]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/utils.py
Update upstream source from tag 'upstream/2019.06.08'
[youtubedl] / youtube_dl / utils.py
index 2340bc306b69652d43246472e7fd3f5e819e2af2..ead9bd8628c88183e5ca0aaa7a0eb9b8fe03d836 100644 (file)
@@ -11,6 +11,7 @@ import contextlib
 import ctypes
 import datetime
 import email.utils
+import email.header
 import errno
 import functools
 import gzip
@@ -21,7 +22,6 @@ import locale
 import math
 import operator
 import os
-import pipes
 import platform
 import random
 import re
@@ -35,9 +35,12 @@ import xml.etree.ElementTree
 import zlib
 
 from .compat import (
+    compat_HTMLParseError,
     compat_HTMLParser,
     compat_basestring,
     compat_chr,
+    compat_cookiejar,
+    compat_ctypes_WINFUNCTYPE,
     compat_etree_fromstring,
     compat_expanduser,
     compat_html_entities,
@@ -47,7 +50,6 @@ from .compat import (
     compat_os_name,
     compat_parse_qs,
     compat_shlex_quote,
-    compat_socket_create_connection,
     compat_str,
     compat_struct_pack,
     compat_struct_unpack,
@@ -80,7 +82,7 @@ def register_socks_protocols():
 compiled_regex_type = type(re.compile(''))
 
 std_headers = {
-    'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/47.0 (Chrome)',
+    'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:59.0) Gecko/20100101 Firefox/59.0',
     'Accept-Charset': 'ISO-8859-1,utf-8;q=0.7,*;q=0.7',
     'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
     'Accept-Encoding': 'gzip, deflate',
@@ -123,8 +125,8 @@ KNOWN_EXTENSIONS = (
 
 # needed for sanitizing filenames in restricted mode
 ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
-                        itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUYP', ['ss'],
-                                        'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuypy')))
+                        itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
+                                        'aaaaaa', ['ae'], 'ceeeeiiiionooooooo', ['oe'], 'uuuuuy', ['th'], 'y')))
 
 DATE_FORMATS = (
     '%d %B %Y',
@@ -158,6 +160,8 @@ DATE_FORMATS = (
     '%Y-%m-%dT%H:%M',
     '%b %d %Y at %H:%M',
     '%b %d %Y at %H:%M:%S',
+    '%B %d %Y at %H:%M',
+    '%B %d %Y at %H:%M:%S',
 )
 
 DATE_FORMATS_DAY_FIRST = list(DATE_FORMATS)
@@ -180,6 +184,7 @@ DATE_FORMATS_MONTH_FIRST.extend([
 ])
 
 PACKED_CODES_RE = r"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)"
+JSON_LD_RE = r'(?is)<script[^>]+type=(["\']?)application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>'
 
 
 def preferredencoding():
@@ -364,9 +369,9 @@ def get_elements_by_attribute(attribute, value, html, escape_value=True):
     retlist = []
     for m in re.finditer(r'''(?xs)
         <([a-zA-Z0-9:._-]+)
-         (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'))*?
+         (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'|))*?
          \s+%s=['"]?%s['"]?
-         (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'))*?
+         (?:\s+[a-zA-Z0-9:._-]+(?:=[a-zA-Z0-9:._-]*|="[^"]*"|='[^']*'|))*?
         \s*>
         (?P<content>.*?)
         </\1>
@@ -408,8 +413,12 @@ def extract_attributes(html_element):
     but the cases in the unit test will work for all of 2.6, 2.7, 3.2-3.5.
     """
     parser = HTMLAttributeParser()
-    parser.feed(html_element)
-    parser.close()
+    try:
+        parser.feed(html_element)
+        parser.close()
+    # Older Python may throw HTMLParseError in case of malformed HTML
+    except compat_HTMLParseError:
+        pass
     return parser.attrs
 
 
@@ -421,8 +430,8 @@ def clean_html(html):
 
     # Newline vs <br />
     html = html.replace('\n', ' ')
-    html = re.sub(r'\s*<\s*br\s*/?\s*>\s*', '\n', html)
-    html = re.sub(r'<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
+    html = re.sub(r'(?u)\s*<\s*br\s*/?\s*>\s*', '\n', html)
+    html = re.sub(r'(?u)<\s*/\s*p\s*>\s*<\s*p[^>]*>', '\n', html)
     # Strip html tags
     html = re.sub('<.*?>', '', html)
     # Replace html entities
@@ -530,10 +539,22 @@ def sanitize_path(s):
     return os.path.join(*sanitized_path)
 
 
-# Prepend protocol-less URLs with `http:` scheme in order to mitigate the number of
-# unwanted failures due to missing protocol
 def sanitize_url(url):
-    return 'http:%s' % url if url.startswith('//') else url
+    # Prepend protocol-less URLs with `http:` scheme in order to mitigate
+    # the number of unwanted failures due to missing protocol
+    if url.startswith('//'):
+        return 'http:%s' % url
+    # Fix some common typos seen so far
+    COMMON_TYPOS = (
+        # https://github.com/ytdl-org/youtube-dl/issues/15649
+        (r'^httpss://', r'https://'),
+        # https://bx1.be/lives/direct-tv/
+        (r'^rmtp([es]?)://', r'rtmp\1://'),
+    )
+    for mistake, fixup in COMMON_TYPOS:
+        if re.match(mistake, url):
+            return re.sub(mistake, fixup, url)
+    return url
 
 
 def sanitized_Request(url, *args, **kwargs):
@@ -575,7 +596,7 @@ def _htmlentity_transform(entity_with_semicolon):
             numstr = '0%s' % numstr
         else:
             base = 10
-        # See https://github.com/rg3/youtube-dl/issues/7518
+        # See https://github.com/ytdl-org/youtube-dl/issues/7518
         try:
             return compat_chr(int(numstr, base))
         except ValueError:
@@ -591,7 +612,7 @@ def unescapeHTML(s):
     assert type(s) == compat_str
 
     return re.sub(
-        r'&([^;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
+        r'&([^&;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
 
 
 def get_subprocess_encoding():
@@ -840,8 +861,8 @@ class XAttrMetadataError(YoutubeDLError):
         self.msg = msg
 
         # Parsing code and msg
-        if (self.code in (errno.ENOSPC, errno.EDQUOT) or
-                'No space left' in self.msg or 'Disk quota excedded' in self.msg):
+        if (self.code in (errno.ENOSPC, errno.EDQUOT)
+                or 'No space left' in self.msg or 'Disk quota excedded' in self.msg):
             self.reason = 'NO_SPACE'
         elif self.code == errno.E2BIG or 'Argument list too long' in self.msg:
             self.reason = 'VALUE_TOO_LONG'
@@ -856,18 +877,56 @@ class XAttrUnavailableError(YoutubeDLError):
 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
     # Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
     # expected HTTP responses to meet HTTP/1.0 or later (see also
-    # https://github.com/rg3/youtube-dl/issues/6727)
+    # https://github.com/ytdl-org/youtube-dl/issues/6727)
     if sys.version_info < (3, 0):
-        kwargs[b'strict'] = True
-    hc = http_class(*args, **kwargs)
+        kwargs['strict'] = True
+    hc = http_class(*args, **compat_kwargs(kwargs))
     source_address = ydl_handler._params.get('source_address')
+
     if source_address is not None:
+        # This is to workaround _create_connection() from socket where it will try all
+        # address data from getaddrinfo() including IPv6. This filters the result from
+        # getaddrinfo() based on the source_address value.
+        # This is based on the cpython socket.create_connection() function.
+        # https://github.com/python/cpython/blob/master/Lib/socket.py#L691
+        def _create_connection(address, timeout=socket._GLOBAL_DEFAULT_TIMEOUT, source_address=None):
+            host, port = address
+            err = None
+            addrs = socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM)
+            af = socket.AF_INET if '.' in source_address[0] else socket.AF_INET6
+            ip_addrs = [addr for addr in addrs if addr[0] == af]
+            if addrs and not ip_addrs:
+                ip_version = 'v4' if af == socket.AF_INET else 'v6'
+                raise socket.error(
+                    "No remote IP%s addresses available for connect, can't use '%s' as source address"
+                    % (ip_version, source_address[0]))
+            for res in ip_addrs:
+                af, socktype, proto, canonname, sa = res
+                sock = None
+                try:
+                    sock = socket.socket(af, socktype, proto)
+                    if timeout is not socket._GLOBAL_DEFAULT_TIMEOUT:
+                        sock.settimeout(timeout)
+                    sock.bind(source_address)
+                    sock.connect(sa)
+                    err = None  # Explicitly break reference cycle
+                    return sock
+                except socket.error as _:
+                    err = _
+                    if sock is not None:
+                        sock.close()
+            if err is not None:
+                raise err
+            else:
+                raise socket.error('getaddrinfo returns an empty list')
+        if hasattr(hc, '_create_connection'):
+            hc._create_connection = _create_connection
         sa = (source_address, 0)
         if hasattr(hc, 'source_address'):  # Python 2.7+
             hc.source_address = sa
         else:  # Python 2.6
             def _hc_connect(self, *args, **kwargs):
-                sock = compat_socket_create_connection(
+                sock = _create_connection(
                     (self.host, self.port), self.timeout, sa)
                 if is_https:
                     self.sock = ssl.wrap_socket(
@@ -931,14 +990,6 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
         except zlib.error:
             return zlib.decompress(data)
 
-    @staticmethod
-    def addinfourl_wrapper(stream, headers, url, code):
-        if hasattr(compat_urllib_request.addinfourl, 'getcode'):
-            return compat_urllib_request.addinfourl(stream, headers, url, code)
-        ret = compat_urllib_request.addinfourl(stream, headers, url)
-        ret.code = code
-        return ret
-
     def http_request(self, req):
         # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
         # always respected by websites, some tend to give out URLs with non percent-encoded
@@ -990,17 +1041,17 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
                     break
                 else:
                     raise original_ioerror
-            resp = self.addinfourl_wrapper(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
+            resp = compat_urllib_request.addinfourl(uncompressed, old_resp.headers, old_resp.url, old_resp.code)
             resp.msg = old_resp.msg
             del resp.headers['Content-encoding']
         # deflate
         if resp.headers.get('Content-encoding', '') == 'deflate':
             gz = io.BytesIO(self.deflate(resp.read()))
-            resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
+            resp = compat_urllib_request.addinfourl(gz, old_resp.headers, old_resp.url, old_resp.code)
             resp.msg = old_resp.msg
             del resp.headers['Content-encoding']
         # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
-        # https://github.com/rg3/youtube-dl/issues/6457).
+        # https://github.com/ytdl-org/youtube-dl/issues/6457).
         if 300 <= resp.code < 400:
             location = resp.headers.get('Location')
             if location:
@@ -1089,6 +1140,49 @@ class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
             req, **kwargs)
 
 
+class YoutubeDLCookieJar(compat_cookiejar.MozillaCookieJar):
+    _HTTPONLY_PREFIX = '#HttpOnly_'
+
+    def save(self, filename=None, ignore_discard=False, ignore_expires=False):
+        # Store session cookies with `expires` set to 0 instead of an empty
+        # string
+        for cookie in self:
+            if cookie.expires is None:
+                cookie.expires = 0
+        compat_cookiejar.MozillaCookieJar.save(self, filename, ignore_discard, ignore_expires)
+
+    def load(self, filename=None, ignore_discard=False, ignore_expires=False):
+        """Load cookies from a file."""
+        if filename is None:
+            if self.filename is not None:
+                filename = self.filename
+            else:
+                raise ValueError(compat_cookiejar.MISSING_FILENAME_TEXT)
+
+        cf = io.StringIO()
+        with open(filename) as f:
+            for line in f:
+                if line.startswith(self._HTTPONLY_PREFIX):
+                    line = line[len(self._HTTPONLY_PREFIX):]
+                cf.write(compat_str(line))
+        cf.seek(0)
+        self._really_load(cf, filename, ignore_discard, ignore_expires)
+        # Session cookies are denoted by either `expires` field set to
+        # an empty string or 0. MozillaCookieJar only recognizes the former
+        # (see [1]). So we need force the latter to be recognized as session
+        # cookies on our own.
+        # Session cookies may be important for cookies-based authentication,
+        # e.g. usually, when user does not check 'Remember me' check box while
+        # logging in on a site, some important cookies are stored as session
+        # cookies so that not recognizing them will result in failed login.
+        # 1. https://bugs.python.org/issue17164
+        for cookie in self:
+            # Treat `expires=0` cookies as session cookies
+            if cookie.expires == 0:
+                cookie.expires = None
+                cookie.discard = True
+
+
 class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
     def __init__(self, cookiejar=None):
         compat_urllib_request.HTTPCookieProcessor.__init__(self, cookiejar)
@@ -1096,7 +1190,7 @@ class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
     def http_response(self, request, response):
         # Python 2 will choke on next HTTP request in row if there are non-ASCII
         # characters in Set-Cookie HTTP header of last response (see
-        # https://github.com/rg3/youtube-dl/issues/6769).
+        # https://github.com/ytdl-org/youtube-dl/issues/6769).
         # In order to at least prevent crashing we will percent encode Set-Cookie
         # header before HTTPCookieProcessor starts processing it.
         # if sys.version_info < (3, 0) and response.headers:
@@ -1186,7 +1280,7 @@ def unified_timestamp(date_str, day_first=True):
     if date_str is None:
         return None
 
-    date_str = date_str.replace(',', ' ')
+    date_str = re.sub(r'[,|]', '', date_str)
 
     pm_delta = 12 if re.search(r'(?i)PM', date_str) else 0
     timezone, date_str = extract_timezone(date_str)
@@ -1194,6 +1288,16 @@ def unified_timestamp(date_str, day_first=True):
     # Remove AM/PM + timezone
     date_str = re.sub(r'(?i)\s*(?:AM|PM)(?:\s+[A-Z]+)?', '', date_str)
 
+    # Remove unrecognized timezones from ISO 8601 alike timestamps
+    m = re.search(r'\d{1,2}:\d{1,2}(?:\.\d+)?(?P<tz>\s*[A-Z]+)$', date_str)
+    if m:
+        date_str = date_str[:-len(m.group('tz'))]
+
+    # Python only supports microseconds, so remove nanoseconds
+    m = re.search(r'^([0-9]{4,}-[0-9]{1,2}-[0-9]{1,2}T[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}\.[0-9]{6})[0-9]+$', date_str)
+    if m:
+        date_str = m.group(1)
+
     for expression in date_formats(day_first):
         try:
             dt = datetime.datetime.strptime(date_str, expression) - timezone + datetime.timedelta(hours=pm_delta)
@@ -1206,7 +1310,7 @@ def unified_timestamp(date_str, day_first=True):
 
 
 def determine_ext(url, default_ext='unknown_video'):
-    if url is None:
+    if url is None or '.' not in url:
         return default_ext
     guess = url.partition('?')[0].rpartition('.')[2]
     if re.match(r'^[A-Za-z0-9]+$', guess):
@@ -1326,31 +1430,31 @@ def _windows_write_string(s, out):
     if fileno not in WIN_OUTPUT_IDS:
         return False
 
-    GetStdHandle = ctypes.WINFUNCTYPE(
+    GetStdHandle = compat_ctypes_WINFUNCTYPE(
         ctypes.wintypes.HANDLE, ctypes.wintypes.DWORD)(
-        (b'GetStdHandle', ctypes.windll.kernel32))
+        ('GetStdHandle', ctypes.windll.kernel32))
     h = GetStdHandle(WIN_OUTPUT_IDS[fileno])
 
-    WriteConsoleW = ctypes.WINFUNCTYPE(
+    WriteConsoleW = compat_ctypes_WINFUNCTYPE(
         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE, ctypes.wintypes.LPWSTR,
         ctypes.wintypes.DWORD, ctypes.POINTER(ctypes.wintypes.DWORD),
-        ctypes.wintypes.LPVOID)((b'WriteConsoleW', ctypes.windll.kernel32))
+        ctypes.wintypes.LPVOID)(('WriteConsoleW', ctypes.windll.kernel32))
     written = ctypes.wintypes.DWORD(0)
 
-    GetFileType = ctypes.WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)((b'GetFileType', ctypes.windll.kernel32))
+    GetFileType = compat_ctypes_WINFUNCTYPE(ctypes.wintypes.DWORD, ctypes.wintypes.DWORD)(('GetFileType', ctypes.windll.kernel32))
     FILE_TYPE_CHAR = 0x0002
     FILE_TYPE_REMOTE = 0x8000
-    GetConsoleMode = ctypes.WINFUNCTYPE(
+    GetConsoleMode = compat_ctypes_WINFUNCTYPE(
         ctypes.wintypes.BOOL, ctypes.wintypes.HANDLE,
         ctypes.POINTER(ctypes.wintypes.DWORD))(
-        (b'GetConsoleMode', ctypes.windll.kernel32))
+        ('GetConsoleMode', ctypes.windll.kernel32))
     INVALID_HANDLE_VALUE = ctypes.wintypes.DWORD(-1).value
 
     def not_a_console(handle):
         if handle == INVALID_HANDLE_VALUE or handle is None:
             return True
-        return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR or
-                GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
+        return ((GetFileType(handle) & ~FILE_TYPE_REMOTE) != FILE_TYPE_CHAR
+                or GetConsoleMode(handle, ctypes.byref(ctypes.wintypes.DWORD())) == 0)
 
     if not_a_console(h):
         return False
@@ -1386,8 +1490,8 @@ def write_string(s, out=None, encoding=None):
         if _windows_write_string(s, out):
             return
 
-    if ('b' in getattr(out, 'mode', '') or
-            sys.version_info[0] < 3):  # Python 2 lies about mode of sys.stderr
+    if ('b' in getattr(out, 'mode', '')
+            or sys.version_info[0] < 3):  # Python 2 lies about mode of sys.stderr
         byt = s.encode(encoding or preferredencoding(), 'ignore')
         out.write(byt)
     elif hasattr(out, 'buffer'):
@@ -1532,7 +1636,7 @@ def shell_quote(args):
         if isinstance(a, bytes):
             # We may get a filename encoded with 'encodeFilename'
             a = a.decode(encoding)
-        quoted_args.append(pipes.quote(a))
+        quoted_args.append(compat_shlex_quote(a))
     return ' '.join(quoted_args)
 
 
@@ -1672,6 +1776,36 @@ def parse_count(s):
     return lookup_unit_table(_UNIT_TABLE, s)
 
 
+def parse_resolution(s):
+    if s is None:
+        return {}
+
+    mobj = re.search(r'\b(?P<w>\d+)\s*[xX×]\s*(?P<h>\d+)\b', s)
+    if mobj:
+        return {
+            'width': int(mobj.group('w')),
+            'height': int(mobj.group('h')),
+        }
+
+    mobj = re.search(r'\b(\d+)[pPiI]\b', s)
+    if mobj:
+        return {'height': int(mobj.group(1))}
+
+    mobj = re.search(r'\b([48])[kK]\b', s)
+    if mobj:
+        return {'height': int(mobj.group(1)) * 540}
+
+    return {}
+
+
+def parse_bitrate(s):
+    if not isinstance(s, compat_str):
+        return
+    mobj = re.search(r'\b(\d+)\s*kbps', s)
+    if mobj:
+        return int(mobj.group(1))
+
+
 def month_by_name(name, lang='en'):
     """ Return the number of a month by (locale-independently) English name """
 
@@ -1758,7 +1892,7 @@ def urljoin(base, path):
         path = path.decode('utf-8')
     if not isinstance(path, compat_str) or not path:
         return None
-    if re.match(r'^(?:https?:)?//', path):
+    if re.match(r'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path):
         return path
     if isinstance(base, bytes):
         base = base.decode('utf-8')
@@ -1788,7 +1922,7 @@ def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
         return default
     try:
         return int(v) * invscale // scale
-    except ValueError:
+    except (ValueError, TypeError):
         return default
 
 
@@ -1809,12 +1943,23 @@ def float_or_none(v, scale=1, invscale=1, default=None):
         return default
     try:
         return float(v) * invscale / scale
-    except ValueError:
+    except (ValueError, TypeError):
         return default
 
 
-def strip_or_none(v):
-    return None if v is None else v.strip()
+def bool_or_none(v, default=None):
+    return v if isinstance(v, bool) else default
+
+
+def strip_or_none(v, default=None):
+    return v.strip() if isinstance(v, compat_str) else default
+
+
+def url_or_none(url):
+    if not url or not isinstance(url, compat_str):
+        return None
+    url = url.strip()
+    return url if re.match(r'^(?:[a-zA-Z][\da-zA-Z.+-]*:)?//', url) else None
 
 
 def parse_duration(s):
@@ -1829,10 +1974,20 @@ def parse_duration(s):
         days, hours, mins, secs, ms = m.groups()
     else:
         m = re.match(
-            r'''(?ix)(?:P?T)?
+            r'''(?ix)(?:P?
+                (?:
+                    [0-9]+\s*y(?:ears?)?\s*
+                )?
+                (?:
+                    [0-9]+\s*m(?:onths?)?\s*
+                )?
+                (?:
+                    [0-9]+\s*w(?:eeks?)?\s*
+                )?
                 (?:
                     (?P<days>[0-9]+)\s*d(?:ays?)?\s*
                 )?
+                T)?
                 (?:
                     (?P<hours>[0-9]+)\s*h(?:ours?)?\s*
                 )?
@@ -1897,7 +2052,7 @@ def get_exe_version(exe, args=['--version'],
     try:
         # STDIN should be redirected too. On UNIX-like systems, ffmpeg triggers
         # SIGTTOU if youtube-dl is run in the background.
-        # See https://github.com/rg3/youtube-dl/issues/955#issuecomment-209789656
+        # See https://github.com/ytdl-org/youtube-dl/issues/955#issuecomment-209789656
         out, _ = subprocess.Popen(
             [encodeArgument(exe)] + args,
             stdin=subprocess.PIPE,
@@ -1927,7 +2082,7 @@ class PagedList(object):
 
 
 class OnDemandPagedList(PagedList):
-    def __init__(self, pagefunc, pagesize, use_cache=False):
+    def __init__(self, pagefunc, pagesize, use_cache=True):
         self._pagefunc = pagefunc
         self._pagesize = pagesize
         self._use_cache = use_cache
@@ -2092,6 +2247,58 @@ def update_Request(req, url=None, data=None, headers={}, query={}):
     return new_req
 
 
+def _multipart_encode_impl(data, boundary):
+    content_type = 'multipart/form-data; boundary=%s' % boundary
+
+    out = b''
+    for k, v in data.items():
+        out += b'--' + boundary.encode('ascii') + b'\r\n'
+        if isinstance(k, compat_str):
+            k = k.encode('utf-8')
+        if isinstance(v, compat_str):
+            v = v.encode('utf-8')
+        # RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
+        # suggests sending UTF-8 directly. Firefox sends UTF-8, too
+        content = b'Content-Disposition: form-data; name="' + k + b'"\r\n\r\n' + v + b'\r\n'
+        if boundary.encode('ascii') in content:
+            raise ValueError('Boundary overlaps with data')
+        out += content
+
+    out += b'--' + boundary.encode('ascii') + b'--\r\n'
+
+    return out, content_type
+
+
+def multipart_encode(data, boundary=None):
+    '''
+    Encode a dict to RFC 7578-compliant form-data
+
+    data:
+        A dict where keys and values can be either Unicode or bytes-like
+        objects.
+    boundary:
+        If specified a Unicode object, it's used as the boundary. Otherwise
+        a random boundary is generated.
+
+    Reference: https://tools.ietf.org/html/rfc7578
+    '''
+    has_specified_boundary = boundary is not None
+
+    while True:
+        if boundary is None:
+            boundary = '---------------' + str(random.randrange(0x0fffffff, 0xffffffff))
+
+        try:
+            out, content_type = _multipart_encode_impl(data, boundary)
+            break
+        except ValueError:
+            if has_specified_boundary:
+                raise
+            boundary = None
+
+    return out, content_type
+
+
 def dict_get(d, key_or_keys, default=None, skip_false_values=True):
     if isinstance(key_or_keys, (list, tuple)):
         for key in key_or_keys:
@@ -2103,13 +2310,30 @@ def dict_get(d, key_or_keys, default=None, skip_false_values=True):
 
 
 def try_get(src, getter, expected_type=None):
-    try:
-        v = getter(src)
-    except (AttributeError, KeyError, TypeError, IndexError):
-        pass
-    else:
-        if expected_type is None or isinstance(v, expected_type):
-            return v
+    if not isinstance(getter, (list, tuple)):
+        getter = [getter]
+    for get in getter:
+        try:
+            v = get(src)
+        except (AttributeError, KeyError, TypeError, IndexError):
+            pass
+        else:
+            if expected_type is None or isinstance(v, expected_type):
+                return v
+
+
+def merge_dicts(*dicts):
+    merged = {}
+    for a_dict in dicts:
+        for k, v in a_dict.items():
+            if v is None:
+                continue
+            if (k not in merged
+                    or (isinstance(v, compat_str) and v
+                        and isinstance(merged[k], compat_str)
+                        and not merged[k])):
+                merged[k] = v
+    return merged
 
 
 def encode_compat_str(string, encoding=preferredencoding(), errors='strict'):
@@ -2145,12 +2369,20 @@ def parse_age_limit(s):
         return int(m.group('age'))
     if s in US_RATINGS:
         return US_RATINGS[s]
-    return TV_PARENTAL_GUIDELINES.get(s)
+    m = re.match(r'^TV[_-]?(%s)$' % '|'.join(k[3:] for k in TV_PARENTAL_GUIDELINES), s)
+    if m:
+        return TV_PARENTAL_GUIDELINES['TV-' + m.group(1)]
+    return None
 
 
 def strip_jsonp(code):
     return re.sub(
-        r'(?s)^[a-zA-Z0-9_.$]+\s*\(\s*(.*)\);?\s*?(?://[^\n]*)*$', r'\1', code)
+        r'''(?sx)^
+            (?:window\.)?(?P<func_name>[a-zA-Z0-9_.$]*)
+            (?:\s*&&\s*(?P=func_name))?
+            \s*\(\s*(?P<callback_data>.*)\);?
+            \s*?(?://[^\n]*)*$''',
+        r'\g<callback_data>', code)
 
 
 def js_to_json(code):
@@ -2188,7 +2420,7 @@ def js_to_json(code):
         "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
         '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
         {comment}|,(?={skip}[\]}}])|
-        [a-zA-Z_][.a-zA-Z_0-9]*|
+        (?:(?<![0-9])[eE]|[a-df-zA-DF-Z_])[.a-zA-Z_0-9]*|
         \b(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:{skip}:)?|
         [0-9]+(?={skip}:)
         '''.format(comment=COMMENT_RE, skip=SKIP_RE), fix_kv, code)
@@ -2270,22 +2502,21 @@ def mimetype2ext(mt):
     return {
         '3gpp': '3gp',
         'smptett+xml': 'tt',
-        'srt': 'srt',
         'ttaf+xml': 'dfxp',
         'ttml+xml': 'ttml',
-        'vtt': 'vtt',
         'x-flv': 'flv',
         'x-mp4-fragmented': 'mp4',
+        'x-ms-sami': 'sami',
         'x-ms-wmv': 'wmv',
         'mpegurl': 'm3u8',
         'x-mpegurl': 'm3u8',
         'vnd.apple.mpegurl': 'm3u8',
         'dash+xml': 'mpd',
-        'f4m': 'f4m',
         'f4m+xml': 'f4m',
         'hds+xml': 'f4m',
         'vnd.ms-sstr+xml': 'ism',
         'quicktime': 'mov',
+        'mp2t': 'ts',
     }.get(res, res)
 
 
@@ -2298,14 +2529,14 @@ def parse_codecs(codecs_str):
     vcodec, acodec = None, None
     for full_codec in splited_codecs:
         codec = full_codec.split('.')[0]
-        if codec in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2', 'h263', 'h264', 'mp4v'):
+        if codec in ('avc1', 'avc2', 'avc3', 'avc4', 'vp9', 'vp8', 'hev1', 'hev2', 'h263', 'h264', 'mp4v', 'hvc1', 'av01'):
             if not vcodec:
                 vcodec = full_codec
-        elif codec in ('mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3'):
+        elif codec in ('mp4a', 'opus', 'vorbis', 'mp3', 'aac', 'ac-3', 'ec-3', 'eac3', 'dtsc', 'dtse', 'dtsh', 'dtsl'):
             if not acodec:
                 acodec = full_codec
         else:
-            write_string('WARNING: Unknown codec %s' % full_codec, sys.stderr)
+            write_string('WARNING: Unknown codec %s\n' % full_codec, sys.stderr)
     if not vcodec and not acodec:
         if len(splited_codecs) == 2:
             return {
@@ -2426,14 +2657,14 @@ def _match_one(filter_part, dct):
     if m:
         op = COMPARISON_OPERATORS[m.group('op')]
         actual_value = dct.get(m.group('key'))
-        if (m.group('quotedstrval') is not None or
-            m.group('strval') is not None or
+        if (m.group('quotedstrval') is not None
+            or m.group('strval') is not None
             # If the original field is a string and matching comparisonvalue is
             # a number we should respect the origin of the original field
             # and process comparison value as a string (see
-            # https://github.com/rg3/youtube-dl/issues/11082).
-            actual_value is not None and m.group('intval') is not None and
-                isinstance(actual_value, compat_str)):
+            # https://github.com/ytdl-org/youtube-dl/issues/11082).
+            or actual_value is not None and m.group('intval') is not None
+                and isinstance(actual_value, compat_str)):
             if m.group('op') not in ('=', '!='):
                 raise ValueError(
                     'Operator %s does not support string values!' % m.group('op'))
@@ -2457,8 +2688,8 @@ def _match_one(filter_part, dct):
         return op(actual_value, comparison_value)
 
     UNARY_OPERATORS = {
-        '': lambda v: v is not None,
-        '!': lambda v: v is None,
+        '': lambda v: (v is True) if isinstance(v, bool) else (v is not None),
+        '!': lambda v: (v is False) if isinstance(v, bool) else (v is None),
     }
     operator_rex = re.compile(r'''(?x)\s*
         (?P<op>%s)\s*(?P<key>[a-z_]+)
@@ -2508,27 +2739,102 @@ def srt_subtitles_timecode(seconds):
 
 
 def dfxp2srt(dfxp_data):
+    '''
+    @param dfxp_data A bytes-like object containing DFXP data
+    @returns A unicode object containing converted SRT data
+    '''
+    LEGACY_NAMESPACES = (
+        (b'http://www.w3.org/ns/ttml', [
+            b'http://www.w3.org/2004/11/ttaf1',
+            b'http://www.w3.org/2006/04/ttaf1',
+            b'http://www.w3.org/2006/10/ttaf1',
+        ]),
+        (b'http://www.w3.org/ns/ttml#styling', [
+            b'http://www.w3.org/ns/ttml#style',
+        ]),
+    )
+
+    SUPPORTED_STYLING = [
+        'color',
+        'fontFamily',
+        'fontSize',
+        'fontStyle',
+        'fontWeight',
+        'textDecoration'
+    ]
+
     _x = functools.partial(xpath_with_ns, ns_map={
+        'xml': 'http://www.w3.org/XML/1998/namespace',
         'ttml': 'http://www.w3.org/ns/ttml',
-        'ttaf1': 'http://www.w3.org/2006/10/ttaf1',
-        'ttaf1_0604': 'http://www.w3.org/2006/04/ttaf1',
+        'tts': 'http://www.w3.org/ns/ttml#styling',
     })
 
+    styles = {}
+    default_style = {}
+
     class TTMLPElementParser(object):
-        out = ''
+        _out = ''
+        _unclosed_elements = []
+        _applied_styles = []
 
         def start(self, tag, attrib):
-            if tag in (_x('ttml:br'), _x('ttaf1:br'), 'br'):
-                self.out += '\n'
+            if tag in (_x('ttml:br'), 'br'):
+                self._out += '\n'
+            else:
+                unclosed_elements = []
+                style = {}
+                element_style_id = attrib.get('style')
+                if default_style:
+                    style.update(default_style)
+                if element_style_id:
+                    style.update(styles.get(element_style_id, {}))
+                for prop in SUPPORTED_STYLING:
+                    prop_val = attrib.get(_x('tts:' + prop))
+                    if prop_val:
+                        style[prop] = prop_val
+                if style:
+                    font = ''
+                    for k, v in sorted(style.items()):
+                        if self._applied_styles and self._applied_styles[-1].get(k) == v:
+                            continue
+                        if k == 'color':
+                            font += ' color="%s"' % v
+                        elif k == 'fontSize':
+                            font += ' size="%s"' % v
+                        elif k == 'fontFamily':
+                            font += ' face="%s"' % v
+                        elif k == 'fontWeight' and v == 'bold':
+                            self._out += '<b>'
+                            unclosed_elements.append('b')
+                        elif k == 'fontStyle' and v == 'italic':
+                            self._out += '<i>'
+                            unclosed_elements.append('i')
+                        elif k == 'textDecoration' and v == 'underline':
+                            self._out += '<u>'
+                            unclosed_elements.append('u')
+                    if font:
+                        self._out += '<font' + font + '>'
+                        unclosed_elements.append('font')
+                    applied_style = {}
+                    if self._applied_styles:
+                        applied_style.update(self._applied_styles[-1])
+                    applied_style.update(style)
+                    self._applied_styles.append(applied_style)
+                self._unclosed_elements.append(unclosed_elements)
 
         def end(self, tag):
-            pass
+            if tag not in (_x('ttml:br'), 'br'):
+                unclosed_elements = self._unclosed_elements.pop()
+                for element in reversed(unclosed_elements):
+                    self._out += '</%s>' % element
+                if unclosed_elements and self._applied_styles:
+                    self._applied_styles.pop()
 
         def data(self, data):
-            self.out += data
+            self._out += data
 
         def close(self):
-            return self.out.strip()
+            return self._out.strip()
 
     def parse_node(node):
         target = TTMLPElementParser()
@@ -2536,13 +2842,47 @@ def dfxp2srt(dfxp_data):
         parser.feed(xml.etree.ElementTree.tostring(node))
         return parser.close()
 
-    dfxp = compat_etree_fromstring(dfxp_data.encode('utf-8'))
+    for k, v in LEGACY_NAMESPACES:
+        for ns in v:
+            dfxp_data = dfxp_data.replace(ns, k)
+
+    dfxp = compat_etree_fromstring(dfxp_data)
     out = []
-    paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall(_x('.//ttaf1:p')) or dfxp.findall(_x('.//ttaf1_0604:p')) or dfxp.findall('.//p')
+    paras = dfxp.findall(_x('.//ttml:p')) or dfxp.findall('.//p')
 
     if not paras:
         raise ValueError('Invalid dfxp/TTML subtitle')
 
+    repeat = False
+    while True:
+        for style in dfxp.findall(_x('.//ttml:style')):
+            style_id = style.get('id') or style.get(_x('xml:id'))
+            if not style_id:
+                continue
+            parent_style_id = style.get('style')
+            if parent_style_id:
+                if parent_style_id not in styles:
+                    repeat = True
+                    continue
+                styles[style_id] = styles[parent_style_id].copy()
+            for prop in SUPPORTED_STYLING:
+                prop_val = style.get(_x('tts:' + prop))
+                if prop_val:
+                    styles.setdefault(style_id, {})[prop] = prop_val
+        if repeat:
+            repeat = False
+        else:
+            break
+
+    for p in ('body', 'div'):
+        ele = xpath_element(dfxp, [_x('.//ttml:' + p), './/' + p])
+        if ele is None:
+            continue
+        style = styles.get(ele.get('style'))
+        if not style:
+            continue
+        default_style.update(style)
+
     for para, index in zip(paras, itertools.count(1)):
         begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
         end_time = parse_dfxp_time_expr(para.attrib.get('end'))
@@ -2571,6 +2911,8 @@ def cli_option(params, command_option, param):
 
 def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
     param = params.get(param)
+    if param is None:
+        return []
     assert isinstance(param, bool)
     if separator:
         return [command_option + separator + (true_value if param else false_value)]
@@ -2650,6 +2992,7 @@ class ISO639Utils(object):
         'gv': 'glv',
         'ha': 'hau',
         'he': 'heb',
+        'iw': 'heb',  # Replaced by he in 1989 revision
         'hi': 'hin',
         'ho': 'hmo',
         'hr': 'hrv',
@@ -2659,6 +3002,7 @@ class ISO639Utils(object):
         'hz': 'her',
         'ia': 'ina',
         'id': 'ind',
+        'in': 'ind',  # Replaced by id in 1989 revision
         'ie': 'ile',
         'ig': 'ibo',
         'ii': 'iii',
@@ -2773,6 +3117,7 @@ class ISO639Utils(object):
         'wo': 'wol',
         'xh': 'xho',
         'yi': 'yid',
+        'ji': 'yid',  # Replaced by yi in 1989 revision
         'yo': 'yor',
         'za': 'zha',
         'zh': 'zho',
@@ -3295,10 +3640,13 @@ class GeoUtils(object):
     }
 
     @classmethod
-    def random_ipv4(cls, code):
-        block = cls._country_ip_map.get(code.upper())
-        if not block:
-            return None
+    def random_ipv4(cls, code_or_block):
+        if len(code_or_block) == 2:
+            block = cls._country_ip_map.get(code_or_block.upper())
+            if not block:
+                return None
+        else:
+            block = code_or_block
         addr, preflen = block.split('/')
         addr_min = compat_struct_unpack('!L', socket.inet_aton(addr))[0]
         addr_max = addr_min | (0xffffffff >> int(preflen))
@@ -3313,7 +3661,7 @@ class PerRequestProxyHandler(compat_urllib_request.ProxyHandler):
             setattr(self, '%s_open' % type,
                     lambda r, proxy='__noproxy__', type=type, meth=self.proxy_open:
                         meth(r, proxy, type))
-        return compat_urllib_request.ProxyHandler.__init__(self, proxies)
+        compat_urllib_request.ProxyHandler.__init__(self, proxies)
 
     def proxy_open(self, req, proxy, type):
         req_proxy = req.headers.get('Ytdl-request-proxy')
@@ -3464,7 +3812,7 @@ def urshift(val, n):
 
 
 # Based on png2str() written by @gdkchan and improved by @yokrysty
-# Originally posted at https://github.com/rg3/youtube-dl/issues/9706
+# Originally posted at https://github.com/ytdl-org/youtube-dl/issues/9706
 def decode_png(png_data):
     # Reference: https://www.w3.org/TR/PNG/
     header = png_data[8:]
@@ -3579,7 +3927,7 @@ def write_xattr(path, key, value):
         if hasattr(xattr, 'set'):  # pyxattr
             # Unicode arguments are not supported in python-pyxattr until
             # version 0.5.0
-            # See https://github.com/rg3/youtube-dl/issues/5498
+            # See https://github.com/ytdl-org/youtube-dl/issues/5498
             pyxattr_required_version = '0.5.0'
             if version_tuple(xattr.__version__) < version_tuple(pyxattr_required_version):
                 # TODO: fallback to CLI tools
@@ -3625,9 +3973,9 @@ def write_xattr(path, key, value):
                     executable = 'xattr'
                     opts = ['-w', key, value]
 
-                cmd = ([encodeFilename(executable, True)] +
-                       [encodeArgument(o) for o in opts] +
-                       [encodeFilename(path, True)])
+                cmd = ([encodeFilename(executable, True)]
+                       + [encodeArgument(o) for o in opts]
+                       [encodeFilename(path, True)])
 
                 try:
                     p = subprocess.Popen(
@@ -3652,3 +4000,15 @@ def write_xattr(path, key, value):
                         "Couldn't find a tool to set the xattrs. "
                         "Install either the python 'xattr' module, "
                         "or the 'xattr' binary.")
+
+
+def random_birthday(year_field, month_field, day_field):
+    start_date = datetime.date(1950, 1, 1)
+    end_date = datetime.date(1995, 12, 31)
+    offset = random.randint(0, (end_date - start_date).days)
+    random_date = start_date + datetime.timedelta(offset)
+    return {
+        year_field: str(random_date.year),
+        month_field: str(random_date.month),
+        day_field: str(random_date.day),
+    }