[jsinterp] Overhaul JSInterp to handle new YT players 4c3f79c5, 324f67b9 (#31170)

* back-port from yt-dlp 8f53dc44a0cc1c2d98c35740b9293462c080f5d0, thanks pukkandan
* also support void, improve <</>> precedence, improve expressions in comma-list
* add more tests

test/test_jsinterp.py

@@ -19,6 +19,9 @@ class TestJSInterpreter(unittest.TestCase):
         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 @@ class TestJSInterpreter(unittest.TestCase):
         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 @@ class TestJSInterpreter(unittest.TestCase):
         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)
@@ -104,18 +114,29 @@ class TestJSInterpreter(unittest.TestCase):
         }''')
         self.assertEqual(jsi.call_function('x'), [20, 20, 30, 40, 50])
 
+    def test_builtins(self):
+        jsi = JSInterpreter('''
+        function x() { return new Date('Wednesday 31 December 1969 18:01:26 MDT') - 0; }
+        ''')
+        self.assertEqual(jsi.call_function('x'), 86000)
+        jsi = JSInterpreter('''
+        function x(dt) { return new Date(dt) - 0; }
+        ''')
+        self.assertEqual(jsi.call_function('x', 'Wednesday 31 December 1969 18:01:26 MDT'), 86000)
+
     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):
         # function x() { a=0; for (i=0; i-10; i++) {a++} a }
         jsi = JSInterpreter('''
-        function x() { a=0; for (i=0; i-10; i = i + 1) {a++} a }
+        function x() { a=0; for (i=0; i-10; i++) {a++} return a }
         ''')
         self.assertEqual(jsi.call_function('x'), 10)
 
@@ -156,19 +177,19 @@ class TestJSInterpreter(unittest.TestCase):
 
     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 +198,22 @@ class TestJSInterpreter(unittest.TestCase):
         function x() { a=5; a -= 1, a+=3; return a }
         ''')
         self.assertEqual(jsi.call_function('x'), 7)
+        jsi = JSInterpreter('''
+        function x() { a=5; return (a -= 1, a+=3, a); }
+        ''')
+        self.assertEqual(jsi.call_function('x'), 7)
+
+    def test_void(self):
+        jsi = JSInterpreter('''
+        function x() { return void 42; }
+        ''')
+        self.assertEqual(jsi.call_function('x'), None)
+
+    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__':

test/test_utils.py

@@ -370,6 +370,9 @@ class TestUtil(unittest.TestCase):
         self.assertEqual(unified_timestamp('Sep 11, 2013 | 5:49 AM'), 1378878540)
         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')

test/test_youtube_signature.py

@@ -90,12 +90,25 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/e06dea74/player_ias.vflset/en_US/base.js',
         'AiuodmaDDYw8d3y4bf', 'ankd8eza2T6Qmw',
     ),
+    (
+        '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',
+    ),
 ]
 
 
 class TestPlayerInfo(unittest.TestCase):
     def test_youtube_extract_player_info(self):
         PLAYER_URLS = (
+            ('https://www.youtube.com/s/player/4c3f79c5/player_ias.vflset/en_US/base.js', '4c3f79c5'),
             ('https://www.youtube.com/s/player/64dddad9/player_ias.vflset/en_US/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player_ias.vflset/fr_FR/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-en_US.vflset/base.js', '64dddad9'),

youtube_dl/compat.py

@@ -2985,7 +2985,6 @@ except ImportError:
     except ImportError:
         compat_filter = filter
 
-
 try:
     from future_builtins import zip as compat_zip
 except ImportError:  # not 2.6+ or is 3.x
@@ -2995,6 +2994,57 @@ except ImportError:  # not 2.6+ or is 3.x
         compat_zip = zip
 
 
+# method renamed between Py2/3
+try:
+    from itertools import zip_longest as compat_itertools_zip_longest
+except ImportError:
+    from itertools import izip_longest as compat_itertools_zip_longest
+
+
+# new class in collections
+try:
+    from collections import ChainMap as compat_collections_chain_map
+except ImportError:
+    # Py < 3.3
+    class compat_collections_chain_map(compat_collections_abc.MutableMapping):
+
+        maps = [{}]
+
+        def __init__(self, *maps):
+            self.maps = list(maps) or [{}]
+
+        def __getitem__(self, k):
+            for m in self.maps:
+                if k in m:
+                    return m[k]
+            raise KeyError(k)
+
+        def __setitem__(self, k, v):
+            self.maps[0].__setitem__(k, v)
+            return
+
+        def __delitem__(self, k):
+            if k in self.maps[0]:
+                del self.maps[0][k]
+                return
+            raise KeyError(k)
+
+        def __iter__(self):
+            return itertools.chain(*reversed(self.maps))
+
+        def __len__(self):
+            return len(iter(self))
+
+        def new_child(self, m=None, **kwargs):
+            m = m or {}
+            m.update(kwargs)
+            return compat_collections_chain_map(m, *self.maps)
+
+        @property
+        def parents(self):
+            return compat_collections_chain_map(*(self.maps[1:]))
+
+
 if sys.version_info < (3, 3):
     def compat_b64decode(s, *args, **kwargs):
         if isinstance(s, compat_str):
@@ -3031,6 +3081,7 @@ __all__ = [
     'compat_basestring',
     'compat_chr',
     'compat_collections_abc',
+    'compat_collections_chain_map',
     'compat_cookiejar',
     'compat_cookiejar_Cookie',
     'compat_cookies',
@@ -3051,6 +3102,7 @@ __all__ = [
     'compat_input',
     'compat_integer_types',
     'compat_itertools_count',
+    'compat_itertools_zip_longest',
     'compat_kwargs',
     'compat_map',
     'compat_numeric_types',

youtube_dl/jsinterp.py

@@ -1,42 +1,87 @@
 from __future__ import unicode_literals
 
+import itertools
 import json
+import math
 import operator
 import re
 
 from .utils import (
+    NO_DEFAULT,
     ExtractorError,
+    js_to_json,
     remove_quotes,
+    unified_timestamp,
 )
 from .compat import (
-    compat_collections_abc,
+    compat_collections_chain_map as ChainMap,
+    compat_itertools_zip_longest as zip_longest,
     compat_str,
 )
-MutableMapping = compat_collections_abc.MutableMapping
 
+_NAME_RE = r'[a-zA-Z_$][\w$]*'
 
-class Nonlocal:
+# (op, definition) in order of binding priority, tightest first
-    pass
+# avoid dict to maintain order
+# definition None => Defined in JSInterpreter._operator
+_DOT_OPERATORS = (
+    ('.', None),
+    # TODO: ('?.', None),
+)
 
-
+_OPERATORS = (
-_OPERATORS = [
     ('|', operator.or_),
     ('^', operator.xor),
     ('&', operator.and_),
     ('>>', operator.rshift),
     ('<<', operator.lshift),
-    ('-', operator.sub),
     ('+', operator.add),
-    ('%', operator.mod),
+    ('-', operator.sub),
-    ('/', operator.truediv),
     ('*', operator.mul),
-]
+    ('/', operator.truediv),
-_ASSIGN_OPERATORS = [(op + '=', opfunc) for op, opfunc in _OPERATORS]
+    ('%', operator.mod),
-_ASSIGN_OPERATORS.append(('=', (lambda cur, right: right)))
+)
 
-_NAME_RE = r'[a-zA-Z_$][a-zA-Z_$0-9]*'
+_COMP_OPERATORS = (
+    ('===', operator.is_),
+    ('==', operator.eq),
+    ('!==', operator.is_not),
+    ('!=', operator.ne),
+    ('<=', operator.le),
+    ('>=', operator.ge),
+    ('<', operator.lt),
+    ('>', operator.gt),
+)
+
+_LOG_OPERATORS = (
+    ('&', operator.and_),
+    ('|', operator.or_),
+    ('^', operator.xor),
+)
+
+_SC_OPERATORS = (
+    ('?', None),
+    ('||', None),
+    ('&&', None),
+    # TODO: ('??', None),
+)
+
+_OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS))
 
 _MATCHING_PARENS = dict(zip(*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
+    try:
+        if math.isnan(cndn):  # NB: NaN cannot be checked by membership
+            return if_false
+    except TypeError:
+        pass
+    return if_true
 
 
 class JS_Break(ExtractorError):
@@ -49,70 +94,77 @@ class JS_Continue(ExtractorError):
         ExtractorError.__init__(self, 'Invalid continue')
 
 
-class LocalNameSpace(MutableMapping):
+class LocalNameSpace(ChainMap):
-    def __init__(self, *stack):
-        self.stack = tuple(stack)
-
-    def __getitem__(self, key):
-        for scope in self.stack:
-            if key in scope:
-                return scope[key]
-        raise KeyError(key)
-
     def __setitem__(self, key, value):
-        for scope in self.stack:
+        for scope in self.maps:
             if key in scope:
                 scope[key] = value
-                break
+                return
-        else:
+        self.maps[0][key] = value
-            self.stack[0][key] = value
-        return value
 
     def __delitem__(self, key):
         raise NotImplementedError('Deleting is not supported')
 
-    def __iter__(self):
-        for scope in self.stack:
-            for scope_item in iter(scope):
-                yield scope_item
-
-    def __len__(self, key):
-        return len(iter(self))
-
     def __repr__(self):
-        return 'LocalNameSpace%s' % (self.stack, )
+        return 'LocalNameSpace%s' % (self.maps, )
 
 
 class JSInterpreter(object):
+    __named_object_counter = 0
+
     def __init__(self, code, objects=None):
-        if objects is None:
+        self.code, self._functions = code, {}
-            objects = {}
+        self._objects = {} if objects is None else objects
-        self.code = code
+
-        self._functions = {}
+    class Exception(ExtractorError):
-        self._objects = objects
+        def __init__(self, msg, *args, **kwargs):
-        self.__named_object_counter = 0
+            expr = kwargs.pop('expr', None)
+            if expr is not None:
+                msg = '{0} in: {1!r}'.format(msg.rstrip(), expr[:100])
+            super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs)
 
     def _named_object(self, namespace, obj):
         self.__named_object_counter += 1
-        name = '__youtube_dl_jsinterp_obj%s' % (self.__named_object_counter, )
+        name = '__youtube_dl_jsinterp_obj%d' % (self.__named_object_counter, )
         namespace[name] = obj
         return name
 
     @staticmethod
-    def _separate(expr, delim=',', max_split=None):
+    def _separate(expr, delim=',', max_split=None, skip_delims=None):
         if not expr:
             return
         counters = {k: 0 for k in _MATCHING_PARENS.values()}
-        start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
+        start, splits, pos, skipping, delim_len = 0, 0, 0, 0, len(delim) - 1
+        in_quote, escaping = None, False
         for idx, char in enumerate(expr):
+            if not in_quote:
                 if char in _MATCHING_PARENS:
                     counters[_MATCHING_PARENS[char]] += 1
                 elif char in counters:
                     counters[char] -= 1
-            if char != delim[pos] or any(counters.values()):
+            if not escaping:
-                pos = 0
+                if char in _QUOTES and in_quote in (char, None):
+                    in_quote = None if in_quote else char
+                else:
+                    escaping = in_quote and char == '\\'
+            else:
+                escaping = False
+
+            if char != delim[pos] or any(counters.values()) or in_quote:
+                pos = skipping = 0
                 continue
-            elif pos != delim_len:
+            elif skipping > 0:
+                skipping -= 1
+                continue
+            elif pos == 0 and skip_delims:
+                here = expr[idx:]
+                for s in skip_delims if isinstance(skip_delims, (list, tuple)) else [skip_delims]:
+                    if here.startswith(s) and s:
+                        skipping = len(s) - 1
+                        break
+                if skipping > 0:
+                    continue
+            if pos < delim_len:
                 pos += 1
                 continue
             yield expr[start: idx - delim_len]
@@ -122,61 +174,108 @@ class JSInterpreter(object):
                 break
         yield expr[start:]
 
-    @staticmethod
+    @classmethod
-    def _separate_at_paren(expr, delim):
+    def _separate_at_paren(cls, expr, delim):
-        separated = list(JSInterpreter._separate(expr, delim, 1))
+        separated = list(cls._separate(expr, delim, 1))
+
         if len(separated) < 2:
-            raise ExtractorError('No terminating paren {0} in {1}'.format(delim, expr))
+            raise cls.Exception('No terminating paren {delim} in {expr}'.format(**locals()))
         return separated[0][1:].strip(), separated[1].strip()
 
+    @staticmethod
+    def _all_operators():
+        return itertools.chain(
+            _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
+
+    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)
+        opfunc = op and next((v for k, v in self._all_operators() if k == op), None)
+        if not opfunc:
+            return right_val
+
+        try:
+            return opfunc(left_val, right_val)
+        except Exception as e:
+            raise self.Exception('Failed to evaluate {left_val!r} {op} {right_val!r}'.format(**locals()), 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('Cannot get index {idx}'.format(**locals()), expr=repr(obj), cause=e)
+
+    def _dump(self, obj, namespace):
+        try:
+            return json.dumps(obj)
+        except TypeError:
+            return self._named_object(namespace, obj)
+
     def interpret_statement(self, stmt, local_vars, allow_recursion=100):
         if allow_recursion < 0:
-            raise ExtractorError('Recursion limit reached')
+            raise self.Exception('Recursion limit reached')
+        allow_recursion -= 1
 
-        sub_statements = list(self._separate(stmt, ';'))
+        should_return = False
-        stmt = (sub_statements or ['']).pop()
+        sub_statements = list(self._separate(stmt, ';')) or ['']
+        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)
+            ret, should_return = self.interpret_statement(sub_stmt, local_vars, allow_recursion)
-            if should_abort:
+            if should_return:
-                return ret
+                return ret, should_return
 
-        should_abort = False
+        m = re.match(r'(?P<var>(?:var|const|let)\s)|return(?:\s+|$)', stmt)
-        stmt = stmt.lstrip()
+        if m:
-        stmt_m = re.match(r'var\s', stmt)
+            expr = stmt[len(m.group(0)):].strip()
-        if stmt_m:
+            should_return = not m.group('var')
-            expr = stmt[len(stmt_m.group(0)):]
+        if not expr:
+            return None, should_return
+
+        if expr[0] in _QUOTES:
+            inner, outer = self._separate(expr, expr[0], 1)
+            inner = json.loads(js_to_json(inner + expr[0]))  # , strict=True))
+            if not outer:
+                return inner, should_return
+            expr = self._named_object(local_vars, inner) + outer
+
+        if expr.startswith('new '):
+            obj = expr[4:]
+            if obj.startswith('Date('):
+                left, right = self._separate_at_paren(obj[4:], ')')
+                left = self.interpret_expression(left, local_vars, allow_recursion)
+                expr = unified_timestamp(left, False)
+                if not expr:
+                    raise self.Exception('Failed to parse date {left!r}'.format(**locals()), expr=expr)
+                expr = self._dump(int(expr * 1000), local_vars) + right
             else:
-            return_m = re.match(r'return(?:\s+|$)', stmt)
+                raise self.Exception('Unsupported object {obj}'.format(**locals()), expr=expr)
-            if return_m:
-                expr = stmt[len(return_m.group(0)):]
-                should_abort = True
-            else:
-                # Try interpreting it as an expression
-                expr = stmt
 
-        v = self.interpret_expression(expr, local_vars, allow_recursion)
+        if expr.startswith('void '):
-        return v, should_abort
+            left = self.interpret_expression(expr[5:], local_vars, allow_recursion)
-
+            return None, should_return
-    def interpret_expression(self, expr, local_vars, allow_recursion):
-        expr = expr.strip()
-        if expr == '':  # Empty expression
-            return None
 
         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)
+            inner, should_abort = self.interpret_statement(inner, local_vars, allow_recursion)
-            if not outer:
+            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, ']')
@@ -185,57 +284,53 @@ class JSInterpreter(object):
                 for item in self._separate(inner)])
             expr = name + outer
 
-        m = re.match(r'try\s*', expr)
+        m = re.match(r'(?P<try>try|finally)\s*|(?:(?P<catch>catch)|(?P<for>for)|(?P<switch>switch))\s*\(', expr)
-        if m:
+        md = m.groupdict() if m else {}
+        if md.get('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 ret, True
-            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
 
-        m = re.match(r'(?:(?P<catch>catch)|(?P<for>for)|(?P<switch>switch))\s*\(', expr)
+        elif md.get('catch'):
-        md = m.groupdict() if m else {}
-        if md.get('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 md.get('for'):
-            def raise_constructor_error(c):
-                raise ExtractorError(
-                    'Premature return in the initialization of a for loop in {0!r}'.format(c))
-
             constructor, remaining = self._separate_at_paren(expr[m.end() - 1:], ')')
             if remaining.startswith('{'):
                 body, expr = self._separate_at_paren(remaining, '}')
             else:
-                m = re.match(r'switch\s*\(', remaining)  # FIXME
+                switch_m = re.match(r'switch\s*\(', remaining)  # FIXME
-                if m:
+                if switch_m:
-                    switch_val, remaining = self._separate_at_paren(remaining[m.end() - 1:], ')')
+                    switch_val, remaining = self._separate_at_paren(remaining[switch_m.end() - 1:], ')')
                     body, expr = self._separate_at_paren(remaining, '}')
                     body = 'switch(%s){%s}' % (switch_val, body)
                 else:
                     body, expr = remaining, ''
             start, cndn, increment = self._separate(constructor, ';')
-            if self.interpret_statement(start, local_vars, allow_recursion - 1)[1]:
+            self.interpret_expression(start, local_vars, allow_recursion)
-                raise_constructor_error(constructor)
             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]:
+                self.interpret_expression(increment, local_vars, allow_recursion)
-                    raise_constructor_error(constructor)
+            ret, should_abort = self.interpret_statement(expr, local_vars, allow_recursion)
-            return self.interpret_statement(expr, local_vars, allow_recursion - 1)[0]
+            return ret, should_abort or should_return
 
         elif md.get('switch'):
             switch_val, remaining = self._separate_at_paren(expr[m.end() - 1:], ')')
@@ -245,7 +340,7 @@ class JSInterpreter(object):
             for default in (False, True):
                 matched = False
                 for item in items:
-                    case, stmt = [i.strip() for i in self._separate(item, ':', 1)]
+                    case, stmt = (i.strip() for i in self._separate(item, ':', 1))
                     if default:
                         matched = matched or case == 'default'
                     elif not matched:
@@ -254,24 +349,28 @@ class JSInterpreter(object):
                     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 ''
+        if len(sub_expressions) > 1:
             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
+            return ret, False
 
         for m in re.finditer(r'''(?x)
-                (?P<pre_sign>\+\+|--)(?P<var1>%(_NAME_RE)s)|
+                (?P<pre_sign>\+\+|--)(?P<var1>{_NAME_RE})|
-                (?P<var2>%(_NAME_RE)s)(?P<post_sign>\+\+|--)''' % globals(), expr):
+                (?P<var2>{_NAME_RE})(?P<post_sign>\+\+|--)'''.format(**globals()), expr):
             var = m.group('var1') or m.group('var2')
             start, end = m.span()
             sign = m.group('pre_sign') or m.group('post_sign')
@@ -279,85 +378,87 @@ class JSInterpreter(object):
             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, should_return
 
-        for op, opfunc in _ASSIGN_OPERATORS:
         m = re.match(r'''(?x)
-                (?P<out>%s)(?:\[(?P<index>[^\]]+?)\])?
+            (?P<assign>
-                \s*%s
+                (?P<out>{_NAME_RE})(?:\[(?P<index>[^\]]+?)\])?\s*
-                (?P<expr>.*)$''' % (_NAME_RE, re.escape(op)), expr)
+                (?P<op>{_OPERATOR_RE})?
-            if not m:
+                =(?P<expr>.*)$
-                continue
+            )|(?P<return>
-            right_val = self.interpret_expression(m.group('expr'), local_vars, allow_recursion)
+                (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
+            )|(?P<indexing>
+                (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
+            )|(?P<attribute>
+                (?P<var>{_NAME_RE})(?:\.(?P<member>[^(]+)|\[(?P<member2>[^\]]+)\])\s*
+            )|(?P<function>
+                (?P<fname>{_NAME_RE})\((?P<args>.*)\)$
+            )'''.format(**globals()), expr)
+        md = m.groupdict() if m else {}
+        if md.get('assign'):
+            left_val = local_vars.get(m.group('out'))
+
+            if not m.group('index'):
+                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('Cannot index undefined variable ' + m.group('out'), expr=expr)
 
-            if m.groupdict().get('index'):
-                lvar = local_vars[m.group('out')]
             idx = self.interpret_expression(m.group('index'), local_vars, allow_recursion)
-                if not isinstance(idx, int):
+            if not isinstance(idx, (int, float)):
-                    raise ExtractorError('List indices must be integers: %s' % (idx, ))
+                raise self.Exception('List index %s must be integer' % (idx, ), expr=expr)
-                cur = lvar[idx]
+            idx = int(idx)
-                val = opfunc(cur, right_val)
+            left_val[idx] = self._operator(
-                lvar[idx] = val
+                m.group('op'), left_val[idx], m.group('expr'), expr, local_vars, allow_recursion)
-                return val
+            return left_val[idx], should_return
-            else:
-                cur = local_vars.get(m.group('out'))
-                val = opfunc(cur, right_val)
-                local_vars[m.group('out')] = val
-                return val
 
-        if expr.isdigit():
+        elif expr.isdigit():
-            return int(expr)
+            return int(expr), should_return
 
-        if expr == 'break':
+        elif expr == 'break':
             raise JS_Break()
         elif expr == 'continue':
             raise JS_Continue()
 
-        var_m = re.match(
+        elif md.get('return'):
-            r'(?!if|return|true|false|null)(?P<name>%s)$' % _NAME_RE,
+            return local_vars[m.group('name')], should_return
-            expr)
-        if var_m:
-            return local_vars[var_m.group('name')]
 
         try:
-            return json.loads(expr)
+            ret = json.loads(js_to_json(expr))  # strict=True)
+            if not md.get('attribute'):
+                return ret, should_return
         except ValueError:
             pass
 
-        m = re.match(
+        if md.get('indexing'):
-            r'(?P<in>%s)\[(?P<idx>.+)\]$' % _NAME_RE, expr)
-        if m:
             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
 
-        def raise_expr_error(where, op, exp):
+        for op, _ in self._all_operators():
-            raise ExtractorError('Premature {0} return of {1} in {2!r}'.format(where, op, exp))
+            # hackety: </> have higher priority than <</>>, but don't confuse them
-
+            skip_delim = (op + op) if op in ('<', '>') else None
-        for op, opfunc in _OPERATORS:
+            separated = list(self._separate(expr, op, skip_delims=skip_delim))
-            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_expr_error('left-side', op, expr)
-            right_val, should_abort = self.interpret_statement(
-                right_val, local_vars, allow_recursion - 1)
-            if should_abort:
-                raise_expr_error('right-side', op, expr)
-            return opfunc(left_val or 0, right_val)
 
-        m = re.match(
+            right_expr = separated.pop()
-            r'(?P<var>%s)(?:\.(?P<member>[^(]+)|\[(?P<member2>[^]]+)\])\s*' % _NAME_RE,
+            while op == '-' and len(separated) > 1 and not separated[-1].strip():
-            expr)
+                right_expr = '-' + right_expr
-        if m:
+                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 md.get('attribute'):
             variable = m.group('var')
-            nl = Nonlocal()
+            member = m.group('member')
-
+            if not member:
-            nl.member = remove_quotes(m.group('member') or m.group('member2'))
+                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, ')')
@@ -367,25 +468,24 @@ class JSInterpreter(object):
             def assertion(cndn, msg):
                 """ assert, but without risk of getting optimized out """
                 if not cndn:
-                    raise ExtractorError('{0} {1}: {2}'.format(nl.member, msg, expr))
+                    raise ExtractorError('{member} {msg}'.format(**locals()), expr=expr)
 
             def eval_method():
-                # nonlocal member
+                if (variable, member) == ('console', 'debug'):
-                member = nl.member
+                    return
-                if variable == 'String':
+                types = {
-                    obj = compat_str
+                    'String': compat_str,
-                elif variable in local_vars:
+                    'Math': float,
-                    obj = local_vars[variable]
+                }
-                else:
+                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]
 
-                if arg_str is None:
                 # Member access
-                    if member == 'length':
+                if arg_str is None:
-                        return len(obj)
+                    return self._index(obj, member)
-                    return obj[member]
 
                 # Function call
                 argvals = [
@@ -396,12 +496,17 @@ class JSInterpreter(object):
                     if member == 'fromCharCode':
                         assertion(argvals, 'takes one or more arguments')
                         return ''.join(map(chr, argvals))
-                    raise ExtractorError('Unsupported string method %s' % (member, ))
+                    raise self.Exception('Unsupported string method ' + member, expr=expr)
+                elif obj == float:
+                    if member == 'pow':
+                        assertion(len(argvals) == 2, 'takes two arguments')
+                        return argvals[0] ** argvals[1]
+                    raise self.Exception('Unsupported Math method ' + member, expr=expr)
 
                 if member == 'split':
                     assertion(argvals, 'takes one or more arguments')
-                    assertion(argvals == [''], 'with arguments is not implemented')
+                    assertion(len(argvals) == 1, 'with limit argument is not implemented')
-                    return list(obj)
+                    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')
@@ -447,7 +552,7 @@ class JSInterpreter(object):
                     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')
@@ -457,32 +562,35 @@ class JSInterpreter(object):
                     except ValueError:
                         return -1
 
-                if isinstance(obj, list):
+                idx = int(member) if isinstance(obj, list) else member
-                    member = int(member)
+                return obj[idx](argvals, allow_recursion=allow_recursion)
-                    nl.member = member
-                return obj[member](argvals)
 
             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
 
-        m = re.match(r'^(?P<func>%s)\((?P<args>[a-zA-Z0-9_$,]*)\)$' % _NAME_RE, expr)
+        elif md.get('function'):
-        if m:
+            fname = m.group('fname')
-            fname = m.group('func')
+            argvals = [self.interpret_expression(v, local_vars, allow_recursion)
-            argvals = tuple([
+                       for v in self._separate(m.group('args'))]
-                int(v) if v.isdigit() else local_vars[v]
-                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
 
-        if expr:
+        raise self.Exception(
-            raise ExtractorError('Unsupported JS expression %r' % expr)
+            'Unsupported JS expression ' + (expr[:40] if expr != stmt else ''), expr=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]+')'''
@@ -494,15 +602,17 @@ class JSInterpreter(object):
                 }\s*;
             ''' % (re.escape(objname), _FUNC_NAME_RE),
             self.code)
+        if not obj_m:
+            raise self.Exception('Could not find object ' + objname)
         fields = obj_m.group('fields')
         # Currently, it only supports function definitions
         fields_m = re.finditer(
             r'''(?x)
-                (?P<key>%s)\s*:\s*function\s*\((?P<args>[a-z,]+)\){(?P<code>[^}]+)}
+                (?P<key>%s)\s*:\s*function\s*\((?P<args>(?:%s|,)*)\){(?P<code>[^}]+)}
-            ''' % _FUNC_NAME_RE,
+            ''' % (_FUNC_NAME_RE, _NAME_RE),
             fields)
         for f in fields_m:
-            argnames = f.group('args').split(',')
+            argnames = self.build_arglist(f.group('args'))
             obj[remove_quotes(f.group('key'))] = self.build_function(argnames, f.group('code'))
 
         return obj
@@ -510,15 +620,19 @@ class JSInterpreter(object):
     def extract_function_code(self, funcname):
         """ @returns argnames, code """
         func_m = re.search(
-            r'''(?x)
+            r'''(?xs)
-                (?:function\s+%(f_n)s|[{;,]\s*%(f_n)s\s*=\s*function|var\s+%(f_n)s\s*=\s*function)\s*
+                (?:
+                    function\s+%(name)s|
+                    [{;,]\s*%(name)s\s*=\s*function|
+                    (?:var|const|let)\s+%(name)s\s*=\s*function
+                )\s*
                 \((?P<args>[^)]*)\)\s*
-                (?P<code>\{(?:(?!};)[^"]|"([^"]|\\")*")+\})''' % {'f_n': re.escape(funcname), },
+                (?P<code>{.+})''' % {'name': re.escape(funcname)},
             self.code)
         code, _ = self._separate_at_paren(func_m.group('code'), '}')  # refine the match
         if func_m is None:
-            raise ExtractorError('Could not find JS function %r' % funcname)
+            raise self.Exception('Could not find JS function "{funcname}"'.format(**locals()))
-        return func_m.group('args').split(','), code
+        return self.build_arglist(func_m.group('args')), code
 
     def extract_function(self, funcname):
         return self.extract_function_from_code(*self.extract_function_code(funcname))
@@ -534,7 +648,7 @@ class JSInterpreter(object):
             name = self._named_object(
                 local_vars,
                 self.extract_function_from_code(
-                    [x.strip() for x in mobj.group('args').split(',')],
+                    self.build_arglist(mobj.group('args')),
                     body, local_vars, *global_stack))
             code = code[:start] + name + remaining
         return self.build_function(argnames, code, local_vars, *global_stack)
@@ -542,17 +656,22 @@ class JSInterpreter(object):
     def call_function(self, funcname, *args):
         return self.extract_function(funcname)(args)
 
+    @classmethod
+    def build_arglist(cls, arg_text):
+        if not arg_text:
+            return []
+        return list(filter(None, (x.strip() or None for x in cls._separate(arg_text))))
+
     def build_function(self, argnames, code, *global_stack):
         global_stack = list(global_stack) or [{}]
-        local_vars = global_stack.pop(0)
+        argnames = tuple(argnames)
 
-        def resf(args, **kwargs):
+        def resf(args, kwargs={}, allow_recursion=100):
-            local_vars.update(dict(zip(argnames, args)))
+            global_stack[0].update(
-            local_vars.update(kwargs)
+                zip_longest(argnames, args, fillvalue=None))
-            var_stack = LocalNameSpace(local_vars, *global_stack)
+            global_stack[0].update(kwargs)
-            for stmt in self._separate(code.replace('\n', ''), ';'):
+            var_stack = LocalNameSpace(*global_stack)
-                ret, should_abort = self.interpret_statement(stmt, var_stack)
+            ret, should_abort = self.interpret_statement(code.replace('\n', ''), var_stack, allow_recursion - 1)
             if should_abort:
-                    break
                 return ret
         return resf

youtube_dl/utils.py

@@ -1696,6 +1696,17 @@ MONTH_NAMES = {
         'juillet', 'août', 'septembre', 'octobre', 'novembre', 'décembre'],
 }
 
+# Timezone names for RFC2822 obs-zone
+# 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
+}
+
 KNOWN_EXTENSIONS = (
     'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
     'flv', 'f4v', 'f4a', 'f4b',
@@ -1735,12 +1746,17 @@ DATE_FORMATS = (
     '%b %dth %Y %I:%M',
     '%Y %m %d',
     '%Y-%m-%d',
+    '%Y.%m.%d.',
     '%Y/%m/%d',
     '%Y/%m/%d %H:%M',
     '%Y/%m/%d %H:%M:%S',
+    '%Y%m%d%H%M',
+    '%Y%m%d%H%M%S',
+    '%Y%m%d',
     '%Y-%m-%d %H:%M',
     '%Y-%m-%d %H:%M:%S',
     '%Y-%m-%d %H:%M:%S.%f',
+    '%Y-%m-%d %H:%M:%S:%f',
     '%d.%m.%Y %H:%M',
     '%d.%m.%Y %H.%M',
     '%Y-%m-%dT%H:%M:%SZ',
@@ -1753,6 +1769,7 @@ DATE_FORMATS = (
     '%b %d %Y at %H:%M:%S',
     '%B %d %Y at %H:%M',
     '%B %d %Y at %H:%M:%S',
+    '%H:%M %d-%b-%Y',
 )
 
 DATE_FORMATS_DAY_FIRST = list(DATE_FORMATS)
@@ -1763,6 +1780,7 @@ DATE_FORMATS_DAY_FIRST.extend([
     '%d/%m/%Y',
     '%d/%m/%y',
     '%d/%m/%Y %H:%M:%S',
+    '%d-%m-%Y %H:%M',
 ])
 
 DATE_FORMATS_MONTH_FIRST = list(DATE_FORMATS)
@@ -2966,10 +2984,22 @@ class YoutubeDLRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
 
 def extract_timezone(date_str):
     m = re.search(
-        r'^.{8,}?(?P<tz>Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
+        r'''(?x)
-        date_str)
+            ^.{8,}?                                              # >=8 char non-TZ prefix, if present
+            (?P<tz>Z|                                            # just the UTC Z, or
+                (?:(?<=.\b\d{4}|\b\d{2}:\d\d)|                   # preceded by 4 digits or hh:mm or
+                   (?<!.\b[a-zA-Z]{3}|[a-zA-Z]{4}|..\b\d\d))     # not preceded by 3 alpha word or >= 4 alpha or 2 digits
+                   [ ]?                                          # optional space
+                (?P<sign>\+|-)                                   # +/-
+                (?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})       # hh[:]mm
+            $)
+        ''', date_str)
     if not m:
-        timezone = datetime.timedelta()
+        m = re.search(r'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', 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'):
@@ -3037,7 +3067,8 @@ 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)
@@ -3063,7 +3094,7 @@ def unified_timestamp(date_str, day_first=True):
             pass
     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'):
@@ -3673,13 +3704,11 @@ def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
     if get_attr:
         if v is not None:
             v = getattr(v, get_attr, None)
-    if v == '':
+    if v in (None, ''):
-        v = None
-    if v is None:
         return default
     try:
         return int(v) * invscale // scale
-    except (ValueError, TypeError):
+    except (ValueError, TypeError, OverflowError):
         return default