From 8f53dc44a0cc1c2d98c35740b9293462c080f5d0 Mon Sep 17 00:00:00 2001 From: pukkandan Date: Sun, 14 Aug 2022 04:51:54 +0530 Subject: [PATCH] [jsinterp] Handle new youtube signature functions Closes #4635 --- test/test_jsinterp.py | 29 ++- test/test_utils.py | 4 + test/test_youtube_signature.py | 8 + yt_dlp/extractor/youtube.py | 3 +- yt_dlp/jsinterp.py | 339 ++++++++++++++++++++++----------- yt_dlp/utils.py | 29 ++- 6 files changed, 287 insertions(+), 125 deletions(-) diff --git a/test/test_jsinterp.py b/test/test_jsinterp.py index 4277cabe0..48e2abcf6 100644 --- a/test/test_jsinterp.py +++ b/test/test_jsinterp.py @@ -19,6 +19,9 @@ def test_basic(self): jsi = JSInterpreter('function x3(){return 42;}') self.assertEqual(jsi.call_function('x3'), 42) + jsi = JSInterpreter('function x3(){42}') + self.assertEqual(jsi.call_function('x3'), None) + jsi = JSInterpreter('var x5 = function(){return 42;}') self.assertEqual(jsi.call_function('x5'), 42) @@ -51,8 +54,11 @@ def test_operators(self): jsi = JSInterpreter('function f(){return 11 >> 2;}') self.assertEqual(jsi.call_function('f'), 2) + jsi = JSInterpreter('function f(){return []? 2+3: 4;}') + self.assertEqual(jsi.call_function('f'), 5) + def test_array_access(self): - jsi = JSInterpreter('function f(){var x = [1,2,3]; x[0] = 4; x[0] = 5; x[2] = 7; return x;}') + jsi = JSInterpreter('function f(){var x = [1,2,3]; x[0] = 4; x[0] = 5; x[2.0] = 7; return x;}') self.assertEqual(jsi.call_function('f'), [5, 2, 7]) def test_parens(self): @@ -62,6 +68,10 @@ def test_parens(self): jsi = JSInterpreter('function f(){return (1 + 2) * 3;}') self.assertEqual(jsi.call_function('f'), 9) + def test_quotes(self): + jsi = JSInterpreter(R'function f(){return "a\"\\("}') + self.assertEqual(jsi.call_function('f'), R'a"\(') + def test_assignments(self): jsi = JSInterpreter('function f(){var x = 20; x = 30 + 1; return x;}') self.assertEqual(jsi.call_function('f'), 31) @@ -107,14 +117,15 @@ def test_precedence(self): def test_call(self): jsi = JSInterpreter(''' function x() { return 2; } - function y(a) { return x() + a; } + function y(a) { return x() + (a?a:0); } function z() { return y(3); } ''') self.assertEqual(jsi.call_function('z'), 5) + self.assertEqual(jsi.call_function('y'), 2) def test_for_loop(self): jsi = JSInterpreter(''' - function x() { a=0; for (i=0; i-10; i++) {a++} a } + function x() { a=0; for (i=0; i-10; i++) {a++} return a } ''') self.assertEqual(jsi.call_function('x'), 10) @@ -155,19 +166,19 @@ def test_try(self): def test_for_loop_continue(self): jsi = JSInterpreter(''' - function x() { a=0; for (i=0; i-10; i++) { continue; a++ } a } + function x() { a=0; for (i=0; i-10; i++) { continue; a++ } return a } ''') self.assertEqual(jsi.call_function('x'), 0) def test_for_loop_break(self): jsi = JSInterpreter(''' - function x() { a=0; for (i=0; i-10; i++) { break; a++ } a } + function x() { a=0; for (i=0; i-10; i++) { break; a++ } return a } ''') self.assertEqual(jsi.call_function('x'), 0) def test_literal_list(self): jsi = JSInterpreter(''' - function x() { [1, 2, "asdf", [5, 6, 7]][3] } + function x() { return [1, 2, "asdf", [5, 6, 7]][3] } ''') self.assertEqual(jsi.call_function('x'), [5, 6, 7]) @@ -177,6 +188,12 @@ def test_comma(self): ''') self.assertEqual(jsi.call_function('x'), 7) + def test_return_function(self): + jsi = JSInterpreter(''' + function x() { return [1, function(){return 1}][1] } + ''') + self.assertEqual(jsi.call_function('x')([]), 1) + if __name__ == '__main__': unittest.main() diff --git a/test/test_utils.py b/test/test_utils.py index 659b071d3..67cd966d8 100644 --- a/test/test_utils.py +++ b/test/test_utils.py @@ -413,6 +413,10 @@ def test_unified_timestamps(self): self.assertEqual(unified_timestamp('December 15, 2017 at 7:49 am'), 1513324140) self.assertEqual(unified_timestamp('2018-03-14T08:32:43.1493874+00:00'), 1521016363) + self.assertEqual(unified_timestamp('December 31 1969 20:00:01 EDT'), 1) + self.assertEqual(unified_timestamp('Wednesday 31 December 1969 18:01:26 MDT'), 86) + self.assertEqual(unified_timestamp('12/31/1969 20:01:18 EDT', False), 78) + def test_determine_ext(self): self.assertEqual(determine_ext('http://example.com/foo/bar.mp4/?download'), 'mp4') self.assertEqual(determine_ext('http://example.com/foo/bar/?download', None), None) diff --git a/test/test_youtube_signature.py b/test/test_youtube_signature.py index 4fc2917e5..559bdfccf 100644 --- a/test/test_youtube_signature.py +++ b/test/test_youtube_signature.py @@ -94,6 +94,14 @@ 'https://www.youtube.com/s/player/5dd88d1d/player-plasma-ias-phone-en_US.vflset/base.js', 'kSxKFLeqzv_ZyHSAt', 'n8gS8oRlHOxPFA', ), + ( + 'https://www.youtube.com/s/player/324f67b9/player_ias.vflset/en_US/base.js', + 'xdftNy7dh9QGnhW', '22qLGxrmX8F1rA', + ), + ( + 'https://www.youtube.com/s/player/4c3f79c5/player_ias.vflset/en_US/base.js', + 'TDCstCG66tEAO5pR9o', 'dbxNtZ14c-yWyw', + ), ] diff --git a/yt_dlp/extractor/youtube.py b/yt_dlp/extractor/youtube.py index b59c8630a..ef289e48c 100644 --- a/yt_dlp/extractor/youtube.py +++ b/yt_dlp/extractor/youtube.py @@ -2653,7 +2653,8 @@ def _extract_n_function(self, video_id, player_url): if self.get_param('youtube_print_sig_code'): self.to_screen(f'Extracted nsig function from {player_id}:\n{func_code[1]}\n') - return lambda s: jsi.extract_function_from_code(*func_code)([s]) + func = jsi.extract_function_from_code(*func_code) + return lambda s: func([s]) def _extract_signature_timestamp(self, video_id, player_url, ytcfg=None, fatal=False): """ diff --git a/yt_dlp/jsinterp.py b/yt_dlp/jsinterp.py index e85371574..1af6ee0aa 100644 --- a/yt_dlp/jsinterp.py +++ b/yt_dlp/jsinterp.py @@ -1,29 +1,62 @@ import collections import contextlib +import itertools import json +import math import operator import re -from .utils import ExtractorError, remove_quotes, truncate_string +from .utils import ( + NO_DEFAULT, + ExtractorError, + js_to_json, + remove_quotes, + truncate_string, + unified_timestamp, + write_string, +) _NAME_RE = r'[a-zA-Z_$][\w$]*' -_OPERATORS = { +_OPERATORS = { # None => Defined in JSInterpreter._operator + '?': None, + + '||': None, + '&&': None, + '&': operator.and_, '|': operator.or_, '^': operator.xor, - '&': operator.and_, + + # FIXME: This should actually be below comparision '>>': operator.rshift, '<<': operator.lshift, - '-': operator.sub, + + '<=': operator.le, + '>=': operator.ge, + '<': operator.lt, + '>': operator.gt, + '+': operator.add, - '%': operator.mod, - '/': operator.truediv, + '-': operator.sub, + '*': operator.mul, + '/': operator.truediv, + '%': operator.mod, } _MATCHING_PARENS = dict(zip('({[', ')}]')) _QUOTES = '\'"' +def _ternary(cndn, if_true=True, if_false=False): + """Simulate JS's ternary operator (cndn?if_true:if_false)""" + if cndn in (False, None, 0, ''): + return if_false + with contextlib.suppress(TypeError): + if math.isnan(cndn): # NB: NaN cannot be checked by membership + return if_false + return if_true + + class JS_Break(ExtractorError): def __init__(self): ExtractorError.__init__(self, 'Invalid break') @@ -46,6 +79,27 @@ def __delitem__(self, key): raise NotImplementedError('Deleting is not supported') +class Debugger: + import sys + ENABLED = 'pytest' in sys.modules + + @staticmethod + def write(*args, level=100): + write_string(f'[debug] JS: {" " * (100 - level)}' + f'{" ".join(truncate_string(str(x), 50, 50) for x in args)}\n') + + @classmethod + def wrap_interpreter(cls, f): + def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs): + if cls.ENABLED and stmt.strip(): + cls.write(stmt, level=allow_recursion) + ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs) + if cls.ENABLED and stmt.strip(): + cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion) + return ret, should_ret + return interpret_statement + + class JSInterpreter: __named_object_counter = 0 @@ -56,7 +110,7 @@ def __init__(self, code, objects=None): class Exception(ExtractorError): def __init__(self, msg, expr=None, *args, **kwargs): if expr is not None: - msg += f' in: {truncate_string(expr, 50, 50)}' + msg = f'{msg.rstrip()} in: {truncate_string(expr, 50, 50)}' super().__init__(msg, *args, **kwargs) def _named_object(self, namespace, obj): @@ -73,9 +127,9 @@ def _separate(expr, delim=',', max_split=None): start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1 in_quote, escaping = None, False for idx, char in enumerate(expr): - if char in _MATCHING_PARENS: + if not in_quote and char in _MATCHING_PARENS: counters[_MATCHING_PARENS[char]] += 1 - elif char in counters: + elif not in_quote and char in counters: counters[char] -= 1 elif not escaping and char in _QUOTES and in_quote in (char, None): in_quote = None if in_quote else char @@ -101,50 +155,91 @@ def _separate_at_paren(cls, expr, delim): raise cls.Exception(f'No terminating paren {delim}', expr) return separated[0][1:].strip(), separated[1].strip() + def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion): + if op in ('||', '&&'): + if (op == '&&') ^ _ternary(left_val): + return left_val # short circuiting + elif op == '?': + right_expr = _ternary(left_val, *self._separate(right_expr, ':', 1)) + + right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) + if not _OPERATORS.get(op): + return right_val + + try: + return _OPERATORS[op](left_val, right_val) + except Exception as e: + raise self.Exception(f'Failed to evaluate {left_val!r} {op} {right_val!r}', expr, cause=e) + + def _index(self, obj, idx): + if idx == 'length': + return len(obj) + try: + return obj[int(idx)] if isinstance(obj, list) else obj[idx] + except Exception as e: + raise self.Exception(f'Cannot get index {idx}', repr(obj), cause=e) + + def _dump(self, obj, namespace): + try: + return json.dumps(obj) + except TypeError: + return self._named_object(namespace, obj) + + @Debugger.wrap_interpreter def interpret_statement(self, stmt, local_vars, allow_recursion=100): if allow_recursion < 0: raise self.Exception('Recursion limit reached') + allow_recursion -= 1 - should_abort = False + should_return = False sub_statements = list(self._separate(stmt, ';')) or [''] - stmt = sub_statements.pop().lstrip() + expr = stmt = sub_statements.pop().strip() for sub_stmt in sub_statements: - ret, should_abort = self.interpret_statement(sub_stmt, local_vars, allow_recursion - 1) - if should_abort: - return ret, should_abort + ret, should_return = self.interpret_statement(sub_stmt, local_vars, allow_recursion) + if should_return: + return ret, should_return m = re.match(r'(?Pvar\s)|return(?:\s+|$)', stmt) - if not m: # Try interpreting it as an expression - expr = stmt - elif m.group('var'): - expr = stmt[len(m.group(0)):] - else: - expr = stmt[len(m.group(0)):] - should_abort = True - - return self.interpret_expression(expr, local_vars, allow_recursion), should_abort - - def interpret_expression(self, expr, local_vars, allow_recursion): - expr = expr.strip() + if m: + expr = stmt[len(m.group(0)):].strip() + should_return = not m.group('var') if not expr: - return None + return None, should_return + + if expr[0] in _QUOTES: + inner, outer = self._separate(expr, expr[0], 1) + inner = json.loads(js_to_json(f'{inner}{expr[0]}', strict=True)) + if not outer: + return inner, should_return + expr = self._named_object(local_vars, inner) + outer + + if expr.startswith('new '): + obj = expr[4:] + if obj.startswith('Date('): + left, right = self._separate_at_paren(obj[4:], ')') + expr = unified_timestamp(left[1:-1], False) + if not expr: + raise self.Exception(f'Failed to parse date {left!r}', expr) + expr = self._dump(int(expr * 1000), local_vars) + right + else: + raise self.Exception(f'Unsupported object {obj}', expr) if expr.startswith('{'): inner, outer = self._separate_at_paren(expr, '}') - inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion - 1) + inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion) if not outer or should_abort: - return inner + return inner, should_abort or should_return else: - expr = json.dumps(inner) + outer + expr = self._dump(inner, local_vars) + outer if expr.startswith('('): inner, outer = self._separate_at_paren(expr, ')') - inner = self.interpret_expression(inner, local_vars, allow_recursion) - if not outer: - return inner + inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion) + if not outer or should_abort: + return inner, should_abort or should_return else: - expr = json.dumps(inner) + outer + expr = self._dump(inner, local_vars) + outer if expr.startswith('['): inner, outer = self._separate_at_paren(expr, ']') @@ -153,21 +248,23 @@ def interpret_expression(self, expr, local_vars, allow_recursion): for item in self._separate(inner)]) expr = name + outer - m = re.match(r'(?Ptry)\s*|(?:(?Pcatch)|(?Pfor)|(?Pswitch))\s*\(', expr) + m = re.match(r'(?Ptry|finally)\s*|(?:(?Pcatch)|(?Pfor)|(?Pswitch))\s*\(', expr) if m and m.group('try'): if expr[m.end()] == '{': try_expr, expr = self._separate_at_paren(expr[m.end():], '}') else: try_expr, expr = expr[m.end() - 1:], '' - ret, should_abort = self.interpret_statement(try_expr, local_vars, allow_recursion - 1) + ret, should_abort = self.interpret_statement(try_expr, local_vars, allow_recursion) if should_abort: - return ret - return self.interpret_statement(expr, local_vars, allow_recursion - 1)[0] + return ret, True + ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion) + return ret, should_abort or should_return elif m and m.group('catch'): # We ignore the catch block _, expr = self._separate_at_paren(expr, '}') - return self.interpret_statement(expr, local_vars, allow_recursion - 1)[0] + ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion) + return ret, should_abort or should_return elif m and m.group('for'): constructor, remaining = self._separate_at_paren(expr[m.end() - 1:], ')') @@ -182,22 +279,21 @@ def interpret_expression(self, expr, local_vars, allow_recursion): else: body, expr = remaining, '' start, cndn, increment = self._separate(constructor, ';') - if self.interpret_statement(start, local_vars, allow_recursion - 1)[1]: - raise self.Exception('Premature return in the initialization of a for loop', constructor) + self.interpret_expression(start, local_vars, allow_recursion) while True: - if not self.interpret_expression(cndn, local_vars, allow_recursion): + if not _ternary(self.interpret_expression(cndn, local_vars, allow_recursion)): break try: - ret, should_abort = self.interpret_statement(body, local_vars, allow_recursion - 1) + ret, should_abort = self.interpret_statement(body, local_vars, allow_recursion) if should_abort: - return ret + return ret, True except JS_Break: break except JS_Continue: pass - if self.interpret_statement(increment, local_vars, allow_recursion - 1)[1]: - raise self.Exception('Premature return in the initialization of a for loop', constructor) - return self.interpret_statement(expr, local_vars, allow_recursion - 1)[0] + self.interpret_expression(increment, local_vars, allow_recursion) + ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion) + return ret, should_abort or should_return elif m and m.group('switch'): switch_val, remaining = self._separate_at_paren(expr[m.end() - 1:], ')') @@ -215,20 +311,23 @@ def interpret_expression(self, expr, local_vars, allow_recursion): if not matched: continue try: - ret, should_abort = self.interpret_statement(stmt, local_vars, allow_recursion - 1) + ret, should_abort = self.interpret_statement(stmt, local_vars, allow_recursion) if should_abort: return ret except JS_Break: break if matched: break - return self.interpret_statement(expr, local_vars, allow_recursion - 1)[0] + ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion) + return ret, should_abort or should_return # Comma separated statements sub_expressions = list(self._separate(expr)) expr = sub_expressions.pop().strip() if sub_expressions else '' for sub_expr in sub_expressions: - self.interpret_expression(sub_expr, local_vars, allow_recursion) + ret, should_abort = self.interpret_statement(sub_expr, local_vars, allow_recursion) + if should_abort: + return ret, True for m in re.finditer(rf'''(?x) (?P\+\+|--)(?P{_NAME_RE})| @@ -240,10 +339,10 @@ def interpret_expression(self, expr, local_vars, allow_recursion): local_vars[var] += 1 if sign[0] == '+' else -1 if m.group('pre_sign'): ret = local_vars[var] - expr = expr[:start] + json.dumps(ret) + expr[end:] + expr = expr[:start] + self._dump(ret, local_vars) + expr[end:] if not expr: - return None + return None, should_return m = re.match(fr'''(?x) (?P @@ -251,36 +350,34 @@ def interpret_expression(self, expr, local_vars, allow_recursion): (?P{"|".join(map(re.escape, _OPERATORS))})? =(?P.*)$ )|(?P - (?!if|return|true|false|null)(?P{_NAME_RE})$ + (?!if|return|true|false|null|undefined)(?P{_NAME_RE})$ )|(?P (?P{_NAME_RE})\[(?P.+)\]$ )|(?P (?P{_NAME_RE})(?:\.(?P[^(]+)|\[(?P[^\]]+)\])\s* )|(?P - (?P{_NAME_RE})\((?P[\w$,]*)\)$ + (?P{_NAME_RE})\((?P.*)\)$ )''', expr) if m and m.group('assign'): - if not m.group('op'): - opfunc = lambda curr, right: right - else: - opfunc = _OPERATORS[m.group('op')] - right_val = self.interpret_expression(m.group('expr'), local_vars, allow_recursion) left_val = local_vars.get(m.group('out')) if not m.group('index'): - local_vars[m.group('out')] = opfunc(left_val, right_val) - return local_vars[m.group('out')] + local_vars[m.group('out')] = self._operator( + m.group('op'), left_val, m.group('expr'), expr, local_vars, allow_recursion) + return local_vars[m.group('out')], should_return elif left_val is None: raise self.Exception(f'Cannot index undefined variable {m.group("out")}', expr) idx = self.interpret_expression(m.group('index'), local_vars, allow_recursion) - if not isinstance(idx, int): + if not isinstance(idx, (int, float)): raise self.Exception(f'List index {idx} must be integer', expr) - left_val[idx] = opfunc(left_val[idx], right_val) - return left_val[idx] + idx = int(idx) + left_val[idx] = self._operator( + m.group('op'), left_val[idx], m.group('expr'), expr, local_vars, allow_recursion) + return left_val[idx], should_return elif expr.isdigit(): - return int(expr) + return int(expr), should_return elif expr == 'break': raise JS_Break() @@ -288,35 +385,33 @@ def interpret_expression(self, expr, local_vars, allow_recursion): raise JS_Continue() elif m and m.group('return'): - return local_vars[m.group('name')] + return local_vars[m.group('name')], should_return with contextlib.suppress(ValueError): - return json.loads(expr) + return json.loads(js_to_json(expr, strict=True)), should_return if m and m.group('indexing'): val = local_vars[m.group('in')] idx = self.interpret_expression(m.group('idx'), local_vars, allow_recursion) - return val[idx] + return self._index(val, idx), should_return - for op, opfunc in _OPERATORS.items(): + for op in _OPERATORS: separated = list(self._separate(expr, op)) if len(separated) < 2: continue - right_val = separated.pop() - left_val = op.join(separated) - left_val, should_abort = self.interpret_statement( - left_val, local_vars, allow_recursion - 1) - if should_abort: - raise self.Exception(f'Premature left-side return of {op}', expr) - right_val, should_abort = self.interpret_statement( - right_val, local_vars, allow_recursion - 1) - if should_abort: - raise self.Exception(f'Premature right-side return of {op}', expr) - return opfunc(left_val or 0, right_val) + right_expr = separated.pop() + while op == '-' and len(separated) > 1 and not separated[-1].strip(): + right_expr = f'-{right_expr}' + separated.pop() + left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion) + return self._operator(op, 0 if left_val is None else left_val, + right_expr, expr, local_vars, allow_recursion), should_return if m and m.group('attribute'): variable = m.group('var') - member = remove_quotes(m.group('member') or m.group('member2')) + member = m.group('member') + if not member: + member = self.interpret_expression(m.group('member2'), local_vars, allow_recursion) arg_str = expr[m.end():] if arg_str.startswith('('): arg_str, remaining = self._separate_at_paren(arg_str, ')') @@ -329,20 +424,24 @@ def assertion(cndn, msg): raise self.Exception(f'{member} {msg}', expr) def eval_method(): - if variable == 'String': - obj = str - elif variable in local_vars: - obj = local_vars[variable] - else: + if (variable, member) == ('console', 'debug'): + if Debugger.ENABLED: + Debugger.write(self.interpret_expression(f'[{arg_str}]', local_vars, allow_recursion)) + return + + types = { + 'String': str, + 'Math': float, + } + obj = local_vars.get(variable, types.get(variable, NO_DEFAULT)) + if obj is NO_DEFAULT: if variable not in self._objects: self._objects[variable] = self.extract_object(variable) obj = self._objects[variable] # Member access if arg_str is None: - if member == 'length': - return len(obj) - return obj[member] + return self._index(obj, member) # Function call argvals = [ @@ -353,12 +452,17 @@ def eval_method(): if member == 'fromCharCode': assertion(argvals, 'takes one or more arguments') return ''.join(map(chr, argvals)) - raise self.Exception(f'Unsupported string method {member}', expr) + raise self.Exception(f'Unsupported String method {member}', expr) + elif obj == float: + if member == 'pow': + assertion(len(argvals) == 2, 'takes two arguments') + return argvals[0] ** argvals[1] + raise self.Exception(f'Unsupported Math method {member}', expr) if member == 'split': assertion(argvals, 'takes one or more arguments') - assertion(argvals == [''], 'with arguments is not implemented') - return list(obj) + assertion(len(argvals) == 1, 'with limit argument is not implemented') + return obj.split(argvals[0]) if argvals[0] else list(obj) elif member == 'join': assertion(isinstance(obj, list), 'must be applied on a list') assertion(len(argvals) == 1, 'takes exactly one argument') @@ -404,7 +508,7 @@ def eval_method(): assertion(argvals, 'takes one or more arguments') assertion(len(argvals) <= 2, 'takes at-most 2 arguments') f, this = (argvals + [''])[:2] - return [f((item, idx, obj), this=this) for idx, item in enumerate(obj)] + return [f((item, idx, obj), {'this': this}, allow_recursion) for idx, item in enumerate(obj)] elif member == 'indexOf': assertion(argvals, 'takes one or more arguments') assertion(len(argvals) <= 2, 'takes at-most 2 arguments') @@ -414,27 +518,35 @@ def eval_method(): except ValueError: return -1 - return obj[int(member) if isinstance(obj, list) else member](argvals) + idx = int(member) if isinstance(obj, list) else member + return obj[idx](argvals, allow_recursion=allow_recursion) if remaining: - return self.interpret_expression( + ret, should_abort = self.interpret_statement( self._named_object(local_vars, eval_method()) + remaining, local_vars, allow_recursion) + return ret, should_return or should_abort else: - return eval_method() + return eval_method(), should_return elif m and m.group('function'): fname = m.group('fname') - argvals = tuple( - int(v) if v.isdigit() else local_vars[v] - for v in self._separate(m.group('args'))) + argvals = [self.interpret_expression(v, local_vars, allow_recursion) + for v in self._separate(m.group('args'))] if fname in local_vars: - return local_vars[fname](argvals) + return local_vars[fname](argvals, allow_recursion=allow_recursion), should_return elif fname not in self._functions: self._functions[fname] = self.extract_function(fname) - return self._functions[fname](argvals) + return self._functions[fname](argvals, allow_recursion=allow_recursion), should_return - raise self.Exception('Unsupported JS expression', expr) + raise self.Exception( + f'Unsupported JS expression {truncate_string(expr, 20, 20) if expr != stmt else ""}', stmt) + + def interpret_expression(self, expr, local_vars, allow_recursion): + ret, should_return = self.interpret_statement(expr, local_vars, allow_recursion) + if should_return: + raise self.Exception('Cannot return from an expression', expr) + return ret def extract_object(self, objname): _FUNC_NAME_RE = r'''(?:[a-zA-Z$0-9]+|"[a-zA-Z$0-9]+"|'[a-zA-Z$0-9]+')''' @@ -446,6 +558,8 @@ def extract_object(self, objname): }\s*; ''' % (re.escape(objname), _FUNC_NAME_RE), self.code) + if not obj_m: + raise self.Exception(f'Could not find object {objname}') fields = obj_m.group('fields') # Currently, it only supports function definitions fields_m = re.finditer( @@ -462,19 +576,19 @@ def extract_object(self, objname): def extract_function_code(self, funcname): """ @returns argnames, code """ func_m = re.search( - r'''(?x) + r'''(?xs) (?: function\s+%(name)s| [{;,]\s*%(name)s\s*=\s*function| var\s+%(name)s\s*=\s*function )\s* \((?P[^)]*)\)\s* - (?P{(?:(?!};)[^"]|"([^"]|\\")*")+})''' % {'name': re.escape(funcname)}, + (?P{.+})''' % {'name': re.escape(funcname)}, self.code) - code, _ = self._separate_at_paren(func_m.group('code'), '}') # refine the match + code, _ = self._separate_at_paren(func_m.group('code'), '}') if func_m is None: raise self.Exception(f'Could not find JS function "{funcname}"') - return func_m.group('args').split(','), code + return [x.strip() for x in func_m.group('args').split(',')], code def extract_function(self, funcname): return self.extract_function_from_code(*self.extract_function_code(funcname)) @@ -498,16 +612,15 @@ def call_function(self, funcname, *args): def build_function(self, argnames, code, *global_stack): global_stack = list(global_stack) or [{}] + argnames = tuple(argnames) - def resf(args, **kwargs): + def resf(args, kwargs={}, allow_recursion=100): global_stack[0].update({ - **dict(zip(argnames, args)), + **dict(itertools.zip_longest(argnames, args, fillvalue=None)), **kwargs }) var_stack = LocalNameSpace(*global_stack) - for stmt in self._separate(code.replace('\n', ''), ';'): - ret, should_abort = self.interpret_statement(stmt, var_stack) - if should_abort: - break - return ret + ret, should_abort = self.interpret_statement(code.replace('\n', ''), var_stack, allow_recursion - 1) + if should_abort: + return ret return resf diff --git a/yt_dlp/utils.py b/yt_dlp/utils.py index 17d6e7335..39a41d5b8 100644 --- a/yt_dlp/utils.py +++ b/yt_dlp/utils.py @@ -150,6 +150,16 @@ def random_user_agent(): 'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'], } +# From https://github.com/python/cpython/blob/3.11/Lib/email/_parseaddr.py#L36-L42 +TIMEZONE_NAMES = { + 'UT': 0, 'UTC': 0, 'GMT': 0, 'Z': 0, + 'AST': -4, 'ADT': -3, # Atlantic (used in Canada) + 'EST': -5, 'EDT': -4, # Eastern + 'CST': -6, 'CDT': -5, # Central + 'MST': -7, 'MDT': -6, # Mountain + 'PST': -8, 'PDT': -7 # Pacific +} + # needed for sanitizing filenames in restricted mode ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ', itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'], @@ -1684,7 +1694,11 @@ def extract_timezone(date_str): $) ''', date_str) if not m: - timezone = datetime.timedelta() + m = re.search(r'\d{1,2}:\d{1,2}(?:\.\d+)?(?P\s*[A-Z]+)$', date_str) + timezone = TIMEZONE_NAMES.get(m and m.group('tz').strip()) + if timezone is not None: + date_str = date_str[:-len(m.group('tz'))] + timezone = datetime.timedelta(hours=timezone or 0) else: date_str = date_str[:-len(m.group('tz'))] if not m.group('sign'): @@ -1746,7 +1760,8 @@ def unified_timestamp(date_str, day_first=True): if date_str is None: return None - date_str = re.sub(r'[,|]', '', date_str) + date_str = re.sub(r'\s+', ' ', re.sub( + r'(?i)[,|]|(mon|tues?|wed(nes)?|thu(rs)?|fri|sat(ur)?)(day)?', '', date_str)) pm_delta = 12 if re.search(r'(?i)PM', date_str) else 0 timezone, date_str = extract_timezone(date_str) @@ -1768,9 +1783,10 @@ def unified_timestamp(date_str, day_first=True): with contextlib.suppress(ValueError): dt = datetime.datetime.strptime(date_str, expression) - timezone + datetime.timedelta(hours=pm_delta) return calendar.timegm(dt.timetuple()) + timetuple = email.utils.parsedate_tz(date_str) if timetuple: - return calendar.timegm(timetuple) + pm_delta * 3600 + return calendar.timegm(timetuple) + pm_delta * 3600 - timezone.total_seconds() def determine_ext(url, default_ext='unknown_video'): @@ -3199,7 +3215,7 @@ def strip_jsonp(code): r'\g', code) -def js_to_json(code, vars={}): +def js_to_json(code, vars={}, *, strict=False): # vars is a dict of var, val pairs to substitute COMMENT_RE = r'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n' SKIP_RE = fr'\s*(?:{COMMENT_RE})?\s*' @@ -3233,14 +3249,17 @@ def fix_kv(m): if v in vars: return vars[v] + if strict: + raise ValueError(f'Unknown value: {v}') return '"%s"' % v def create_map(mobj): return json.dumps(dict(json.loads(js_to_json(mobj.group(1) or '[]', vars=vars)))) - code = re.sub(r'new Date\((".+")\)', r'\g<1>', code) code = re.sub(r'new Map\((\[.*?\])?\)', create_map, code) + if not strict: + code = re.sub(r'new Date\((".+")\)', r'\g<1>', code) return re.sub(r'''(?sx) "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|