import errno
import functools
import gzip
-import itertools
import io
+import itertools
import json
import locale
import math
import pipes
import platform
import re
-import ssl
import socket
-import struct
+import ssl
import subprocess
import sys
import tempfile
compat_parse_qs,
compat_socket_create_connection,
compat_str,
+ compat_struct_pack,
compat_urllib_error,
compat_urllib_parse,
compat_urllib_parse_urlencode,
shlex_quote,
)
+from .socks import (
+ ProxyType,
+ sockssocket,
+)
+
+
+def register_socks_protocols():
+ # "Register" SOCKS protocols
+ # In Python < 2.6.5, urlsplit() suffers from bug https://bugs.python.org/issue7904
+ # URLs with protocols not in urlparse.uses_netloc are not handled correctly
+ for scheme in ('socks', 'socks4', 'socks4a', 'socks5'):
+ if scheme not in compat_urlparse.uses_netloc:
+ compat_urlparse.uses_netloc.append(scheme)
+
# This is not clearly defined otherwise
compiled_regex_type = type(re.compile(''))
'wav',
'f4f', 'f4m', 'm3u8', 'smil')
+# needed for sanitizing filenames in restricted mode
+ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþÿ',
+ itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOUUUUYP', ['ss'],
+ 'aaaaaa', ['ae'], 'ceeeeiiiionoooooouuuuypy')))
+
def preferredencoding():
"""Get preferred encoding.
m = re.search(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>
Set is_id if this is not an arbitrary string, but an ID that should be kept if possible
"""
def replace_insane(char):
+ if restricted and char in ACCENT_CHARS:
+ return ACCENT_CHARS[char]
if char == '?' or ord(char) < 32 or ord(char) == 127:
return ''
elif char == '"':
self._params = params
def http_open(self, req):
+ conn_class = compat_http_client.HTTPConnection
+
+ socks_proxy = req.headers.get('Ytdl-socks-proxy')
+ if socks_proxy:
+ 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, compat_http_client.HTTPConnection, False),
+ _create_http_connection, self, conn_class, False),
req)
@staticmethod
https_response = http_response
+def make_socks_conn_class(base_class, socks_proxy):
+ assert issubclass(base_class, (
+ compat_http_client.HTTPConnection, compat_http_client.HTTPSConnection))
+
+ url_components = compat_urlparse.urlparse(socks_proxy)
+ if url_components.scheme.lower() == 'socks5':
+ socks_type = ProxyType.SOCKS5
+ elif url_components.scheme.lower() in ('socks', 'socks4'):
+ socks_type = ProxyType.SOCKS4
+ elif url_components.scheme.lower() == 'socks4a':
+ socks_type = ProxyType.SOCKS4A
+
+ proxy_args = (
+ socks_type,
+ url_components.hostname, url_components.port or 1080,
+ True, # Remote DNS
+ url_components.username, url_components.password
+ )
+
+ class SocksConnection(base_class):
+ def connect(self):
+ self.sock = sockssocket()
+ self.sock.setproxy(*proxy_args)
+ if type(self.timeout) in (int, float):
+ self.sock.settimeout(self.timeout)
+ self.sock.connect((self.host, self.port))
+
+ if isinstance(self, compat_http_client.HTTPSConnection):
+ if hasattr(self, '_context'): # Python > 2.6
+ self.sock = self._context.wrap_socket(
+ self.sock, server_hostname=self.host)
+ else:
+ self.sock = ssl.wrap_socket(self.sock)
+
+ return SocksConnection
+
+
class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
def __init__(self, params, https_conn_class=None, *args, **kwargs):
compat_urllib_request.HTTPSHandler.__init__(self, *args, **kwargs)
def https_open(self, req):
kwargs = {}
+ conn_class = self._https_conn_class
+
if hasattr(self, '_context'): # python > 2.6
kwargs['context'] = self._context
if hasattr(self, '_check_hostname'): # python 3.x
kwargs['check_hostname'] = self._check_hostname
+
+ socks_proxy = req.headers.get('Ytdl-socks-proxy')
+ if socks_proxy:
+ 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, self._https_conn_class, True),
+ _create_http_connection, self, conn_class, True),
req, **kwargs)
def intlist_to_bytes(xs):
if not xs:
return b''
- return struct_pack('%dB' % len(xs), *xs)
+ return compat_struct_pack('%dB' % len(xs), *xs)
# Cross-platform file locking
fragment=escape_rfc3986(url_parsed.fragment)
).geturl()
-try:
- struct.pack('!I', 0)
-except TypeError:
- # In Python 2.6 and 2.7.x < 2.7.7, struct requires a bytes argument
- # See https://bugs.python.org/issue19099
- def struct_pack(spec, *args):
- if isinstance(spec, compat_str):
- spec = spec.encode('ascii')
- return struct.pack(spec, *args)
-
- def struct_unpack(spec, *args):
- if isinstance(spec, compat_str):
- spec = spec.encode('ascii')
- return struct.unpack(spec, *args)
-else:
- struct_pack = struct.pack
- struct_unpack = struct.unpack
-
def read_batch_urls(batch_fd):
def fixup(url):
if proxy == '__noproxy__':
return None # No Proxy
+ if compat_urlparse.urlparse(proxy).scheme.lower() in ('socks', 'socks4', 'socks4a', 'socks5'):
+ req.add_header('Ytdl-socks-proxy', proxy)
+ # youtube-dl's http/https handlers do wrapping the socket with socks
+ return None
return compat_urllib_request.ProxyHandler.proxy_open(
self, req, proxy, type)