
756 if remaining:
757 ret, should_abort = self.interpret_statement(
758 self._named_object(local_vars, eval_method()) + remaining,
759 local_vars, allow_recursion)
760 return ret, should_return or should_abort
761 else:
762 return eval_method(), should_return
763
764 elif md.get('function'):
765 fname = m.group('fname')
766 argvals = [self.interpret_expression(v, local_vars, allow_recursion)
767 for v in self._separate(m.group('args'))]
768 if fname in local_vars:
769 return local_vars[fname](argvals, allow_recursion=allow_recursion), should_return
770 elif fname not in self._functions:
771 self._functions[fname] = self.extract_function(fname)
772 return self._functions[fname](argvals, allow_recursion=allow_recursion), should_return
773
774 raise self.Exception(
775 'Unsupported JS expression ' + (expr[:40] if expr != stmt else ''), expr=stmt)
776
777 def interpret_expression(self, expr, local_vars, allow_recursion):
778 ret, should_return = self.interpret_statement(expr, local_vars, allow_recursion)
779 if should_return:
780 raise self.Exception('Cannot return from an expression', expr)
781 return ret
782
783 def extract_object(self, objname):
784 _FUNC_NAME_RE = r'''(?:[a-zA-Z$0-9]+|"[a-zA-Z$0-9]+"|'[a-zA-Z$0-9]+')'''
785 obj = {}
786 obj_m = re.search(
787 r'''(?x)
788 (?<!this\.)%s\s*=\s*{\s*
789 (?P<fields>(%s\s*:\s*function\s*\(.*?\)\s*{.*?}(?:,\s*)?)*)
790 }\s*;
791 ''' % (re.escape(objname), _FUNC_NAME_RE),
792 self.code)
793 if not obj_m:
794 raise self.Exception('Could not find object ' + objname)
795 fields = obj_m.group('fields')
796 # Currently, it only supports function definitions
797 fields_m = re.finditer(
798 r'''(?x)
799 (?P<key>%s)\s*:\s*function\s*\((?P<args>(?:%s|,)*)\){(?P<code>[^}]+)}
800 ''' % (_FUNC_NAME_RE, _NAME_RE),
801 fields)
802 for f in fields_m:
803 argnames = self.build_arglist(f.group('args'))
804 obj[remove_quotes(f.group('key'))] = self.build_function(argnames, f.group('code'))
805
806 return obj
807
808 def extract_function_code(self, funcname):
809 """ @returns argnames, code """
810 func_m = re.search(
811 r'''(?xs)
812 (?:
813 function\s+%(name)s|
814 [{;,]\s*%(name)s\s*=\s*function|
815 (?:var|const|let)\s+%(name)s\s*=\s*function
816 )\s*
817 \((?P<args>[^)]*)\)\s*
818 (?P<code>{.+})''' % {'name': re.escape(funcname)},
819 self.code)
820 code, _ = self._separate_at_paren(func_m.group('code')) # refine the match
821 if func_m is None:
822 raise self.Exception('Could not find JS function "{funcname}"'.format(**locals()))
823 return self.build_arglist(func_m.group('args')), code
824
825 def extract_function(self, funcname):
826 return self.extract_function_from_code(*self.extract_function_code(funcname))
827
828 def extract_function_from_code(self, argnames, code, *global_stack):
829 local_vars = {}
830 while True:
831 mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code)
832 if mobj is None:
833 break
834 start, body_start = mobj.span()
835 body, remaining = self._separate_at_paren(code[body_start - 1:], '}')
836 name = self._named_object(local_vars, self.extract_function_from_code(
837 [x.strip() for x in mobj.group('args').split(',')],
838 body, local_vars, *global_stack))
839 code = code[:start] + name + remaining
840 return self.build_function(argnames, code, local_vars, *global_stack)
841
842 def call_function(self, funcname, *args):
843 return self.extract_function(funcname)(args)
844
845 @classmethod
846 def build_arglist(cls, arg_text):
847 if not arg_text:
848 return []
849
850 def valid_arg(y):
851 y = y.strip()
852 if not y:
853 raise cls.Exception('Missing arg in "%s"' % (arg_text, ))
854 return y
855
856 return [valid_arg(x) for x in cls._separate(arg_text)]
857
858 def build_function(self, argnames, code, *global_stack):
859 global_stack = list(global_stack) or [{}]
860 argnames = tuple(argnames)
861
862 def resf(args, kwargs={}, allow_recursion=100):
863 global_stack[0].update(
864 zip_longest(argnames, args, fillvalue=None))
865 global_stack[0].update(kwargs)
866 var_stack = LocalNameSpace(*global_stack)
867 ret, should_abort = self.interpret_statement(code.replace('\n', ' '), var_stack, allow_recursion - 1)
868 if should_abort:
869 return ret
870 return resf
|
