def sanitize_url(url):
# Prepend protocol-less URLs with `http:` scheme in order to mitigate
# the number of unwanted failures due to missing protocol
- if url.startswith('//'):
+ if url is None:
+ return
+ elif url.startswith('//'):
return 'http:%s' % url
# Fix some common typos seen so far
COMMON_TYPOS = (
parts.hostname if parts.port is None
else '%s:%d' % (parts.hostname, parts.port))))
auth_payload = base64.b64encode(
- ('%s:%s' % (parts.username, parts.password or '')).encode('utf-8'))
- return url, 'Basic ' + auth_payload.decode('utf-8')
+ ('%s:%s' % (parts.username, parts.password or '')).encode())
+ return url, f'Basic {auth_payload.decode()}'
def sanitized_Request(url, *args, **kwargs):
context.check_hostname = opts_check_certificate
if params.get('legacyserverconnect'):
context.options |= 4 # SSL_OP_LEGACY_SERVER_CONNECT
+ # Allow use of weaker ciphers in Python 3.10+. See https://bugs.python.org/issue43998
+ context.set_ciphers('DEFAULT')
context.verify_mode = ssl.CERT_REQUIRED if opts_check_certificate else ssl.CERT_NONE
if opts_check_certificate:
if has_certifi and 'no-certifi' not in params.get('compat_opts', []):
except ssl.SSLError:
# enum_certificates is not present in mingw python. See https://github.com/yt-dlp/yt-dlp/issues/1151
if sys.platform == 'win32' and hasattr(ssl, 'enum_certificates'):
- # Create a new context to discard any certificates that were already loaded
- context = ssl.SSLContext(ssl.PROTOCOL_TLS_CLIENT)
- context.check_hostname, context.verify_mode = True, ssl.CERT_REQUIRED
for storename in ('CA', 'ROOT'):
_ssl_load_windows_store_certs(context, storename)
context.set_default_verify_paths()
+ client_certfile = params.get('client_certificate')
+ if client_certfile:
+ try:
+ context.load_cert_chain(
+ client_certfile, keyfile=params.get('client_certificate_key'),
+ password=params.get('client_certificate_password'))
+ except ssl.SSLError:
+ raise YoutubeDLError('Unable to load client certificate')
return YoutubeDLHTTPSHandler(params, context=context, **kwargs)
location = resp.headers.get('Location')
if location:
# As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
- location = location.encode('iso-8859-1').decode('utf-8')
+ location = location.encode('iso-8859-1').decode()
location_escaped = escape_url(location)
if location != location_escaped:
del resp.headers['Location']
conn_class = make_socks_conn_class(conn_class, socks_proxy)
del req.headers['Ytdl-socks-proxy']
- return self.do_open(functools.partial(
- _create_http_connection, self, conn_class, True),
- req, **kwargs)
+ try:
+ return self.do_open(
+ functools.partial(_create_http_connection, self, conn_class, True), req, **kwargs)
+ except urllib.error.URLError as e:
+ if (isinstance(e.reason, ssl.SSLError)
+ and getattr(e.reason, 'reason', None) == 'SSLV3_ALERT_HANDSHAKE_FAILURE'):
+ raise YoutubeDLError('SSLV3_ALERT_HANDSHAKE_FAILURE: Try using --legacy-server-connect')
+ raise
class YoutubeDLCookieJar(compat_cookiejar.MozillaCookieJar):
'CookieFileEntry',
('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value'))
- def save(self, filename=None, ignore_discard=False, ignore_expires=False):
+ def __init__(self, filename=None, *args, **kwargs):
+ super().__init__(None, *args, **kwargs)
+ if self.is_path(filename):
+ filename = os.fspath(filename)
+ self.filename = filename
+
+ @staticmethod
+ def _true_or_false(cndn):
+ return 'TRUE' if cndn else 'FALSE'
+
+ @staticmethod
+ def is_path(file):
+ return isinstance(file, (str, bytes, os.PathLike))
+
+ @contextlib.contextmanager
+ def open(self, file, *, write=False):
+ if self.is_path(file):
+ with open(file, 'w' if write else 'r', encoding='utf-8') as f:
+ yield f
+ else:
+ if write:
+ file.truncate(0)
+ yield file
+
+ def _really_save(self, f, ignore_discard=False, ignore_expires=False):
+ now = time.time()
+ for cookie in self:
+ if (not ignore_discard and cookie.discard
+ or not ignore_expires and cookie.is_expired(now)):
+ continue
+ name, value = cookie.name, cookie.value
+ if value is None:
+ # cookies.txt regards 'Set-Cookie: foo' as a cookie
+ # with no name, whereas http.cookiejar regards it as a
+ # cookie with no value.
+ name, value = '', name
+ f.write('%s\n' % '\t'.join((
+ cookie.domain,
+ self._true_or_false(cookie.domain.startswith('.')),
+ cookie.path,
+ self._true_or_false(cookie.secure),
+ str_or_none(cookie.expires, default=''),
+ name, value
+ )))
+
+ def save(self, filename=None, *args, **kwargs):
"""
Save cookies to a file.
+ Code is taken from CPython 3.6
+ https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """
- Most of the code is taken from CPython 3.8 and slightly adapted
- to support cookie files with UTF-8 in both python 2 and 3.
- """
if filename is None:
if self.filename is not None:
filename = self.filename
else:
raise ValueError(compat_cookiejar.MISSING_FILENAME_TEXT)
- # Store session cookies with `expires` set to 0 instead of an empty
- # string
+ # 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
- with open(filename, 'w', encoding='utf-8') as f:
+ with self.open(filename, write=True) as f:
f.write(self._HEADER)
- now = time.time()
- for cookie in self:
- if not ignore_discard and cookie.discard:
- continue
- if not ignore_expires and cookie.is_expired(now):
- continue
- if cookie.secure:
- secure = 'TRUE'
- else:
- secure = 'FALSE'
- if cookie.domain.startswith('.'):
- initial_dot = 'TRUE'
- else:
- initial_dot = 'FALSE'
- if cookie.expires is not None:
- expires = compat_str(cookie.expires)
- else:
- expires = ''
- if cookie.value is None:
- # cookies.txt regards 'Set-Cookie: foo' as a cookie
- # with no name, whereas http.cookiejar regards it as a
- # cookie with no value.
- name = ''
- value = cookie.name
- else:
- name = cookie.name
- value = cookie.value
- f.write(
- '\t'.join([cookie.domain, initial_dot, cookie.path,
- secure, expires, name, value]) + '\n')
+ self._really_save(f, *args, **kwargs)
def load(self, filename=None, ignore_discard=False, ignore_expires=False):
"""Load cookies from a file."""
return line
cf = io.StringIO()
- with open(filename, encoding='utf-8') as f:
+ with self.open(filename) as f:
for line in f:
try:
cf.write(prepare_line(line))
CONTENT_HEADERS = ("content-length", "content-type")
# NB: don't use dict comprehension for python 2.6 compatibility
newheaders = {k: v for k, v in req.headers.items() if k.lower() not in CONTENT_HEADERS}
+
+ # A 303 must either use GET or HEAD for subsequent request
+ # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4
+ if code == 303 and m != 'HEAD':
+ m = 'GET'
+ # 301 and 302 redirects are commonly turned into a GET from a POST
+ # for subsequent requests by browsers, so we'll do the same.
+ # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.2
+ # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.3
+ if code in (301, 302) and m == 'POST':
+ m = 'GET'
+
return compat_urllib_request.Request(
newurl, headers=newheaders, origin_req_host=req.origin_req_host,
- unverifiable=True)
+ unverifiable=True, method=m)
def extract_timezone(date_str):
def datetime_from_str(date_str, precision='auto', format='%Y%m%d'):
- """
- Return a datetime object from a string in the format YYYYMMDD or
- (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
-
- format: string date format used to return datetime object from
- precision: round the time portion of a datetime object.
- auto|microsecond|second|minute|hour|day.
- auto: round to the unit provided in date_str (if applicable).
+ R"""
+ Return a datetime object from a string.
+ Supported format:
+ (now|today|yesterday|DATE)([+-]\d+(microsecond|second|minute|hour|day|week|month|year)s?)?
+
+ @param format strftime format of DATE
+ @param precision Round the datetime object: auto|microsecond|second|minute|hour|day
+ auto: round to the unit provided in date_str (if applicable).
"""
auto_precision = False
if precision == 'auto':
if date_str == 'yesterday':
return today - datetime.timedelta(days=1)
match = re.match(
- r'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)(s)?',
+ r'(?P<start>.+)(?P<sign>[+-])(?P<time>\d+)(?P<unit>microsecond|second|minute|hour|day|week|month|year)s?',
date_str)
if match is not None:
start_time = datetime_from_str(match.group('start'), precision, format)
def date_from_str(date_str, format='%Y%m%d', strict=False):
- """
- Return a datetime object from a string in the format YYYYMMDD or
- (now|today|yesterday|date)[+-][0-9](microsecond|second|minute|hour|day|week|month|year)(s)?
+ R"""
+ Return a date object from a string using datetime_from_str
- If "strict", only (now|today)[+-][0-9](day|week|month|year)(s)? is allowed
-
- format: string date format used to return datetime object from
+ @param strict Restrict allowed patterns to "YYYYMMDD" and
+ (now|today|yesterday)(-\d+(day|week|month|year)s?)?
"""
- if strict and not re.fullmatch(r'\d{8}|(now|today)[+-]\d+(day|week|month|year)(s)?', date_str):
- raise ValueError(f'Invalid date format {date_str}')
+ if strict and not re.fullmatch(r'\d{8}|(now|today|yesterday)(-\d+(day|week|month|year)s?)?', date_str):
+ raise ValueError(f'Invalid date format "{date_str}"')
return datetime_from_str(date_str, precision='microsecond', format=format).date()
assert isinstance(s, str)
out = out or sys.stderr
- from .compat import WINDOWS_VT_MODE # Must be imported locally
- if WINDOWS_VT_MODE:
+ if compat_os_name == 'nt' and supports_terminal_sequences(out):
s = re.sub(r'([\r\n]+)', r' \1', s)
if 'b' in getattr(out, 'mode', ''):
# a bytestring, but since unicode_literals turns
# every string into a unicode string, it fails.
return
- title_bytes = title.encode('utf-8')
+ title_bytes = title.encode()
buf = ctypes.create_string_buffer(len(title_bytes))
buf.value = title_bytes
try:
def urljoin(base, path):
if isinstance(path, bytes):
- path = path.decode('utf-8')
+ path = path.decode()
if not isinstance(path, compat_str) or not path:
return None
if re.match(r'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path):
return path
if isinstance(base, bytes):
- base = base.decode('utf-8')
+ base = base.decode()
if not isinstance(base, compat_str) or not re.match(
r'^(?:https?:)?//', base):
return None
class LazyList(collections.abc.Sequence):
- ''' Lazy immutable list from an iterable
- Note that slices of a LazyList are lists and not LazyList'''
+ """Lazy immutable list from an iterable
+ Note that slices of a LazyList are lists and not LazyList"""
class IndexError(IndexError):
pass
def __init__(self, iterable, *, reverse=False, _cache=None):
- self.__iterable = iter(iterable)
- self.__cache = [] if _cache is None else _cache
- self.__reversed = reverse
+ self._iterable = iter(iterable)
+ self._cache = [] if _cache is None else _cache
+ self._reversed = reverse
def __iter__(self):
- if self.__reversed:
+ if self._reversed:
# We need to consume the entire iterable to iterate in reverse
yield from self.exhaust()
return
- yield from self.__cache
- for item in self.__iterable:
- self.__cache.append(item)
+ yield from self._cache
+ for item in self._iterable:
+ self._cache.append(item)
yield item
- def __exhaust(self):
- self.__cache.extend(self.__iterable)
- # Discard the emptied iterable to make it pickle-able
- self.__iterable = []
- return self.__cache
+ def _exhaust(self):
+ self._cache.extend(self._iterable)
+ self._iterable = [] # Discard the emptied iterable to make it pickle-able
+ return self._cache
def exhaust(self):
- ''' Evaluate the entire iterable '''
- return self.__exhaust()[::-1 if self.__reversed else 1]
+ """Evaluate the entire iterable"""
+ return self._exhaust()[::-1 if self._reversed else 1]
@staticmethod
- def __reverse_index(x):
+ def _reverse_index(x):
return None if x is None else -(x + 1)
def __getitem__(self, idx):
if isinstance(idx, slice):
- if self.__reversed:
- idx = slice(self.__reverse_index(idx.start), self.__reverse_index(idx.stop), -(idx.step or 1))
+ if self._reversed:
+ idx = slice(self._reverse_index(idx.start), self._reverse_index(idx.stop), -(idx.step or 1))
start, stop, step = idx.start, idx.stop, idx.step or 1
elif isinstance(idx, int):
- if self.__reversed:
- idx = self.__reverse_index(idx)
+ if self._reversed:
+ idx = self._reverse_index(idx)
start, stop, step = idx, idx, 0
else:
raise TypeError('indices must be integers or slices')
or (stop is None and step > 0)):
# We need to consume the entire iterable to be able to slice from the end
# Obviously, never use this with infinite iterables
- self.__exhaust()
+ self._exhaust()
try:
- return self.__cache[idx]
+ return self._cache[idx]
except IndexError as e:
raise self.IndexError(e) from e
- n = max(start or 0, stop or 0) - len(self.__cache) + 1
+ n = max(start or 0, stop or 0) - len(self._cache) + 1
if n > 0:
- self.__cache.extend(itertools.islice(self.__iterable, n))
+ self._cache.extend(itertools.islice(self._iterable, n))
try:
- return self.__cache[idx]
+ return self._cache[idx]
except IndexError as e:
raise self.IndexError(e) from e
def __bool__(self):
try:
- self[-1] if self.__reversed else self[0]
+ self[-1] if self._reversed else self[0]
except self.IndexError:
return False
return True
def __len__(self):
- self.__exhaust()
- return len(self.__cache)
+ self._exhaust()
+ return len(self._cache)
def __reversed__(self):
- return type(self)(self.__iterable, reverse=not self.__reversed, _cache=self.__cache)
+ return type(self)(self._iterable, reverse=not self._reversed, _cache=self._cache)
def __copy__(self):
- return type(self)(self.__iterable, reverse=self.__reversed, _cache=self.__cache)
+ return type(self)(self._iterable, reverse=self._reversed, _cache=self._cache)
def __repr__(self):
# repr and str should mimic a list. So we exhaust the iterable
for k, v in data.items():
out += b'--' + boundary.encode('ascii') + b'\r\n'
if isinstance(k, compat_str):
- k = k.encode('utf-8')
+ k = k.encode()
if isinstance(v, compat_str):
- v = v.encode('utf-8')
+ v = v.encode()
# 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'
(b'\xff\xfe', 'utf-16-le'),
(b'\xfe\xff', 'utf-16-be'),
]
+
+ encoding = 'utf-8'
for bom, enc in BOMS:
- if first_bytes.startswith(bom):
- s = first_bytes[len(bom):].decode(enc, 'replace')
- break
- else:
- s = first_bytes.decode('utf-8', 'replace')
+ while first_bytes.startswith(bom):
+ encoding, first_bytes = enc, first_bytes[len(bom):]
- return re.match(r'^\s*<', s)
+ return re.match(r'^\s*<', first_bytes.decode(encoding, 'replace'))
def determine_protocol(info_dict):
'Couldn\'t find a tool to set the xattrs. Install either the python "xattr" or "pyxattr" modules or the '
+ ('"xattr" binary' if sys.platform != 'linux' else 'GNU "attr" package (which contains the "setfattr" tool)'))
- value = value.decode('utf-8')
+ value = value.decode()
try:
p = Popen(
[exe, '-w', key, value, path] if exe == 'xattr' else [exe, '-n', key, '-v', value, path],
net_location += ':' + urllib.parse.quote(iri_parts.password, safe=r"!$%&'()*+,~")
net_location += '@'
- net_location += iri_parts.hostname.encode('idna').decode('utf-8') # Punycode for Unicode hostnames.
+ net_location += iri_parts.hostname.encode('idna').decode() # Punycode for Unicode hostnames.
# The 'idna' encoding produces ASCII text.
if iri_parts.port is not None and iri_parts.port != 80:
net_location += ':' + str(iri_parts.port)
}
if headers:
header_data.update(headers)
- header_b64 = base64.b64encode(json.dumps(header_data).encode('utf-8'))
- payload_b64 = base64.b64encode(json.dumps(payload_data).encode('utf-8'))
- h = hmac.new(key.encode('utf-8'), header_b64 + b'.' + payload_b64, hashlib.sha256)
+ header_b64 = base64.b64encode(json.dumps(header_data).encode())
+ payload_b64 = base64.b64encode(json.dumps(payload_data).encode())
+ h = hmac.new(key.encode(), header_b64 + b'.' + payload_b64, hashlib.sha256)
signature_b64 = base64.b64encode(h.digest())
token = header_b64 + b'.' + payload_b64 + b'.' + signature_b64
return token
class Config:
own_args = None
+ parsed_args = None
filename = None
__initialized = False
def __init__(self, parser, label=None):
- self._parser, self.label = parser, label
+ self.parser, self.label = parser, label
self._loaded_paths, self.configs = set(), []
def init(self, args=None, filename=None):
return False
self._loaded_paths.add(location)
- self.__initialized = True
- self.own_args, self.filename = args, filename
- for location in self._parser.parse_args(args)[0].config_locations or []:
+ self.own_args, self.__initialized = args, True
+ opts, _ = self.parser.parse_known_args(args)
+ self.parsed_args, self.filename = args, filename
+
+ for location in opts.config_locations or []:
location = os.path.join(directory, expand_path(location))
if os.path.isdir(location):
location = os.path.join(location, 'yt-dlp.conf')
if not os.path.exists(location):
- self._parser.error(f'config location {location} does not exist')
+ self.parser.error(f'config location {location} does not exist')
self.append_config(self.read_file(location), location)
return True
return opts
def append_config(self, *args, label=None):
- config = type(self)(self._parser, label)
+ config = type(self)(self.parser, label)
config._loaded_paths = self._loaded_paths
if config.init(*args):
self.configs.append(config)
def all_args(self):
for config in reversed(self.configs):
yield from config.all_args
- yield from self.own_args or []
+ yield from self.parsed_args or []
+
+ def parse_known_args(self, **kwargs):
+ return self.parser.parse_known_args(self.all_args, **kwargs)
def parse_args(self):
- return self._parser.parse_args(self.all_args)
+ return self.parser.parse_args(self.all_args)
class WebSocketsWrapper():
class classproperty:
def __init__(self, f):
+ functools.update_wrapper(self, f)
self.f = f
def __get__(self, _, cls):
return self.f(cls)
-def Namespace(**kwargs):
- return collections.namedtuple('Namespace', kwargs)(**kwargs)
+class Namespace:
+ """Immutable namespace"""
+
+ def __init__(self, **kwargs):
+ self._dict = kwargs
+
+ def __getattr__(self, attr):
+ return self._dict[attr]
+
+ def __contains__(self, item):
+ return item in self._dict.values()
+
+ def __iter__(self):
+ return iter(self._dict.items())
+
+ def __repr__(self):
+ return f'{type(self).__name__}({", ".join(f"{k}={v}" for k, v in self)})'
# Deprecated