]>
Commit | Line | Data |
---|---|---|
1 | import collections | |
2 | import contextlib | |
3 | import itertools | |
4 | import json | |
5 | import math | |
6 | import operator | |
7 | import re | |
8 | ||
9 | from .utils import ( | |
10 | NO_DEFAULT, | |
11 | ExtractorError, | |
12 | js_to_json, | |
13 | remove_quotes, | |
14 | truncate_string, | |
15 | unified_timestamp, | |
16 | write_string, | |
17 | ) | |
18 | ||
19 | ||
20 | def _js_bit_op(op): | |
21 | def zeroise(x): | |
22 | return 0 if x in (None, JS_Undefined) else x | |
23 | ||
24 | def wrapped(a, b): | |
25 | return op(zeroise(a), zeroise(b)) & 0xffffffff | |
26 | ||
27 | return wrapped | |
28 | ||
29 | ||
30 | def _js_arith_op(op): | |
31 | ||
32 | def wrapped(a, b): | |
33 | if JS_Undefined in (a, b): | |
34 | return float('nan') | |
35 | return op(a or 0, b or 0) | |
36 | ||
37 | return wrapped | |
38 | ||
39 | ||
40 | def _js_div(a, b): | |
41 | if JS_Undefined in (a, b) or not (a and b): | |
42 | return float('nan') | |
43 | return (a or 0) / b if b else float('inf') | |
44 | ||
45 | ||
46 | def _js_mod(a, b): | |
47 | if JS_Undefined in (a, b) or not b: | |
48 | return float('nan') | |
49 | return (a or 0) % b | |
50 | ||
51 | ||
52 | def _js_exp(a, b): | |
53 | if not b: | |
54 | return 1 # even 0 ** 0 !! | |
55 | elif JS_Undefined in (a, b): | |
56 | return float('nan') | |
57 | return (a or 0) ** b | |
58 | ||
59 | ||
60 | def _js_eq_op(op): | |
61 | ||
62 | def wrapped(a, b): | |
63 | if {a, b} <= {None, JS_Undefined}: | |
64 | return op(a, a) | |
65 | return op(a, b) | |
66 | ||
67 | return wrapped | |
68 | ||
69 | ||
70 | def _js_comp_op(op): | |
71 | ||
72 | def wrapped(a, b): | |
73 | if JS_Undefined in (a, b): | |
74 | return False | |
75 | if isinstance(a, str) or isinstance(b, str): | |
76 | return op(str(a or 0), str(b or 0)) | |
77 | return op(a or 0, b or 0) | |
78 | ||
79 | return wrapped | |
80 | ||
81 | ||
82 | def _js_ternary(cndn, if_true=True, if_false=False): | |
83 | """Simulate JS's ternary operator (cndn?if_true:if_false)""" | |
84 | if cndn in (False, None, 0, '', JS_Undefined): | |
85 | return if_false | |
86 | with contextlib.suppress(TypeError): | |
87 | if math.isnan(cndn): # NB: NaN cannot be checked by membership | |
88 | return if_false | |
89 | return if_true | |
90 | ||
91 | ||
92 | # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence | |
93 | _OPERATORS = { # None => Defined in JSInterpreter._operator | |
94 | '?': None, | |
95 | '??': None, | |
96 | '||': None, | |
97 | '&&': None, | |
98 | ||
99 | '|': _js_bit_op(operator.or_), | |
100 | '^': _js_bit_op(operator.xor), | |
101 | '&': _js_bit_op(operator.and_), | |
102 | ||
103 | '===': operator.is_, | |
104 | '!==': operator.is_not, | |
105 | '==': _js_eq_op(operator.eq), | |
106 | '!=': _js_eq_op(operator.ne), | |
107 | ||
108 | '<=': _js_comp_op(operator.le), | |
109 | '>=': _js_comp_op(operator.ge), | |
110 | '<': _js_comp_op(operator.lt), | |
111 | '>': _js_comp_op(operator.gt), | |
112 | ||
113 | '>>': _js_bit_op(operator.rshift), | |
114 | '<<': _js_bit_op(operator.lshift), | |
115 | ||
116 | '+': _js_arith_op(operator.add), | |
117 | '-': _js_arith_op(operator.sub), | |
118 | ||
119 | '*': _js_arith_op(operator.mul), | |
120 | '%': _js_mod, | |
121 | '/': _js_div, | |
122 | '**': _js_exp, | |
123 | } | |
124 | ||
125 | _COMP_OPERATORS = {'===', '!==', '==', '!=', '<=', '>=', '<', '>'} | |
126 | ||
127 | _NAME_RE = r'[a-zA-Z_$][\w$]*' | |
128 | _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]'))) | |
129 | _QUOTES = '\'"/' | |
130 | ||
131 | ||
132 | class JS_Undefined: | |
133 | pass | |
134 | ||
135 | ||
136 | class JS_Break(ExtractorError): | |
137 | def __init__(self): | |
138 | ExtractorError.__init__(self, 'Invalid break') | |
139 | ||
140 | ||
141 | class JS_Continue(ExtractorError): | |
142 | def __init__(self): | |
143 | ExtractorError.__init__(self, 'Invalid continue') | |
144 | ||
145 | ||
146 | class JS_Throw(ExtractorError): | |
147 | def __init__(self, e): | |
148 | self.error = e | |
149 | ExtractorError.__init__(self, f'Uncaught exception {e}') | |
150 | ||
151 | ||
152 | class LocalNameSpace(collections.ChainMap): | |
153 | def __setitem__(self, key, value): | |
154 | for scope in self.maps: | |
155 | if key in scope: | |
156 | scope[key] = value | |
157 | return | |
158 | self.maps[0][key] = value | |
159 | ||
160 | def __delitem__(self, key): | |
161 | raise NotImplementedError('Deleting is not supported') | |
162 | ||
163 | ||
164 | class Debugger: | |
165 | import sys | |
166 | ENABLED = False and 'pytest' in sys.modules | |
167 | ||
168 | @staticmethod | |
169 | def write(*args, level=100): | |
170 | write_string(f'[debug] JS: {" " * (100 - level)}' | |
171 | f'{" ".join(truncate_string(str(x), 50, 50) for x in args)}\n') | |
172 | ||
173 | @classmethod | |
174 | def wrap_interpreter(cls, f): | |
175 | def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs): | |
176 | if cls.ENABLED and stmt.strip(): | |
177 | cls.write(stmt, level=allow_recursion) | |
178 | try: | |
179 | ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs) | |
180 | except Exception as e: | |
181 | if cls.ENABLED: | |
182 | if isinstance(e, ExtractorError): | |
183 | e = e.orig_msg | |
184 | cls.write('=> Raises:', e, '<-|', stmt, level=allow_recursion) | |
185 | raise | |
186 | if cls.ENABLED and stmt.strip(): | |
187 | cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion) | |
188 | return ret, should_ret | |
189 | return interpret_statement | |
190 | ||
191 | ||
192 | class JSInterpreter: | |
193 | __named_object_counter = 0 | |
194 | ||
195 | _RE_FLAGS = { | |
196 | # special knowledge: Python's re flags are bitmask values, current max 128 | |
197 | # invent new bitmask values well above that for literal parsing | |
198 | # TODO: new pattern class to execute matches with these flags | |
199 | 'd': 1024, # Generate indices for substring matches | |
200 | 'g': 2048, # Global search | |
201 | 'i': re.I, # Case-insensitive search | |
202 | 'm': re.M, # Multi-line search | |
203 | 's': re.S, # Allows . to match newline characters | |
204 | 'u': re.U, # Treat a pattern as a sequence of unicode code points | |
205 | 'y': 4096, # Perform a "sticky" search that matches starting at the current position in the target string | |
206 | } | |
207 | ||
208 | _EXC_NAME = '__yt_dlp_exception__' | |
209 | ||
210 | def __init__(self, code, objects=None): | |
211 | self.code, self._functions = code, {} | |
212 | self._objects = {} if objects is None else objects | |
213 | ||
214 | class Exception(ExtractorError): | |
215 | def __init__(self, msg, expr=None, *args, **kwargs): | |
216 | if expr is not None: | |
217 | msg = f'{msg.rstrip()} in: {truncate_string(expr, 50, 50)}' | |
218 | super().__init__(msg, *args, **kwargs) | |
219 | ||
220 | def _named_object(self, namespace, obj): | |
221 | self.__named_object_counter += 1 | |
222 | name = f'__yt_dlp_jsinterp_obj{self.__named_object_counter}' | |
223 | namespace[name] = obj | |
224 | return name | |
225 | ||
226 | @classmethod | |
227 | def _regex_flags(cls, expr): | |
228 | flags = 0 | |
229 | if not expr: | |
230 | return flags, expr | |
231 | for idx, ch in enumerate(expr): | |
232 | if ch not in cls._RE_FLAGS: | |
233 | break | |
234 | flags |= cls._RE_FLAGS[ch] | |
235 | return flags, expr[idx + 1:] | |
236 | ||
237 | @staticmethod | |
238 | def _separate(expr, delim=',', max_split=None): | |
239 | OP_CHARS = '+-*/%&|^=<>!,;{}:[' | |
240 | if not expr: | |
241 | return | |
242 | counters = {k: 0 for k in _MATCHING_PARENS.values()} | |
243 | start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1 | |
244 | in_quote, escaping, after_op, in_regex_char_group = None, False, True, False | |
245 | for idx, char in enumerate(expr): | |
246 | if not in_quote and char in _MATCHING_PARENS: | |
247 | counters[_MATCHING_PARENS[char]] += 1 | |
248 | elif not in_quote and char in counters: | |
249 | # Something's wrong if we get negative, but ignore it anyway | |
250 | if counters[char]: | |
251 | counters[char] -= 1 | |
252 | elif not escaping: | |
253 | if char in _QUOTES and in_quote in (char, None): | |
254 | if in_quote or after_op or char != '/': | |
255 | in_quote = None if in_quote and not in_regex_char_group else char | |
256 | elif in_quote == '/' and char in '[]': | |
257 | in_regex_char_group = char == '[' | |
258 | escaping = not escaping and in_quote and char == '\\' | |
259 | after_op = not in_quote and char in OP_CHARS or (char.isspace() and after_op) | |
260 | ||
261 | if char != delim[pos] or any(counters.values()) or in_quote: | |
262 | pos = 0 | |
263 | continue | |
264 | elif pos != delim_len: | |
265 | pos += 1 | |
266 | continue | |
267 | yield expr[start: idx - delim_len] | |
268 | start, pos = idx + 1, 0 | |
269 | splits += 1 | |
270 | if max_split and splits >= max_split: | |
271 | break | |
272 | yield expr[start:] | |
273 | ||
274 | @classmethod | |
275 | def _separate_at_paren(cls, expr, delim=None): | |
276 | if delim is None: | |
277 | delim = expr and _MATCHING_PARENS[expr[0]] | |
278 | separated = list(cls._separate(expr, delim, 1)) | |
279 | if len(separated) < 2: | |
280 | raise cls.Exception(f'No terminating paren {delim}', expr) | |
281 | return separated[0][1:].strip(), separated[1].strip() | |
282 | ||
283 | def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion): | |
284 | if op in ('||', '&&'): | |
285 | if (op == '&&') ^ _js_ternary(left_val): | |
286 | return left_val # short circuiting | |
287 | elif op == '??': | |
288 | if left_val not in (None, JS_Undefined): | |
289 | return left_val | |
290 | elif op == '?': | |
291 | right_expr = _js_ternary(left_val, *self._separate(right_expr, ':', 1)) | |
292 | ||
293 | right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) | |
294 | if not _OPERATORS.get(op): | |
295 | return right_val | |
296 | ||
297 | try: | |
298 | return _OPERATORS[op](left_val, right_val) | |
299 | except Exception as e: | |
300 | raise self.Exception(f'Failed to evaluate {left_val!r} {op} {right_val!r}', expr, cause=e) | |
301 | ||
302 | def _index(self, obj, idx, allow_undefined=False): | |
303 | if idx == 'length': | |
304 | return len(obj) | |
305 | try: | |
306 | return obj[int(idx)] if isinstance(obj, list) else obj[idx] | |
307 | except Exception as e: | |
308 | if allow_undefined: | |
309 | return JS_Undefined | |
310 | raise self.Exception(f'Cannot get index {idx}', repr(obj), cause=e) | |
311 | ||
312 | def _dump(self, obj, namespace): | |
313 | try: | |
314 | return json.dumps(obj) | |
315 | except TypeError: | |
316 | return self._named_object(namespace, obj) | |
317 | ||
318 | @Debugger.wrap_interpreter | |
319 | def interpret_statement(self, stmt, local_vars, allow_recursion=100): | |
320 | if allow_recursion < 0: | |
321 | raise self.Exception('Recursion limit reached') | |
322 | allow_recursion -= 1 | |
323 | ||
324 | should_return = False | |
325 | sub_statements = list(self._separate(stmt, ';')) or [''] | |
326 | expr = stmt = sub_statements.pop().strip() | |
327 | ||
328 | for sub_stmt in sub_statements: | |
329 | ret, should_return = self.interpret_statement(sub_stmt, local_vars, allow_recursion) | |
330 | if should_return: | |
331 | return ret, should_return | |
332 | ||
333 | m = re.match(r'(?P<var>(?:var|const|let)\s)|return(?:\s+|(?=["\'])|$)|(?P<throw>throw\s+)', stmt) | |
334 | if m: | |
335 | expr = stmt[len(m.group(0)):].strip() | |
336 | if m.group('throw'): | |
337 | raise JS_Throw(self.interpret_expression(expr, local_vars, allow_recursion)) | |
338 | should_return = not m.group('var') | |
339 | if not expr: | |
340 | return None, should_return | |
341 | ||
342 | if expr[0] in _QUOTES: | |
343 | inner, outer = self._separate(expr, expr[0], 1) | |
344 | if expr[0] == '/': | |
345 | flags, outer = self._regex_flags(outer) | |
346 | # Avoid https://github.com/python/cpython/issues/74534 | |
347 | inner = re.compile(inner[1:].replace('[[', r'[\['), flags=flags) | |
348 | else: | |
349 | inner = json.loads(js_to_json(f'{inner}{expr[0]}', strict=True)) | |
350 | if not outer: | |
351 | return inner, should_return | |
352 | expr = self._named_object(local_vars, inner) + outer | |
353 | ||
354 | if expr.startswith('new '): | |
355 | obj = expr[4:] | |
356 | if obj.startswith('Date('): | |
357 | left, right = self._separate_at_paren(obj[4:]) | |
358 | expr = unified_timestamp( | |
359 | self.interpret_expression(left, local_vars, allow_recursion), False) | |
360 | if not expr: | |
361 | raise self.Exception(f'Failed to parse date {left!r}', expr) | |
362 | expr = self._dump(int(expr * 1000), local_vars) + right | |
363 | else: | |
364 | raise self.Exception(f'Unsupported object {obj}', expr) | |
365 | ||
366 | if expr.startswith('void '): | |
367 | left = self.interpret_expression(expr[5:], local_vars, allow_recursion) | |
368 | return None, should_return | |
369 | ||
370 | if expr.startswith('{'): | |
371 | inner, outer = self._separate_at_paren(expr) | |
372 | # try for object expression (Map) | |
373 | sub_expressions = [list(self._separate(sub_expr.strip(), ':', 1)) for sub_expr in self._separate(inner)] | |
374 | if all(len(sub_expr) == 2 for sub_expr in sub_expressions): | |
375 | def dict_item(key, val): | |
376 | val = self.interpret_expression(val, local_vars, allow_recursion) | |
377 | if re.match(_NAME_RE, key): | |
378 | return key, val | |
379 | return self.interpret_expression(key, local_vars, allow_recursion), val | |
380 | ||
381 | return dict(dict_item(k, v) for k, v in sub_expressions), should_return | |
382 | ||
383 | inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion) | |
384 | if not outer or should_abort: | |
385 | return inner, should_abort or should_return | |
386 | else: | |
387 | expr = self._dump(inner, local_vars) + outer | |
388 | ||
389 | if expr.startswith('('): | |
390 | inner, outer = self._separate_at_paren(expr) | |
391 | inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion) | |
392 | if not outer or should_abort: | |
393 | return inner, should_abort or should_return | |
394 | else: | |
395 | expr = self._dump(inner, local_vars) + outer | |
396 | ||
397 | if expr.startswith('['): | |
398 | inner, outer = self._separate_at_paren(expr) | |
399 | name = self._named_object(local_vars, [ | |
400 | self.interpret_expression(item, local_vars, allow_recursion) | |
401 | for item in self._separate(inner)]) | |
402 | expr = name + outer | |
403 | ||
404 | m = re.match(r'''(?x) | |
405 | (?P<try>try)\s*\{| | |
406 | (?P<switch>switch)\s*\(| | |
407 | (?P<for>for)\s*\( | |
408 | ''', expr) | |
409 | md = m.groupdict() if m else {} | |
410 | if md.get('try'): | |
411 | try_expr, expr = self._separate_at_paren(expr[m.end() - 1:]) | |
412 | err = None | |
413 | try: | |
414 | ret, should_abort = self.interpret_statement(try_expr, local_vars, allow_recursion) | |
415 | if should_abort: | |
416 | return ret, True | |
417 | except Exception as e: | |
418 | # XXX: This works for now, but makes debugging future issues very hard | |
419 | err = e | |
420 | ||
421 | pending = (None, False) | |
422 | m = re.match(r'catch\s*(?P<err>\(\s*{_NAME_RE}\s*\))?\{{'.format(**globals()), expr) | |
423 | if m: | |
424 | sub_expr, expr = self._separate_at_paren(expr[m.end() - 1:]) | |
425 | if err: | |
426 | catch_vars = {} | |
427 | if m.group('err'): | |
428 | catch_vars[m.group('err')] = err.error if isinstance(err, JS_Throw) else err | |
429 | catch_vars = local_vars.new_child(catch_vars) | |
430 | err, pending = None, self.interpret_statement(sub_expr, catch_vars, allow_recursion) | |
431 | ||
432 | m = re.match(r'finally\s*\{', expr) | |
433 | if m: | |
434 | sub_expr, expr = self._separate_at_paren(expr[m.end() - 1:]) | |
435 | ret, should_abort = self.interpret_statement(sub_expr, local_vars, allow_recursion) | |
436 | if should_abort: | |
437 | return ret, True | |
438 | ||
439 | ret, should_abort = pending | |
440 | if should_abort: | |
441 | return ret, True | |
442 | ||
443 | if err: | |
444 | raise err | |
445 | ||
446 | elif md.get('for'): | |
447 | constructor, remaining = self._separate_at_paren(expr[m.end() - 1:]) | |
448 | if remaining.startswith('{'): | |
449 | body, expr = self._separate_at_paren(remaining) | |
450 | else: | |
451 | switch_m = re.match(r'switch\s*\(', remaining) # FIXME | |
452 | if switch_m: | |
453 | switch_val, remaining = self._separate_at_paren(remaining[switch_m.end() - 1:]) | |
454 | body, expr = self._separate_at_paren(remaining, '}') | |
455 | body = 'switch(%s){%s}' % (switch_val, body) | |
456 | else: | |
457 | body, expr = remaining, '' | |
458 | start, cndn, increment = self._separate(constructor, ';') | |
459 | self.interpret_expression(start, local_vars, allow_recursion) | |
460 | while True: | |
461 | if not _js_ternary(self.interpret_expression(cndn, local_vars, allow_recursion)): | |
462 | break | |
463 | try: | |
464 | ret, should_abort = self.interpret_statement(body, local_vars, allow_recursion) | |
465 | if should_abort: | |
466 | return ret, True | |
467 | except JS_Break: | |
468 | break | |
469 | except JS_Continue: | |
470 | pass | |
471 | self.interpret_expression(increment, local_vars, allow_recursion) | |
472 | ||
473 | elif md.get('switch'): | |
474 | switch_val, remaining = self._separate_at_paren(expr[m.end() - 1:]) | |
475 | switch_val = self.interpret_expression(switch_val, local_vars, allow_recursion) | |
476 | body, expr = self._separate_at_paren(remaining, '}') | |
477 | items = body.replace('default:', 'case default:').split('case ')[1:] | |
478 | for default in (False, True): | |
479 | matched = False | |
480 | for item in items: | |
481 | case, stmt = (i.strip() for i in self._separate(item, ':', 1)) | |
482 | if default: | |
483 | matched = matched or case == 'default' | |
484 | elif not matched: | |
485 | matched = (case != 'default' | |
486 | and switch_val == self.interpret_expression(case, local_vars, allow_recursion)) | |
487 | if not matched: | |
488 | continue | |
489 | try: | |
490 | ret, should_abort = self.interpret_statement(stmt, local_vars, allow_recursion) | |
491 | if should_abort: | |
492 | return ret | |
493 | except JS_Break: | |
494 | break | |
495 | if matched: | |
496 | break | |
497 | ||
498 | if md: | |
499 | ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion) | |
500 | return ret, should_abort or should_return | |
501 | ||
502 | # Comma separated statements | |
503 | sub_expressions = list(self._separate(expr)) | |
504 | if len(sub_expressions) > 1: | |
505 | for sub_expr in sub_expressions: | |
506 | ret, should_abort = self.interpret_statement(sub_expr, local_vars, allow_recursion) | |
507 | if should_abort: | |
508 | return ret, True | |
509 | return ret, False | |
510 | ||
511 | for m in re.finditer(rf'''(?x) | |
512 | (?P<pre_sign>\+\+|--)(?P<var1>{_NAME_RE})| | |
513 | (?P<var2>{_NAME_RE})(?P<post_sign>\+\+|--)''', expr): | |
514 | var = m.group('var1') or m.group('var2') | |
515 | start, end = m.span() | |
516 | sign = m.group('pre_sign') or m.group('post_sign') | |
517 | ret = local_vars[var] | |
518 | local_vars[var] += 1 if sign[0] == '+' else -1 | |
519 | if m.group('pre_sign'): | |
520 | ret = local_vars[var] | |
521 | expr = expr[:start] + self._dump(ret, local_vars) + expr[end:] | |
522 | ||
523 | if not expr: | |
524 | return None, should_return | |
525 | ||
526 | m = re.match(fr'''(?x) | |
527 | (?P<assign> | |
528 | (?P<out>{_NAME_RE})(?:\[(?P<index>[^\]]+?)\])?\s* | |
529 | (?P<op>{"|".join(map(re.escape, set(_OPERATORS) - _COMP_OPERATORS))})? | |
530 | =(?!=)(?P<expr>.*)$ | |
531 | )|(?P<return> | |
532 | (?!if|return|true|false|null|undefined|NaN)(?P<name>{_NAME_RE})$ | |
533 | )|(?P<indexing> | |
534 | (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$ | |
535 | )|(?P<attribute> | |
536 | (?P<var>{_NAME_RE})(?:(?P<nullish>\?)?\.(?P<member>[^(]+)|\[(?P<member2>[^\]]+)\])\s* | |
537 | )|(?P<function> | |
538 | (?P<fname>{_NAME_RE})\((?P<args>.*)\)$ | |
539 | )''', expr) | |
540 | if m and m.group('assign'): | |
541 | left_val = local_vars.get(m.group('out')) | |
542 | ||
543 | if not m.group('index'): | |
544 | local_vars[m.group('out')] = self._operator( | |
545 | m.group('op'), left_val, m.group('expr'), expr, local_vars, allow_recursion) | |
546 | return local_vars[m.group('out')], should_return | |
547 | elif left_val in (None, JS_Undefined): | |
548 | raise self.Exception(f'Cannot index undefined variable {m.group("out")}', expr) | |
549 | ||
550 | idx = self.interpret_expression(m.group('index'), local_vars, allow_recursion) | |
551 | if not isinstance(idx, (int, float)): | |
552 | raise self.Exception(f'List index {idx} must be integer', expr) | |
553 | idx = int(idx) | |
554 | left_val[idx] = self._operator( | |
555 | m.group('op'), self._index(left_val, idx), m.group('expr'), expr, local_vars, allow_recursion) | |
556 | return left_val[idx], should_return | |
557 | ||
558 | elif expr.isdigit(): | |
559 | return int(expr), should_return | |
560 | ||
561 | elif expr == 'break': | |
562 | raise JS_Break() | |
563 | elif expr == 'continue': | |
564 | raise JS_Continue() | |
565 | elif expr == 'undefined': | |
566 | return JS_Undefined, should_return | |
567 | elif expr == 'NaN': | |
568 | return float('NaN'), should_return | |
569 | ||
570 | elif m and m.group('return'): | |
571 | return local_vars.get(m.group('name'), JS_Undefined), should_return | |
572 | ||
573 | with contextlib.suppress(ValueError): | |
574 | return json.loads(js_to_json(expr, strict=True)), should_return | |
575 | ||
576 | if m and m.group('indexing'): | |
577 | val = local_vars[m.group('in')] | |
578 | idx = self.interpret_expression(m.group('idx'), local_vars, allow_recursion) | |
579 | return self._index(val, idx), should_return | |
580 | ||
581 | for op in _OPERATORS: | |
582 | separated = list(self._separate(expr, op)) | |
583 | right_expr = separated.pop() | |
584 | while True: | |
585 | if op in '?<>*-' and len(separated) > 1 and not separated[-1].strip(): | |
586 | separated.pop() | |
587 | elif not (separated and op == '?' and right_expr.startswith('.')): | |
588 | break | |
589 | right_expr = f'{op}{right_expr}' | |
590 | if op != '-': | |
591 | right_expr = f'{separated.pop()}{op}{right_expr}' | |
592 | if not separated: | |
593 | continue | |
594 | left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion) | |
595 | return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), should_return | |
596 | ||
597 | if m and m.group('attribute'): | |
598 | variable, member, nullish = m.group('var', 'member', 'nullish') | |
599 | if not member: | |
600 | member = self.interpret_expression(m.group('member2'), local_vars, allow_recursion) | |
601 | arg_str = expr[m.end():] | |
602 | if arg_str.startswith('('): | |
603 | arg_str, remaining = self._separate_at_paren(arg_str) | |
604 | else: | |
605 | arg_str, remaining = None, arg_str | |
606 | ||
607 | def assertion(cndn, msg): | |
608 | """ assert, but without risk of getting optimized out """ | |
609 | if not cndn: | |
610 | raise self.Exception(f'{member} {msg}', expr) | |
611 | ||
612 | def eval_method(): | |
613 | if (variable, member) == ('console', 'debug'): | |
614 | if Debugger.ENABLED: | |
615 | Debugger.write(self.interpret_expression(f'[{arg_str}]', local_vars, allow_recursion)) | |
616 | return | |
617 | ||
618 | types = { | |
619 | 'String': str, | |
620 | 'Math': float, | |
621 | } | |
622 | obj = local_vars.get(variable, types.get(variable, NO_DEFAULT)) | |
623 | if obj is NO_DEFAULT: | |
624 | if variable not in self._objects: | |
625 | try: | |
626 | self._objects[variable] = self.extract_object(variable) | |
627 | except self.Exception: | |
628 | if not nullish: | |
629 | raise | |
630 | obj = self._objects.get(variable, JS_Undefined) | |
631 | ||
632 | if nullish and obj is JS_Undefined: | |
633 | return JS_Undefined | |
634 | ||
635 | # Member access | |
636 | if arg_str is None: | |
637 | return self._index(obj, member, nullish) | |
638 | ||
639 | # Function call | |
640 | argvals = [ | |
641 | self.interpret_expression(v, local_vars, allow_recursion) | |
642 | for v in self._separate(arg_str)] | |
643 | ||
644 | if obj == str: | |
645 | if member == 'fromCharCode': | |
646 | assertion(argvals, 'takes one or more arguments') | |
647 | return ''.join(map(chr, argvals)) | |
648 | raise self.Exception(f'Unsupported String method {member}', expr) | |
649 | elif obj == float: | |
650 | if member == 'pow': | |
651 | assertion(len(argvals) == 2, 'takes two arguments') | |
652 | return argvals[0] ** argvals[1] | |
653 | raise self.Exception(f'Unsupported Math method {member}', expr) | |
654 | ||
655 | if member == 'split': | |
656 | assertion(argvals, 'takes one or more arguments') | |
657 | assertion(len(argvals) == 1, 'with limit argument is not implemented') | |
658 | return obj.split(argvals[0]) if argvals[0] else list(obj) | |
659 | elif member == 'join': | |
660 | assertion(isinstance(obj, list), 'must be applied on a list') | |
661 | assertion(len(argvals) == 1, 'takes exactly one argument') | |
662 | return argvals[0].join(obj) | |
663 | elif member == 'reverse': | |
664 | assertion(not argvals, 'does not take any arguments') | |
665 | obj.reverse() | |
666 | return obj | |
667 | elif member == 'slice': | |
668 | assertion(isinstance(obj, list), 'must be applied on a list') | |
669 | assertion(len(argvals) == 1, 'takes exactly one argument') | |
670 | return obj[argvals[0]:] | |
671 | elif member == 'splice': | |
672 | assertion(isinstance(obj, list), 'must be applied on a list') | |
673 | assertion(argvals, 'takes one or more arguments') | |
674 | index, howMany = map(int, (argvals + [len(obj)])[:2]) | |
675 | if index < 0: | |
676 | index += len(obj) | |
677 | add_items = argvals[2:] | |
678 | res = [] | |
679 | for i in range(index, min(index + howMany, len(obj))): | |
680 | res.append(obj.pop(index)) | |
681 | for i, item in enumerate(add_items): | |
682 | obj.insert(index + i, item) | |
683 | return res | |
684 | elif member == 'unshift': | |
685 | assertion(isinstance(obj, list), 'must be applied on a list') | |
686 | assertion(argvals, 'takes one or more arguments') | |
687 | for item in reversed(argvals): | |
688 | obj.insert(0, item) | |
689 | return obj | |
690 | elif member == 'pop': | |
691 | assertion(isinstance(obj, list), 'must be applied on a list') | |
692 | assertion(not argvals, 'does not take any arguments') | |
693 | if not obj: | |
694 | return | |
695 | return obj.pop() | |
696 | elif member == 'push': | |
697 | assertion(argvals, 'takes one or more arguments') | |
698 | obj.extend(argvals) | |
699 | return obj | |
700 | elif member == 'forEach': | |
701 | assertion(argvals, 'takes one or more arguments') | |
702 | assertion(len(argvals) <= 2, 'takes at-most 2 arguments') | |
703 | f, this = (argvals + [''])[:2] | |
704 | return [f((item, idx, obj), {'this': this}, allow_recursion) for idx, item in enumerate(obj)] | |
705 | elif member == 'indexOf': | |
706 | assertion(argvals, 'takes one or more arguments') | |
707 | assertion(len(argvals) <= 2, 'takes at-most 2 arguments') | |
708 | idx, start = (argvals + [0])[:2] | |
709 | try: | |
710 | return obj.index(idx, start) | |
711 | except ValueError: | |
712 | return -1 | |
713 | elif member == 'charCodeAt': | |
714 | assertion(isinstance(obj, str), 'must be applied on a string') | |
715 | assertion(len(argvals) == 1, 'takes exactly one argument') | |
716 | idx = argvals[0] if isinstance(argvals[0], int) else 0 | |
717 | if idx >= len(obj): | |
718 | return None | |
719 | return ord(obj[idx]) | |
720 | ||
721 | idx = int(member) if isinstance(obj, list) else member | |
722 | return obj[idx](argvals, allow_recursion=allow_recursion) | |
723 | ||
724 | if remaining: | |
725 | ret, should_abort = self.interpret_statement( | |
726 | self._named_object(local_vars, eval_method()) + remaining, | |
727 | local_vars, allow_recursion) | |
728 | return ret, should_return or should_abort | |
729 | else: | |
730 | return eval_method(), should_return | |
731 | ||
732 | elif m and m.group('function'): | |
733 | fname = m.group('fname') | |
734 | argvals = [self.interpret_expression(v, local_vars, allow_recursion) | |
735 | for v in self._separate(m.group('args'))] | |
736 | if fname in local_vars: | |
737 | return local_vars[fname](argvals, allow_recursion=allow_recursion), should_return | |
738 | elif fname not in self._functions: | |
739 | self._functions[fname] = self.extract_function(fname) | |
740 | return self._functions[fname](argvals, allow_recursion=allow_recursion), should_return | |
741 | ||
742 | raise self.Exception( | |
743 | f'Unsupported JS expression {truncate_string(expr, 20, 20) if expr != stmt else ""}', stmt) | |
744 | ||
745 | def interpret_expression(self, expr, local_vars, allow_recursion): | |
746 | ret, should_return = self.interpret_statement(expr, local_vars, allow_recursion) | |
747 | if should_return: | |
748 | raise self.Exception('Cannot return from an expression', expr) | |
749 | return ret | |
750 | ||
751 | def extract_object(self, objname): | |
752 | _FUNC_NAME_RE = r'''(?:[a-zA-Z$0-9]+|"[a-zA-Z$0-9]+"|'[a-zA-Z$0-9]+')''' | |
753 | obj = {} | |
754 | obj_m = re.search( | |
755 | r'''(?x) | |
756 | (?<!this\.)%s\s*=\s*{\s* | |
757 | (?P<fields>(%s\s*:\s*function\s*\(.*?\)\s*{.*?}(?:,\s*)?)*) | |
758 | }\s*; | |
759 | ''' % (re.escape(objname), _FUNC_NAME_RE), | |
760 | self.code) | |
761 | if not obj_m: | |
762 | raise self.Exception(f'Could not find object {objname}') | |
763 | fields = obj_m.group('fields') | |
764 | # Currently, it only supports function definitions | |
765 | fields_m = re.finditer( | |
766 | r'''(?x) | |
767 | (?P<key>%s)\s*:\s*function\s*\((?P<args>(?:%s|,)*)\){(?P<code>[^}]+)} | |
768 | ''' % (_FUNC_NAME_RE, _NAME_RE), | |
769 | fields) | |
770 | for f in fields_m: | |
771 | argnames = f.group('args').split(',') | |
772 | obj[remove_quotes(f.group('key'))] = self.build_function(argnames, f.group('code')) | |
773 | ||
774 | return obj | |
775 | ||
776 | def extract_function_code(self, funcname): | |
777 | """ @returns argnames, code """ | |
778 | func_m = re.search( | |
779 | r'''(?xs) | |
780 | (?: | |
781 | function\s+%(name)s| | |
782 | [{;,]\s*%(name)s\s*=\s*function| | |
783 | (?:var|const|let)\s+%(name)s\s*=\s*function | |
784 | )\s* | |
785 | \((?P<args>[^)]*)\)\s* | |
786 | (?P<code>{.+})''' % {'name': re.escape(funcname)}, | |
787 | self.code) | |
788 | code, _ = self._separate_at_paren(func_m.group('code')) | |
789 | if func_m is None: | |
790 | raise self.Exception(f'Could not find JS function "{funcname}"') | |
791 | return [x.strip() for x in func_m.group('args').split(',')], code | |
792 | ||
793 | def extract_function(self, funcname): | |
794 | return self.extract_function_from_code(*self.extract_function_code(funcname)) | |
795 | ||
796 | def extract_function_from_code(self, argnames, code, *global_stack): | |
797 | local_vars = {} | |
798 | while True: | |
799 | mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code) | |
800 | if mobj is None: | |
801 | break | |
802 | start, body_start = mobj.span() | |
803 | body, remaining = self._separate_at_paren(code[body_start - 1:]) | |
804 | name = self._named_object(local_vars, self.extract_function_from_code( | |
805 | [x.strip() for x in mobj.group('args').split(',')], | |
806 | body, local_vars, *global_stack)) | |
807 | code = code[:start] + name + remaining | |
808 | return self.build_function(argnames, code, local_vars, *global_stack) | |
809 | ||
810 | def call_function(self, funcname, *args): | |
811 | return self.extract_function(funcname)(args) | |
812 | ||
813 | def build_function(self, argnames, code, *global_stack): | |
814 | global_stack = list(global_stack) or [{}] | |
815 | argnames = tuple(argnames) | |
816 | ||
817 | def resf(args, kwargs={}, allow_recursion=100): | |
818 | global_stack[0].update(itertools.zip_longest(argnames, args, fillvalue=None)) | |
819 | global_stack[0].update(kwargs) | |
820 | var_stack = LocalNameSpace(*global_stack) | |
821 | ret, should_abort = self.interpret_statement(code.replace('\n', ' '), var_stack, allow_recursion - 1) | |
822 | if should_abort: | |
823 | return ret | |
824 | return resf |