[YouTube] Simplify pattern for nsig function name extraction

* fix signature code extraction
* raise if n function returns input value
* add new tests from yt-dlp

Co-authored-by: bashonly

test/test_jsinterp.py

@@ -160,7 +160,6 @@ class TestJSInterpreter(unittest.TestCase):
         self._test('function f(){var x = 20; x += 30 + 1; return x;}', 51)
         self._test('function f(){var x = 20; x -= 30 + 1; return x;}', -11)
 
-    @unittest.skip('Not yet fully implemented')
     def test_comments(self):
         self._test('''
             function f() {
@@ -179,6 +178,15 @@ class TestJSInterpreter(unittest.TestCase):
             }
         ''', 3)
 
+        self._test('''
+            function f() {
+                var x = ( /* 1 + */ 2 +
+                          /* 30 * 40 */
+                          50);
+                return x;
+            }
+        ''', 52)
+
     def test_precedence(self):
         self._test('''
             function f() {
@@ -483,6 +491,13 @@ class TestJSInterpreter(unittest.TestCase):
             self._test(jsi, 't-e-s-t', args=[test_input, '-'])
             self._test(jsi, '', args=[[], '-'])
 
+        self._test('function f(){return '
+                   '[1, 1.0, "abc", {a: 1}, null, undefined, Infinity, NaN].join()}',
+                   '1,1,abc,[object Object],,,Infinity,NaN')
+        self._test('function f(){return '
+                   '[1, 1.0, "abc", {a: 1}, null, undefined, Infinity, NaN].join("~")}',
+                   '1~1~abc~[object Object]~~~Infinity~NaN')
+
     def test_split(self):
         test_result = list('test')
         tests = [
@@ -496,6 +511,18 @@ class TestJSInterpreter(unittest.TestCase):
             self._test(jsi, test_result, args=['t-e-s-t', '-'])
             self._test(jsi, [''], args=['', '-'])
             self._test(jsi, [], args=['', ''])
+        # RegExp split
+        self._test('function f(){return "test".split(/(?:)/)}',
+                   ['t', 'e', 's', 't'])
+        self._test('function f(){return "t-e-s-t".split(/[es-]+/)}',
+                   ['t', 't'])
+        # from MDN: surrogate pairs aren't handled: case 1 fails
+        # self._test('function f(){return "😄😄".split(/(?:)/)}',
+        #            ['\ud83d', '\ude04', '\ud83d', '\ude04'])
+        # case 2 beats Py3.2: it gets the case 1 result
+        if sys.version_info >= (2, 6) and not ((3, 0) <= sys.version_info < (3, 3)):
+            self._test('function f(){return "😄😄".split(/(?:)/u)}',
+                       ['😄', '😄'])
 
     def test_slice(self):
         self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice()}', [0, 1, 2, 3, 4, 5, 6, 7, 8])
@@ -525,6 +552,40 @@ class TestJSInterpreter(unittest.TestCase):
         self._test('function f(){return "012345678".slice(-1, 1)}', '')
         self._test('function f(){return "012345678".slice(-3, -1)}', '67')
 
+    def test_pop(self):
+        # pop
+        self._test('function f(){var a = [0, 1, 2, 3, 4, 5, 6, 7, 8]; return [a.pop(), a]}',
+                   [8, [0, 1, 2, 3, 4, 5, 6, 7]])
+        self._test('function f(){return [].pop()}', JS_Undefined)
+        # push
+        self._test('function f(){var a = [0, 1, 2]; return [a.push(3, 4), a]}',
+                   [5, [0, 1, 2, 3, 4]])
+        self._test('function f(){var a = [0, 1, 2]; return [a.push(), a]}',
+                   [3, [0, 1, 2]])
+
+    def test_shift(self):
+        # shift
+        self._test('function f(){var a = [0, 1, 2, 3, 4, 5, 6, 7, 8]; return [a.shift(), a]}',
+                   [0, [1, 2, 3, 4, 5, 6, 7, 8]])
+        self._test('function f(){return [].shift()}', JS_Undefined)
+        # unshift
+        self._test('function f(){var a = [0, 1, 2]; return [a.unshift(3, 4), a]}',
+                   [5, [3, 4, 0, 1, 2]])
+        self._test('function f(){var a = [0, 1, 2]; return [a.unshift(), a]}',
+                   [3, [0, 1, 2]])
+
+    def test_forEach(self):
+        self._test('function f(){var ret = []; var l = [4, 2]; '
+                   'var log = function(e,i,a){ret.push([e,i,a]);}; '
+                   'l.forEach(log); '
+                   'return [ret.length, ret[0][0], ret[1][1], ret[0][2]]}',
+                   [2, 4, 1, [4, 2]])
+        self._test('function f(){var ret = []; var l = [4, 2]; '
+                   'var log = function(e,i,a){this.push([e,i,a]);}; '
+                   'l.forEach(log, ret); '
+                   'return [ret.length, ret[0][0], ret[1][1], ret[0][2]]}',
+                   [2, 4, 1, [4, 2]])
+
 
 if __name__ == '__main__':
     unittest.main()

test/test_youtube_signature.py

@@ -1,4 +1,5 @@
 #!/usr/bin/env python
+# coding: utf-8
 
 from __future__ import unicode_literals
 
@@ -12,6 +13,7 @@ import re
 import string
 
 from youtube_dl.compat import (
+    compat_contextlib_suppress,
     compat_open as open,
     compat_str,
     compat_urlretrieve,
@@ -50,23 +52,33 @@ _SIG_TESTS = [
     (
         'https://s.ytimg.com/yts/jsbin/html5player-en_US-vflBb0OQx.js',
         84,
-        '123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQ0STUVWXYZ!"#$%&\'()*+,@./:;<=>'
+        '123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQ0STUVWXYZ!"#$%&\'()*+,@./:;<=>',
     ),
     (
         'https://s.ytimg.com/yts/jsbin/html5player-en_US-vfl9FYC6l.js',
         83,
-        '123456789abcdefghijklmnopqr0tuvwxyzABCDETGHIJKLMNOPQRS>UVWXYZ!"#$%&\'()*+,-./:;<=F'
+        '123456789abcdefghijklmnopqr0tuvwxyzABCDETGHIJKLMNOPQRS>UVWXYZ!"#$%&\'()*+,-./:;<=F',
     ),
     (
         'https://s.ytimg.com/yts/jsbin/html5player-en_US-vflCGk6yw/html5player.js',
         '4646B5181C6C3020DF1D9C7FCFEA.AD80ABF70C39BD369CCCAE780AFBB98FA6B6CB42766249D9488C288',
-        '82C8849D94266724DC6B6AF89BBFA087EACCD963.B93C07FBA084ACAEFCF7C9D1FD0203C6C1815B6B'
+        '82C8849D94266724DC6B6AF89BBFA087EACCD963.B93C07FBA084ACAEFCF7C9D1FD0203C6C1815B6B',
     ),
     (
         'https://s.ytimg.com/yts/jsbin/html5player-en_US-vflKjOTVq/html5player.js',
         '312AA52209E3623129A412D56A40F11CB0AF14AE.3EE09501CB14E3BCDC3B2AE808BF3F1D14E7FBF12',
         '112AA5220913623229A412D56A40F11CB0AF14AE.3EE0950FCB14EEBCDC3B2AE808BF331D14E7FBF3',
-    )
+    ),
+    (
+        'https://www.youtube.com/s/player/6ed0d907/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'AOq0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xx8j7v1pDL2QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0',
+    ),
+    (
+        'https://www.youtube.com/s/player/3bb1f723/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'MyOSJXtKI3m-uME_jv7-pT12gOFC02RFkGoqWpzE0Cs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+    ),
 ]
 
 _NSIG_TESTS = [
@@ -142,6 +154,10 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/5a3b6271/player_ias.vflset/en_US/base.js',
         'B2j7f_UPT4rfje85Lu_e', 'm5DmNymaGQ5RdQ',
     ),
+    (
+        'https://www.youtube.com/s/player/7a062b77/player_ias.vflset/en_US/base.js',
+        'NRcE3y3mVtm_cV-W', 'VbsCYUATvqlt5w',
+    ),
     (
         'https://www.youtube.com/s/player/dac945fd/player_ias.vflset/en_US/base.js',
         'o8BkRxXhuYsBCWi6RplPdP', '3Lx32v_hmzTm6A',
@@ -154,6 +170,10 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/cfa9e7cb/player_ias.vflset/en_US/base.js',
         'qO0NiMtYQ7TeJnfFG2', 'k9cuJDHNS5O7kQ',
     ),
+    (
+        'https://www.youtube.com/s/player/8c7583ff/player_ias.vflset/en_US/base.js',
+        '1wWCVpRR96eAmMI87L', 'KSkWAVv1ZQxC3A',
+    ),
     (
         'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js',
         '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ',
@@ -182,6 +202,14 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/b12cc44b/player_ias.vflset/en_US/base.js',
         'keLa5R2U00sR9SQK', 'N1OGyujjEwMnLw',
     ),
+    (
+        'https://www.youtube.com/s/player/3bb1f723/player_ias.vflset/en_US/base.js',
+        'gK15nzVyaXE9RsMP3z', 'ZFFWFLPWx9DEgQ',
+    ),
+    (
+        'https://www.youtube.com/s/player/f8f53e1a/player_ias.vflset/en_US/base.js',
+        'VTQOUOv0mCIeJ7i8kZB', 'kcfD8wy0sNLyNQ',
+    ),
 ]
 
 
@@ -216,11 +244,9 @@ class TestSignature(unittest.TestCase):
             os.mkdir(self.TESTDATA_DIR)
 
     def tearDown(self):
-        try:
+        with compat_contextlib_suppress(OSError):
             for f in os.listdir(self.TESTDATA_DIR):
                 os.remove(f)
-        except OSError:
-            pass
 
 
 def t_factory(name, sig_func, url_pattern):
@@ -254,11 +280,12 @@ def signature(jscode, sig_input):
 
 def n_sig(jscode, sig_input):
     funcname = YoutubeIE(FakeYDL())._extract_n_function_name(jscode)
-    return JSInterpreter(jscode).call_function(funcname, sig_input)
+    return JSInterpreter(jscode).call_function(
+        funcname, sig_input, _ytdl_do_not_return=sig_input)
 
 
 make_sig_test = t_factory(
-    'signature', signature, re.compile(r'.*-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.[a-z]+$'))
+    'signature', signature, re.compile(r'.*(?:-|/player/)(?P<id>[a-zA-Z0-9_-]+)(?:/.+\.js|(?:/watch_as3|/html5player)?\.[a-z]+)$'))
 for test_spec in _SIG_TESTS:
     make_sig_test(*test_spec)
 

youtube_dl/extractor/youtube.py

@@ -1579,19 +1579,26 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         self.to_screen('Extracted signature function:\n' + code)
 
     def _parse_sig_js(self, jscode):
+        # Examples where `sig` is funcname:
+        # sig=function(a){a=a.split(""); ... ;return a.join("")};
+        # ;c&&(c=sig(decodeURIComponent(c)),a.set(b,encodeURIComponent(c)));return a};
+        # {var l=f,m=h.sp,n=sig(decodeURIComponent(h.s));l.set(m,encodeURIComponent(n))}
+        # sig=function(J){J=J.split(""); ... ;return J.join("")};
+        # ;N&&(N=sig(decodeURIComponent(N)),J.set(R,encodeURIComponent(N)));return J};
+        # {var H=u,k=f.sp,v=sig(decodeURIComponent(f.s));H.set(k,encodeURIComponent(v))}
         funcname = self._search_regex(
-            (r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P<sig>[a-zA-Z0-9$]+)\(',
+            (r'\b(?P<var>[a-zA-Z0-9$]+)&&\((?P=var)=(?P<sig>[a-zA-Z0-9$]{2,})\(decodeURIComponent\((?P=var)\)\)',
+             r'(?P<sig>[a-zA-Z0-9$]+)\s*=\s*function\(\s*(?P<arg>[a-zA-Z0-9$]+)\s*\)\s*{\s*(?P=arg)\s*=\s*(?P=arg)\.split\(\s*""\s*\)\s*;\s*[^}]+;\s*return\s+(?P=arg)\.join\(\s*""\s*\)',
+             r'(?:\b|[^a-zA-Z0-9$])(?P<sig>[a-zA-Z0-9$]{2,})\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)(?:;[a-zA-Z0-9$]{2}\.[a-zA-Z0-9$]{2}\(a,\d+\))?',
+             # Old patterns
+             r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P<sig>[a-zA-Z0-9$]+)\(',
              r'\b[a-zA-Z0-9]+\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*encodeURIComponent\s*\(\s*(?P<sig>[a-zA-Z0-9$]+)\(',
              r'\bm=(?P<sig>[a-zA-Z0-9$]{2,})\(decodeURIComponent\(h\.s\)\)',
-             r'\bc&&\(c=(?P<sig>[a-zA-Z0-9$]{2,})\(decodeURIComponent\(c\)\)',
-             r'(?:\b|[^a-zA-Z0-9$])(?P<sig>[a-zA-Z0-9$]{2,})\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)(?:;[a-zA-Z0-9$]{2}\.[a-zA-Z0-9$]{2}\(a,\d+\))?',
-             r'(?P<sig>[a-zA-Z0-9$]+)\s*=\s*function\(\s*a\s*\)\s*{\s*a\s*=\s*a\.split\(\s*""\s*\)',
              # Obsolete patterns
              r'("|\')signature\1\s*,\s*(?P<sig>[a-zA-Z0-9$]+)\(',
              r'\.sig\|\|(?P<sig>[a-zA-Z0-9$]+)\(',
              r'yt\.akamaized\.net/\)\s*\|\|\s*.*?\s*[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?:encodeURIComponent\s*\()?\s*(?P<sig>[a-zA-Z0-9$]+)\(',
              r'\b[cs]\s*&&\s*[adf]\.set\([^,]+\s*,\s*(?P<sig>[a-zA-Z0-9$]+)\(',
-             r'\b[a-zA-Z0-9]+\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*(?P<sig>[a-zA-Z0-9$]+)\(',
              r'\bc\s*&&\s*[a-zA-Z0-9]+\.set\([^,]+\s*,\s*\([^)]*\)\s*\(\s*(?P<sig>[a-zA-Z0-9$]+)\('),
             jscode, 'Initial JS player signature function name', group='sig')
 
@@ -1658,36 +1665,29 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
     def _extract_n_function_name(self, jscode):
         func_name, idx = self._search_regex(
-            # new: (b=String.fromCharCode(110),c=a.get(b))&&c=nfunc[idx](c)
+            # (y=NuD(),Mw(k),q=k.Z[y]||null)&&(q=narray[idx](q),k.set(y,q),k.V||NuD(''))}};
-            # or:  (b="nn"[+a.D],c=a.get(b))&&(c=nfunc[idx](c)
+            # (R="nn"[+J.Z],mW(J),N=J.K[R]||null)&&(N=narray[idx](N),J.set(R,N))}};
-            # or:  (PL(a),b=a.j.n||null)&&(b=nfunc[idx](b)
+            # or:  (b=String.fromCharCode(110),c=a.get(b))&&c=narray[idx](c)
+            # or:  (b="nn"[+a.D],c=a.get(b))&&(c=narray[idx](c)
+            # or:  (PL(a),b=a.j.n||null)&&(b=narray[idx](b)
             # or:  (b="nn"[+a.D],vL(a),c=a.j[b]||null)&&(c=narray[idx](c),a.set(b,c),narray.length||nfunc("")
-            # old: (b=a.get("n"))&&(b=nfunc[idx](b)(?P<c>[a-z])\s*=\s*[a-z]\s*
+            # old: (b=a.get("n"))&&(b=narray[idx](b)(?P<c>[a-z])\s*=\s*[a-z]\s*
             # older: (b=a.get("n"))&&(b=nfunc(b)
             r'''(?x)
-                \((?:[\w$()\s]+,)*?\s*      # (
+                # (expr, ...,
-                (?P<b>[a-z])\s*=\s*         # b=
+                \((?:(?:\s*[\w$]+\s*=)?(?:[\w$"+\.\s(\[]+(?:[)\]]\s*)?),)*
-                (?:
+                  # b=...
-                    (?:                     # expect ,c=a.get(b) (etc)
+                  (?P<b>[\w$]+)\s*=\s*(?!(?P=b)[^\w$])[\w$]+\s*(?:(?:
-                        String\s*\.\s*fromCharCode\s*\(\s*110\s*\)|
+                    \.\s*[\w$]+ |
-                        "n+"\[\s*\+?s*[\w$.]+\s*]
+                    \[\s*[\w$]+\s*\] |
-                    )\s*(?:,[\w$()\s]+(?=,))*|
+                    \.\s*get\s*\(\s*[\w$"]+\s*\)
-                       (?P<old>[\w$]+)      # a (old[er])
+                  )\s*){,2}(?:\s*\|\|\s*null(?=\s*\)))?\s*
-                   )\s*
+                \)\s*&&\s*\(        # ...)&&(
-                   (?(old)
+                # b = nfunc, b = narray[idx]
-                                            # b.get("n")
+                (?P=b)\s*=\s*(?P<nfunc>[\w$]+)\s*
-                       (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*?
+                    (?:\[\s*(?P<idx>[\w$]+)\s*\]\s*)?
-                       (?:\.\s*n|\[\s*"n"\s*]|\.\s*get\s*\(\s*"n"\s*\))
+                    # (...)
-                       |                    # ,c=a.get(b)
+                    \(\s*[\w$]+\s*\)
-                       ,\s*(?P<c>[a-z])\s*=\s*[a-z]\s*
-                       (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*?
-                       (?:\[\s*(?P=b)\s*]|\.\s*get\s*\(\s*(?P=b)\s*\))
-                   )
-                                            # interstitial junk
-                   \s*(?:\|\|\s*null\s*)?(?:\)\s*)?&&\s*(?:\(\s*)?
-               (?(c)(?P=c)|(?P=b))\s*=\s*   # [c|b]=
-                                            # nfunc|nfunc[idx]
-                   (?P<nfunc>[a-zA-Z_$][\w$]*)(?:\s*\[(?P<idx>\d+)\])?\s*\(\s*[\w$]+\s*\)
             ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'),
             default=(None, None))
         # thx bashonly: yt-dlp/yt-dlp/pull/10611
@@ -1697,15 +1697,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 r'''(?xs)
                     (?:(?<=[^\w$])|^)       # instead of \b, which ignores $
                     (?P<name>(?!\d)[a-zA-Z\d_$]+)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\)
-                    \s*\{(?:(?!};).)+?["']enhanced_except_
+                    \s*\{(?:(?!};).)+?(?:
+                        ["']enhanced_except_ |
+                        return\s*(?P<q>"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[a-zA-Z0-9_$]+
+                    )
                 ''', jscode, 'Initial JS player n function name', group='name')
         if not idx:
             return func_name
 
-        return self._parse_json(self._search_regex(
+        return self._search_json(
-            r'var\s+{0}\s*=\s*(\[.+?\])\s*[,;]'.format(re.escape(func_name)), jscode,
+            r'var\s+{0}\s*='.format(re.escape(func_name)), jscode,
-            'Initial JS player n function list ({0}.{1})'.format(func_name, idx)),
+            'Initial JS player n function list ({0}.{1})'.format(func_name, idx),
-            func_name, transform_source=js_to_json)[int(idx)]
+            func_name, contains_pattern=r'\[[\s\S]+\]', end_pattern='[,;]',
+            transform_source=js_to_json)[int(idx)]
 
     def _extract_n_function_code(self, video_id, player_url):
         player_id = self._extract_player_info(player_url)
@@ -1728,13 +1732,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
         def extract_nsig(s):
             try:
-                ret = func([s])
+                ret = func([s], kwargs={'_ytdl_do_not_return': s})
             except JSInterpreter.Exception:
                 raise
             except Exception as e:
                 raise JSInterpreter.Exception(traceback.format_exc(), cause=e)
 
-            if ret.startswith('enhanced_except_'):
+            if ret.startswith('enhanced_except_') or ret.endswith(s):
                 raise JSInterpreter.Exception('Signature function returned an exception')
             return ret
 

youtube_dl/jsinterp.py

@@ -368,7 +368,7 @@ class Debugger(object):
                 raise
             if cls.ENABLED and stmt.strip():
                 if should_ret or repr(ret) != stmt:
-                    cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion)
+                    cls.write(['->', '=>'][bool(should_ret)], repr(ret), '<-|', stmt, level=allow_recursion)
             return ret, should_ret
         return interpret_statement
 
@@ -397,6 +397,9 @@ class JSInterpreter(object):
         RE_FLAGS = {
             # special knowledge: Python's re flags are bitmask values, current max 128
             # invent new bitmask values well above that for literal parsing
+            # JS 'u' flag is effectively always set (surrogate pairs aren't seen),
+            # but \u{...} and \p{...} escapes aren't handled); no additional JS 'v'
+            # features are supported
             # TODO: execute matches with these flags (remaining: d, y)
             'd': 1024,  # Generate indices for substring matches
             'g': 2048,  # Global search
@@ -404,6 +407,7 @@ class JSInterpreter(object):
             'm': re.M,  # Multi-line search
             's': re.S,  # Allows . to match newline characters
             'u': re.U,  # Treat a pattern as a sequence of unicode code points
+            'v': re.U,  # Like 'u' with extended character class and \p{} syntax
             'y': 4096,  # Perform a "sticky" search that matches starting at the current position in the target string
         }
 
@@ -484,9 +488,18 @@ class JSInterpreter(object):
         skipping = 0
         if skip_delims:
             skip_delims = variadic(skip_delims)
+        skip_txt = None
         for idx, char in enumerate(expr):
+            if skip_txt and idx <= skip_txt[1]:
+                continue
             paren_delta = 0
             if not in_quote:
+                if char == '/' and expr[idx:idx + 2] == '/*':
+                    # skip a comment
+                    skip_txt = expr[idx:].find('*/', 2)
+                    skip_txt = [idx, idx + skip_txt + 1] if skip_txt >= 2 else None
+                    if skip_txt:
+                        continue
                 if char in _MATCHING_PARENS:
                     counters[_MATCHING_PARENS[char]] += 1
                     paren_delta = 1
@@ -519,11 +532,18 @@ class JSInterpreter(object):
             if pos < delim_len:
                 pos += 1
                 continue
+            if skip_txt and skip_txt[0] >= start and skip_txt[1] <= idx - delim_len:
+                yield expr[start:skip_txt[0]] + expr[skip_txt[1] + 1: idx - delim_len]
+            else:
                 yield expr[start: idx - delim_len]
+            skip_txt = None
             start, pos = idx + 1, 0
             splits += 1
             if max_split and splits >= max_split:
                 break
+        if skip_txt and skip_txt[0] >= start:
+            yield expr[start:skip_txt[0]] + expr[skip_txt[1] + 1:]
+        else:
             yield expr[start:]
 
     @classmethod
@@ -583,7 +603,7 @@ class JSInterpreter(object):
 
     # used below
     _VAR_RET_THROW_RE = re.compile(r'''(?x)
-        (?P<var>(?:var|const|let)\s)|return(?:\s+|(?=["'])|$)|(?P<throw>throw\s+)
+        (?:(?P<var>var|const|let)\s+|(?P<ret>return)(?:\s+|(?=["'])|$)|(?P<throw>throw)\s+)
         ''')
     _COMPOUND_RE = re.compile(r'''(?x)
         (?P<try>try)\s*\{|
@@ -663,7 +683,7 @@ class JSInterpreter(object):
             expr = stmt[len(m.group(0)):].strip()
             if m.group('throw'):
                 raise JS_Throw(self.interpret_expression(expr, local_vars, allow_recursion))
-            should_return = not m.group('var')
+            should_return = 'return' if m.group('ret') else False
         if not expr:
             return None, should_return
 
@@ -948,14 +968,20 @@ class JSInterpreter(object):
             return _Infinity, should_return
 
         elif md.get('return'):
-            return local_vars[m.group('name')], should_return
+            ret = local_vars[m.group('name')]
+            # challenge may try to force returning the original value
+            # use an optional internal var to block this
+            if should_return == 'return':
+                if '_ytdl_do_not_return' not in local_vars:
+                    return ret, True
+                return (ret, True) if ret != local_vars['_ytdl_do_not_return'] else (ret, False)
+            else:
+                return ret, should_return
 
-        try:
+        with compat_contextlib_suppress(ValueError):
             ret = json.loads(js_to_json(expr))  # strict=True)
             if not md.get('attribute'):
                 return ret, should_return
-        except ValueError:
-            pass
 
         if md.get('indexing'):
             val = local_vars[m.group('in')]
@@ -1047,13 +1073,47 @@ class JSInterpreter(object):
                     raise self.Exception('Unsupported Math method ' + member, expr=expr)
 
                 if member == 'split':
-                    assertion(argvals, 'takes one or more arguments')
+                    assertion(len(argvals) <= 2, 'takes at most two arguments')
-                    assertion(len(argvals) == 1, 'with limit argument is not implemented')
+                    if len(argvals) > 1:
-                    return obj.split(argvals[0]) if argvals[0] else list(obj)
+                        limit = argvals[1]
+                        assertion(isinstance(limit, int) and limit >= 0, 'integer limit >= 0')
+                        if limit == 0:
+                            return []
+                    else:
+                        limit = 0
+                    if len(argvals) == 0:
+                        argvals = [JS_Undefined]
+                    elif isinstance(argvals[0], self.JS_RegExp):
+                        # avoid re.split(), similar but not enough
+
+                        def where():
+                            for m in argvals[0].finditer(obj):
+                                yield m.span(0)
+                            yield (None, None)
+
+                        def splits(limit=limit):
+                            i = 0
+                            for j, jj in where():
+                                if j == jj == 0:
+                                    continue
+                                if j is None and i >= len(obj):
+                                    break
+                                yield obj[i:j]
+                                if jj is None or limit == 1:
+                                    break
+                                limit -= 1
+                                i = jj
+
+                        return list(splits())
+                    return (
+                        obj.split(argvals[0], limit - 1) if argvals[0] and argvals[0] != JS_Undefined
+                        else list(obj)[:limit or None])
                 elif member == 'join':
                     assertion(isinstance(obj, list), 'must be applied on a list')
-                    assertion(len(argvals) == 1, 'takes exactly one argument')
+                    assertion(len(argvals) <= 1, 'takes at most one argument')
-                    return argvals[0].join(obj)
+                    return (',' if len(argvals) == 0 else argvals[0]).join(
+                        ('' if x in (None, JS_Undefined) else _js_toString(x))
+                        for x in obj)
                 elif member == 'reverse':
                     assertion(not argvals, 'does not take any arguments')
                     obj.reverse()
@@ -1075,37 +1135,31 @@ class JSInterpreter(object):
                     index, how_many = map(int, (argvals + [len(obj)])[:2])
                     if index < 0:
                         index += len(obj)
-                    add_items = argvals[2:]
+                    res = [obj.pop(index)
-                    res = []
+                           for _ in range(index, min(index + how_many, len(obj)))]
-                    for _ in range(index, min(index + how_many, len(obj))):
+                    obj[index:index] = argvals[2:]
-                        res.append(obj.pop(index))
-                    for i, item in enumerate(add_items):
-                        obj.insert(index + i, item)
                     return res
-                elif member == 'unshift':
+                elif member in ('shift', 'pop'):
-                    assertion(isinstance(obj, list), 'must be applied on a list')
-                    assertion(argvals, 'takes one or more arguments')
-                    for item in reversed(argvals):
-                        obj.insert(0, item)
-                    return obj
-                elif member == 'pop':
                     assertion(isinstance(obj, list), 'must be applied on a list')
                     assertion(not argvals, 'does not take any arguments')
-                    if not obj:
+                    return obj.pop(0 if member == 'shift' else -1) if len(obj) > 0 else JS_Undefined
-                        return
+                elif member == 'unshift':
-                    return obj.pop()
+                    assertion(isinstance(obj, list), 'must be applied on a list')
+                    # not enforced: assertion(argvals, 'takes one or more arguments')
+                    obj[0:0] = argvals
+                    return len(obj)
                 elif member == 'push':
-                    assertion(argvals, 'takes one or more arguments')
+                    # not enforced: assertion(argvals, 'takes one or more arguments')
                     obj.extend(argvals)
-                    return obj
+                    return len(obj)
                 elif member == 'forEach':
                     assertion(argvals, 'takes one or more arguments')
-                    assertion(len(argvals) <= 2, 'takes at-most 2 arguments')
+                    assertion(len(argvals) <= 2, 'takes at most 2 arguments')
                     f, this = (argvals + [''])[:2]
                     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')
+                    assertion(len(argvals) <= 2, 'takes at most 2 arguments')
                     idx, start = (argvals + [0])[:2]
                     try:
                         return obj.index(idx, start)
@@ -1114,7 +1168,7 @@ class JSInterpreter(object):
                 elif member == 'charCodeAt':
                     assertion(isinstance(obj, compat_str), 'must be applied on a string')
                     # assertion(len(argvals) == 1, 'takes exactly one argument') # but not enforced
-                    idx = argvals[0] if isinstance(argvals[0], int) else 0
+                    idx = argvals[0] if len(argvals) > 0 and isinstance(argvals[0], int) else 0
                     if idx >= len(obj):
                         return None
                     return ord(obj[idx])
@@ -1165,7 +1219,7 @@ class JSInterpreter(object):
             yield self.interpret_expression(v, local_vars, allow_recursion)
 
     def extract_object(self, objname):
-        _FUNC_NAME_RE = r'''(?:[a-zA-Z$0-9]+|"[a-zA-Z$0-9]+"|'[a-zA-Z$0-9]+')'''
+        _FUNC_NAME_RE = r'''(?:{n}|"{n}"|'{n}')'''.format(n=_NAME_RE)
         obj = {}
         fields = next(filter(None, (
             obj_m.group('fields') for obj_m in re.finditer(
@@ -1224,6 +1278,7 @@ class JSInterpreter(object):
 
     def extract_function_from_code(self, argnames, code, *global_stack):
         local_vars = {}
+
         while True:
             mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code)
             if mobj is None:
@@ -1234,10 +1289,11 @@ class JSInterpreter(object):
                 [x.strip() for x in mobj.group('args').split(',')],
                 body, local_vars, *global_stack))
             code = code[:start] + name + remaining
+
         return self.build_function(argnames, code, local_vars, *global_stack)
 
-    def call_function(self, funcname, *args):
+    def call_function(self, funcname, *args, **kw_global_vars):
-        return self.extract_function(funcname)(args)
+        return self.extract_function(funcname)(args, kw_global_vars)
 
     @classmethod
     def build_arglist(cls, arg_text):
@@ -1256,8 +1312,9 @@ class JSInterpreter(object):
         global_stack = list(global_stack) or [{}]
         argnames = tuple(argnames)
 
-        def resf(args, kwargs={}, allow_recursion=100):
+        def resf(args, kwargs=None, allow_recursion=100):
-            global_stack[0].update(zip_longest(argnames, args, fillvalue=None))
+            kwargs = kwargs or {}
+            global_stack[0].update(zip_longest(argnames, args, fillvalue=JS_Undefined))
             global_stack[0].update(kwargs)
             var_stack = LocalNameSpace(*global_stack)
             ret, should_abort = self.interpret_statement(code.replace('\n', ' '), var_stack, allow_recursion - 1)