1 from __future__ import unicode_literals
16 import urllib.request as compat_urllib_request
17 except ImportError: # Python 2
18 import urllib2 as compat_urllib_request
21 import urllib.error as compat_urllib_error
22 except ImportError: # Python 2
23 import urllib2 as compat_urllib_error
26 import urllib.parse as compat_urllib_parse
27 except ImportError: # Python 2
28 import urllib as compat_urllib_parse
31 from urllib.parse import urlparse as compat_urllib_parse_urlparse
32 except ImportError: # Python 2
33 from urlparse import urlparse as compat_urllib_parse_urlparse
36 import urllib.parse as compat_urlparse
37 except ImportError: # Python 2
38 import urlparse as compat_urlparse
41 import http.cookiejar as compat_cookiejar
42 except ImportError: # Python 2
43 import cookielib as compat_cookiejar
46 import html.entities as compat_html_entities
47 except ImportError: # Python 2
48 import htmlentitydefs as compat_html_entities
51 import http.client as compat_http_client
52 except ImportError: # Python 2
53 import httplib as compat_http_client
56 from urllib.error import HTTPError as compat_HTTPError
57 except ImportError: # Python 2
58 from urllib2 import HTTPError as compat_HTTPError
61 from urllib.request import urlretrieve as compat_urlretrieve
62 except ImportError: # Python 2
63 from urllib import urlretrieve as compat_urlretrieve
67 from subprocess import DEVNULL
68 compat_subprocess_get_DEVNULL = lambda: DEVNULL
70 compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w')
73 import http.server as compat_http_server
75 import BaseHTTPServer as compat_http_server
78 from urllib.parse import unquote as compat_urllib_parse_unquote
80 def compat_urllib_parse_unquote_to_bytes(string):
81 """unquote_to_bytes('abc%20def') -> b'abc def'."""
82 # Note: strings are encoded as UTF-8. This is only an issue if it contains
83 # unescaped non-ASCII characters, which URIs should not.
85 # Is it a string-like object?
88 if isinstance(string, str):
89 string = string.encode('utf-8')
90 # string = encode('utf-8')
92 # python3 -> 2: must implicitly convert to bits
93 bits = bytes(string).split(b'%')
100 for item in bits[1:]:
102 append(item[:2].decode('hex'))
109 compat_urllib_parse_asciire = re.compile('([\x00-\x7f]+)')
111 def compat_urllib_parse_unquote(string, encoding='utf-8', errors='replace'):
112 """Replace %xx escapes by their single-character equivalent. The optional
113 encoding and errors parameters specify how to decode percent-encoded
114 sequences into Unicode characters, as accepted by the bytes.decode()
116 By default, percent-encoded sequences are decoded with UTF-8, and invalid
117 sequences are replaced by a placeholder character.
119 unquote('abc%20def') -> 'abc def'.
122 if '%' not in string:
130 bits = compat_urllib_parse_asciire.split(string)
133 for i in range(1, len(bits), 2):
134 foo = compat_urllib_parse_unquote_to_bytes(bits[i])
135 foo = foo.decode(encoding, errors)
140 if not isinstance(bar, unicode):
141 bar = bar.decode('utf-8')
146 compat_str = unicode # Python 2
151 compat_basestring = basestring # Python 2
153 compat_basestring = str
156 compat_chr = unichr # Python 2
161 from xml.etree.ElementTree import ParseError as compat_xml_parse_error
162 except ImportError: # Python 2.6
163 from xml.parsers.expat import ExpatError as compat_xml_parse_error
167 from urllib.parse import parse_qs as compat_parse_qs
168 except ImportError: # Python 2
169 # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib.
170 # Python 2's version is apparently totally broken
172 def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
173 encoding='utf-8', errors='replace'):
174 qs, _coerce_result = qs, compat_str
175 pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')]
177 for name_value in pairs:
178 if not name_value and not strict_parsing:
180 nv = name_value.split('=', 1)
183 raise ValueError("bad query field: %r" % (name_value,))
184 # Handle case of a control-name with no equal sign
185 if keep_blank_values:
189 if len(nv[1]) or keep_blank_values:
190 name = nv[0].replace('+', ' ')
191 name = compat_urllib_parse_unquote(
192 name, encoding=encoding, errors=errors)
193 name = _coerce_result(name)
194 value = nv[1].replace('+', ' ')
195 value = compat_urllib_parse_unquote(
196 value, encoding=encoding, errors=errors)
197 value = _coerce_result(value)
198 r.append((name, value))
201 def compat_parse_qs(qs, keep_blank_values=False, strict_parsing=False,
202 encoding='utf-8', errors='replace'):
204 pairs = _parse_qsl(qs, keep_blank_values, strict_parsing,
205 encoding=encoding, errors=errors)
206 for name, value in pairs:
207 if name in parsed_result:
208 parsed_result[name].append(value)
210 parsed_result[name] = [value]
214 from shlex import quote as shlex_quote
215 except ImportError: # Python < 3.3
217 if re.match(r'^[-_\w./]+$', s):
220 return "'" + s.replace("'", "'\"'\"'") + "'"
230 if sys.version_info >= (3, 0):
231 compat_getenv = os.getenv
232 compat_expanduser = os.path.expanduser
234 # Environment variables should be decoded with filesystem encoding.
235 # Otherwise it will fail if any non-ASCII characters present (see #3854 #3217 #2918)
237 def compat_getenv(key, default=None):
238 from .utils import get_filesystem_encoding
239 env = os.getenv(key, default)
241 env = env.decode(get_filesystem_encoding())
244 # HACK: The default implementations of os.path.expanduser from cpython do not decode
245 # environment variables with filesystem encoding. We will work around this by
246 # providing adjusted implementations.
247 # The following are os.path.expanduser implementations from cpython 2.7.8 stdlib
248 # for different platforms with correct environment variables decoding.
250 if os.name == 'posix':
251 def compat_expanduser(path):
252 """Expand ~ and ~user constructions. If user or $HOME is unknown,
254 if not path.startswith('~'):
256 i = path.find('/', 1)
260 if 'HOME' not in os.environ:
262 userhome = pwd.getpwuid(os.getuid()).pw_dir
264 userhome = compat_getenv('HOME')
268 pwent = pwd.getpwnam(path[1:i])
271 userhome = pwent.pw_dir
272 userhome = userhome.rstrip('/')
273 return (userhome + path[i:]) or '/'
274 elif os.name == 'nt' or os.name == 'ce':
275 def compat_expanduser(path):
276 """Expand ~ and ~user constructs.
278 If user or $HOME is unknown, do nothing."""
282 while i < n and path[i] not in '/\\':
285 if 'HOME' in os.environ:
286 userhome = compat_getenv('HOME')
287 elif 'USERPROFILE' in os.environ:
288 userhome = compat_getenv('USERPROFILE')
289 elif 'HOMEPATH' not in os.environ:
293 drive = compat_getenv('HOMEDRIVE')
296 userhome = os.path.join(drive, compat_getenv('HOMEPATH'))
299 userhome = os.path.join(os.path.dirname(userhome), path[1:i])
301 return userhome + path[i:]
303 compat_expanduser = os.path.expanduser
306 if sys.version_info < (3, 0):
308 from .utils import preferredencoding
309 print(s.encode(preferredencoding(), 'xmlcharrefreplace'))
312 assert isinstance(s, compat_str)
317 subprocess_check_output = subprocess.check_output
318 except AttributeError:
319 def subprocess_check_output(*args, **kwargs):
320 assert 'input' not in kwargs
321 p = subprocess.Popen(*args, stdout=subprocess.PIPE, **kwargs)
322 output, _ = p.communicate()
325 raise subprocess.CalledProcessError(ret, p.args, output=output)
328 if sys.version_info < (3, 0) and sys.platform == 'win32':
329 def compat_getpass(prompt, *args, **kwargs):
330 if isinstance(prompt, compat_str):
331 from .utils import preferredencoding
332 prompt = prompt.encode(preferredencoding())
333 return getpass.getpass(prompt, *args, **kwargs)
335 compat_getpass = getpass.getpass
337 # Old 2.6 and 2.7 releases require kwargs to be bytes
341 _testfunc(**{'x': 0})
343 def compat_kwargs(kwargs):
344 return dict((bytes(k), v) for k, v in kwargs.items())
346 compat_kwargs = lambda kwargs: kwargs
349 if sys.version_info < (2, 7):
350 def compat_socket_create_connection(address, timeout, source_address=None):
353 for res in socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM):
354 af, socktype, proto, canonname, sa = res
357 sock = socket.socket(af, socktype, proto)
358 sock.settimeout(timeout)
360 sock.bind(source_address)
363 except socket.error as _:
370 raise socket.error("getaddrinfo returns an empty list")
372 compat_socket_create_connection = socket.create_connection
375 # Fix https://github.com/rg3/youtube-dl/issues/4223
376 # See http://bugs.python.org/issue9161 for what is broken
377 def workaround_optparse_bug9161():
378 op = optparse.OptionParser()
379 og = optparse.OptionGroup(op, 'foo')
383 real_add_option = optparse.OptionGroup.add_option
385 def _compat_add_option(self, *args, **kwargs):
387 v.encode('ascii', 'replace') if isinstance(v, compat_str)
389 bargs = [enc(a) for a in args]
391 (k, enc(v)) for k, v in kwargs.items())
392 return real_add_option(self, *bargs, **bkwargs)
393 optparse.OptionGroup.add_option = _compat_add_option
395 if hasattr(shutil, 'get_terminal_size'): # Python >= 3.3
396 compat_get_terminal_size = shutil.get_terminal_size
398 _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines'])
400 def compat_get_terminal_size():
401 columns = compat_getenv('COLUMNS', None)
403 columns = int(columns)
406 lines = compat_getenv('LINES', None)
413 sp = subprocess.Popen(
415 stdout=subprocess.PIPE, stderr=subprocess.PIPE)
416 out, err = sp.communicate()
417 lines, columns = map(int, out.split())
420 return _terminal_size(columns, lines)
423 itertools.count(start=0, step=1)
424 compat_itertools_count = itertools.count
425 except TypeError: # Python 2.6
426 def compat_itertools_count(start=0, step=1):
438 'compat_get_terminal_size',
441 'compat_html_entities',
442 'compat_http_client',
443 'compat_http_server',
444 'compat_itertools_count',
449 'compat_socket_create_connection',
451 'compat_subprocess_get_DEVNULL',
452 'compat_urllib_error',
453 'compat_urllib_parse',
454 'compat_urllib_parse_asciire',
455 'compat_urllib_parse_unquote',
456 'compat_urllib_parse_unquote_to_bytes',
457 'compat_urllib_parse_urlparse',
458 'compat_urllib_request',
460 'compat_urlretrieve',
461 'compat_xml_parse_error',
463 'subprocess_check_output',
464 'workaround_optparse_bug9161',