1 from __future__ import unicode_literals
17 import xml.etree.ElementTree
21 import urllib.request as compat_urllib_request
22 except ImportError: # Python 2
23 import urllib2 as compat_urllib_request
26 import urllib.error as compat_urllib_error
27 except ImportError: # Python 2
28 import urllib2 as compat_urllib_error
31 import urllib.parse as compat_urllib_parse
32 except ImportError: # Python 2
33 import urllib as compat_urllib_parse
36 from urllib.parse import urlparse as compat_urllib_parse_urlparse
37 except ImportError: # Python 2
38 from urlparse import urlparse as compat_urllib_parse_urlparse
41 import urllib.parse as compat_urlparse
42 except ImportError: # Python 2
43 import urlparse as compat_urlparse
46 import urllib.response as compat_urllib_response
47 except ImportError: # Python 2
48 import urllib as compat_urllib_response
51 import http.cookiejar as compat_cookiejar
52 except ImportError: # Python 2
53 import cookielib as compat_cookiejar
56 import http.cookies as compat_cookies
57 except ImportError: # Python 2
58 import Cookie as compat_cookies
61 import html.entities as compat_html_entities
62 except ImportError: # Python 2
63 import htmlentitydefs as compat_html_entities
66 import http.client as compat_http_client
67 except ImportError: # Python 2
68 import httplib as compat_http_client
71 from urllib.error import HTTPError as compat_HTTPError
72 except ImportError: # Python 2
73 from urllib2 import HTTPError as compat_HTTPError
76 from urllib.request import urlretrieve as compat_urlretrieve
77 except ImportError: # Python 2
78 from urllib import urlretrieve as compat_urlretrieve
81 from html.parser import HTMLParser as compat_HTMLParser
82 except ImportError: # Python 2
83 from HTMLParser import HTMLParser as compat_HTMLParser
87 from subprocess import DEVNULL
88 compat_subprocess_get_DEVNULL = lambda: DEVNULL
90 compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w')
93 import http.server as compat_http_server
95 import BaseHTTPServer as compat_http_server
98 compat_str = unicode # Python 2
103 from urllib.parse import unquote_to_bytes as compat_urllib_parse_unquote_to_bytes
104 from urllib.parse import unquote as compat_urllib_parse_unquote
105 from urllib.parse import unquote_plus as compat_urllib_parse_unquote_plus
106 except ImportError: # Python 2
107 _asciire = (compat_urllib_parse._asciire if hasattr(compat_urllib_parse, '_asciire')
108 else re.compile('([\x00-\x7f]+)'))
110 # HACK: The following are the correct unquote_to_bytes, unquote and unquote_plus
111 # implementations from cpython 3.4.3's stdlib. Python 2's version
112 # is apparently broken (see https://github.com/rg3/youtube-dl/pull/6244)
114 def compat_urllib_parse_unquote_to_bytes(string):
115 """unquote_to_bytes('abc%20def') -> b'abc def'."""
116 # Note: strings are encoded as UTF-8. This is only an issue if it contains
117 # unescaped non-ASCII characters, which URIs should not.
119 # Is it a string-like object?
122 if isinstance(string, compat_str):
123 string = string.encode('utf-8')
124 bits = string.split(b'%')
129 for item in bits[1:]:
131 append(compat_urllib_parse._hextochr[item[:2]])
138 def compat_urllib_parse_unquote(string, encoding='utf-8', errors='replace'):
139 """Replace %xx escapes by their single-character equivalent. The optional
140 encoding and errors parameters specify how to decode percent-encoded
141 sequences into Unicode characters, as accepted by the bytes.decode()
143 By default, percent-encoded sequences are decoded with UTF-8, and invalid
144 sequences are replaced by a placeholder character.
146 unquote('abc%20def') -> 'abc def'.
148 if '%' not in string:
155 bits = _asciire.split(string)
158 for i in range(1, len(bits), 2):
159 append(compat_urllib_parse_unquote_to_bytes(bits[i]).decode(encoding, errors))
163 def compat_urllib_parse_unquote_plus(string, encoding='utf-8', errors='replace'):
164 """Like unquote(), but also replace plus signs by spaces, as required for
165 unquoting HTML form values.
167 unquote_plus('%7e/abc+def') -> '~/abc def'
169 string = string.replace('+', ' ')
170 return compat_urllib_parse_unquote(string, encoding, errors)
173 from urllib.parse import urlencode as compat_urllib_parse_urlencode
174 except ImportError: # Python 2
175 # Python 2 will choke in urlencode on mixture of byte and unicode strings.
176 # Possible solutions are to either port it from python 3 with all
177 # the friends or manually ensure input query contains only byte strings.
178 # We will stick with latter thus recursively encoding the whole query.
179 def compat_urllib_parse_urlencode(query, doseq=0, encoding='utf-8'):
181 if isinstance(e, dict):
183 elif isinstance(e, (list, tuple,)):
184 list_e = encode_list(e)
185 e = tuple(list_e) if isinstance(e, tuple) else list_e
186 elif isinstance(e, compat_str):
187 e = e.encode(encoding)
191 return dict((encode_elem(k), encode_elem(v)) for k, v in d.items())
194 return [encode_elem(e) for e in l]
196 return compat_urllib_parse.urlencode(encode_elem(query), doseq=doseq)
199 from urllib.request import DataHandler as compat_urllib_request_DataHandler
200 except ImportError: # Python < 3.4
201 # Ported from CPython 98774:1733b3bd46db, Lib/urllib/request.py
202 class compat_urllib_request_DataHandler(compat_urllib_request.BaseHandler):
203 def data_open(self, req):
204 # data URLs as specified in RFC 2397.
206 # ignores POSTed data
209 # dataurl := "data:" [ mediatype ] [ ";base64" ] "," data
210 # mediatype := [ type "/" subtype ] *( ";" parameter )
212 # parameter := attribute "=" value
213 url = req.get_full_url()
215 scheme, data = url.split(':', 1)
216 mediatype, data = data.split(',', 1)
218 # even base64 encoded data URLs might be quoted so unquote in any case:
219 data = compat_urllib_parse_unquote_to_bytes(data)
220 if mediatype.endswith(';base64'):
221 data = binascii.a2b_base64(data)
222 mediatype = mediatype[:-7]
225 mediatype = 'text/plain;charset=US-ASCII'
227 headers = email.message_from_string(
228 'Content-type: %s\nContent-length: %d\n' % (mediatype, len(data)))
230 return compat_urllib_response.addinfourl(io.BytesIO(data), headers, url)
233 compat_basestring = basestring # Python 2
235 compat_basestring = str
238 compat_chr = unichr # Python 2
243 from xml.etree.ElementTree import ParseError as compat_xml_parse_error
244 except ImportError: # Python 2.6
245 from xml.parsers.expat import ExpatError as compat_xml_parse_error
247 if sys.version_info[0] >= 3:
248 compat_etree_fromstring = xml.etree.ElementTree.fromstring
250 # python 2.x tries to encode unicode strings with ascii (see the
251 # XMLParser._fixtext method)
252 etree = xml.etree.ElementTree
255 _etree_iter = etree.Element.iter
256 except AttributeError: # Python <=2.6
257 def _etree_iter(root):
258 for el in root.findall('*'):
260 for sub in _etree_iter(el):
263 # on 2.6 XML doesn't have a parser argument, function copied from CPython
265 def _XML(text, parser=None):
267 parser = etree.XMLParser(target=etree.TreeBuilder())
269 return parser.close()
271 def _element_factory(*args, **kwargs):
272 el = etree.Element(*args, **kwargs)
273 for k, v in el.items():
274 if isinstance(v, bytes):
275 el.set(k, v.decode('utf-8'))
278 def compat_etree_fromstring(text):
279 doc = _XML(text, parser=etree.XMLParser(target=etree.TreeBuilder(element_factory=_element_factory)))
280 for el in _etree_iter(doc):
281 if el.text is not None and isinstance(el.text, bytes):
282 el.text = el.text.decode('utf-8')
285 if sys.version_info < (2, 7):
286 # Here comes the crazy part: In 2.6, if the xpath is a unicode,
287 # .//node does not match if a node is a direct child of . !
288 def compat_xpath(xpath):
289 if isinstance(xpath, compat_str):
290 xpath = xpath.encode('ascii')
293 compat_xpath = lambda xpath: xpath
296 from urllib.parse import parse_qs as compat_parse_qs
297 except ImportError: # Python 2
298 # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib.
299 # Python 2's version is apparently totally broken
301 def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
302 encoding='utf-8', errors='replace'):
303 qs, _coerce_result = qs, compat_str
304 pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')]
306 for name_value in pairs:
307 if not name_value and not strict_parsing:
309 nv = name_value.split('=', 1)
312 raise ValueError('bad query field: %r' % (name_value,))
313 # Handle case of a control-name with no equal sign
314 if keep_blank_values:
318 if len(nv[1]) or keep_blank_values:
319 name = nv[0].replace('+', ' ')
320 name = compat_urllib_parse_unquote(
321 name, encoding=encoding, errors=errors)
322 name = _coerce_result(name)
323 value = nv[1].replace('+', ' ')
324 value = compat_urllib_parse_unquote(
325 value, encoding=encoding, errors=errors)
326 value = _coerce_result(value)
327 r.append((name, value))
330 def compat_parse_qs(qs, keep_blank_values=False, strict_parsing=False,
331 encoding='utf-8', errors='replace'):
333 pairs = _parse_qsl(qs, keep_blank_values, strict_parsing,
334 encoding=encoding, errors=errors)
335 for name, value in pairs:
336 if name in parsed_result:
337 parsed_result[name].append(value)
339 parsed_result[name] = [value]
343 from shlex import quote as shlex_quote
344 except ImportError: # Python < 3.3
346 if re.match(r'^[-_\w./]+$', s):
349 return "'" + s.replace("'", "'\"'\"'") + "'"
352 if sys.version_info >= (2, 7, 3):
353 compat_shlex_split = shlex.split
355 # Working around shlex issue with unicode strings on some python 2
356 # versions (see http://bugs.python.org/issue1548891)
357 def compat_shlex_split(s, comments=False, posix=True):
358 if isinstance(s, compat_str):
359 s = s.encode('utf-8')
360 return shlex.split(s, comments, posix)
370 compat_os_name = os._name if os.name == 'java' else os.name
373 if sys.version_info >= (3, 0):
374 compat_getenv = os.getenv
375 compat_expanduser = os.path.expanduser
377 def compat_setenv(key, value, env=os.environ):
380 # Environment variables should be decoded with filesystem encoding.
381 # Otherwise it will fail if any non-ASCII characters present (see #3854 #3217 #2918)
383 def compat_getenv(key, default=None):
384 from .utils import get_filesystem_encoding
385 env = os.getenv(key, default)
387 env = env.decode(get_filesystem_encoding())
390 def compat_setenv(key, value, env=os.environ):
392 from .utils import get_filesystem_encoding
393 return v.encode(get_filesystem_encoding()) if isinstance(v, compat_str) else v
394 env[encode(key)] = encode(value)
396 # HACK: The default implementations of os.path.expanduser from cpython do not decode
397 # environment variables with filesystem encoding. We will work around this by
398 # providing adjusted implementations.
399 # The following are os.path.expanduser implementations from cpython 2.7.8 stdlib
400 # for different platforms with correct environment variables decoding.
402 if compat_os_name == 'posix':
403 def compat_expanduser(path):
404 """Expand ~ and ~user constructions. If user or $HOME is unknown,
406 if not path.startswith('~'):
408 i = path.find('/', 1)
412 if 'HOME' not in os.environ:
414 userhome = pwd.getpwuid(os.getuid()).pw_dir
416 userhome = compat_getenv('HOME')
420 pwent = pwd.getpwnam(path[1:i])
423 userhome = pwent.pw_dir
424 userhome = userhome.rstrip('/')
425 return (userhome + path[i:]) or '/'
426 elif compat_os_name == 'nt' or compat_os_name == 'ce':
427 def compat_expanduser(path):
428 """Expand ~ and ~user constructs.
430 If user or $HOME is unknown, do nothing."""
434 while i < n and path[i] not in '/\\':
437 if 'HOME' in os.environ:
438 userhome = compat_getenv('HOME')
439 elif 'USERPROFILE' in os.environ:
440 userhome = compat_getenv('USERPROFILE')
441 elif 'HOMEPATH' not in os.environ:
445 drive = compat_getenv('HOMEDRIVE')
448 userhome = os.path.join(drive, compat_getenv('HOMEPATH'))
451 userhome = os.path.join(os.path.dirname(userhome), path[1:i])
453 return userhome + path[i:]
455 compat_expanduser = os.path.expanduser
458 if sys.version_info < (3, 0):
460 from .utils import preferredencoding
461 print(s.encode(preferredencoding(), 'xmlcharrefreplace'))
464 assert isinstance(s, compat_str)
469 subprocess_check_output = subprocess.check_output
470 except AttributeError:
471 def subprocess_check_output(*args, **kwargs):
472 assert 'input' not in kwargs
473 p = subprocess.Popen(*args, stdout=subprocess.PIPE, **kwargs)
474 output, _ = p.communicate()
477 raise subprocess.CalledProcessError(ret, p.args, output=output)
480 if sys.version_info < (3, 0) and sys.platform == 'win32':
481 def compat_getpass(prompt, *args, **kwargs):
482 if isinstance(prompt, compat_str):
483 from .utils import preferredencoding
484 prompt = prompt.encode(preferredencoding())
485 return getpass.getpass(prompt, *args, **kwargs)
487 compat_getpass = getpass.getpass
489 # Python < 2.6.5 require kwargs to be bytes
493 _testfunc(**{'x': 0})
495 def compat_kwargs(kwargs):
496 return dict((bytes(k), v) for k, v in kwargs.items())
498 compat_kwargs = lambda kwargs: kwargs
501 if sys.version_info < (2, 7):
502 def compat_socket_create_connection(address, timeout, source_address=None):
505 for res in socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM):
506 af, socktype, proto, canonname, sa = res
509 sock = socket.socket(af, socktype, proto)
510 sock.settimeout(timeout)
512 sock.bind(source_address)
515 except socket.error as _:
522 raise socket.error('getaddrinfo returns an empty list')
524 compat_socket_create_connection = socket.create_connection
527 # Fix https://github.com/rg3/youtube-dl/issues/4223
528 # See http://bugs.python.org/issue9161 for what is broken
529 def workaround_optparse_bug9161():
530 op = optparse.OptionParser()
531 og = optparse.OptionGroup(op, 'foo')
535 real_add_option = optparse.OptionGroup.add_option
537 def _compat_add_option(self, *args, **kwargs):
539 v.encode('ascii', 'replace') if isinstance(v, compat_str)
541 bargs = [enc(a) for a in args]
543 (k, enc(v)) for k, v in kwargs.items())
544 return real_add_option(self, *bargs, **bkwargs)
545 optparse.OptionGroup.add_option = _compat_add_option
547 if hasattr(shutil, 'get_terminal_size'): # Python >= 3.3
548 compat_get_terminal_size = shutil.get_terminal_size
550 _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines'])
552 def compat_get_terminal_size(fallback=(80, 24)):
553 columns = compat_getenv('COLUMNS')
555 columns = int(columns)
558 lines = compat_getenv('LINES')
564 if columns is None or lines is None or columns <= 0 or lines <= 0:
566 sp = subprocess.Popen(
568 stdout=subprocess.PIPE, stderr=subprocess.PIPE)
569 out, err = sp.communicate()
570 _lines, _columns = map(int, out.split())
572 _columns, _lines = _terminal_size(*fallback)
574 if columns is None or columns <= 0:
576 if lines is None or lines <= 0:
578 return _terminal_size(columns, lines)
581 itertools.count(start=0, step=1)
582 compat_itertools_count = itertools.count
583 except TypeError: # Python 2.6
584 def compat_itertools_count(start=0, step=1):
590 if sys.version_info >= (3, 0):
591 from tokenize import tokenize as compat_tokenize_tokenize
593 from tokenize import generate_tokens as compat_tokenize_tokenize
602 'compat_etree_fromstring',
604 'compat_get_terminal_size',
607 'compat_html_entities',
608 'compat_http_client',
609 'compat_http_server',
610 'compat_itertools_count',
617 'compat_shlex_split',
618 'compat_socket_create_connection',
620 'compat_subprocess_get_DEVNULL',
621 'compat_tokenize_tokenize',
622 'compat_urllib_error',
623 'compat_urllib_parse',
624 'compat_urllib_parse_unquote',
625 'compat_urllib_parse_unquote_plus',
626 'compat_urllib_parse_unquote_to_bytes',
627 'compat_urllib_parse_urlencode',
628 'compat_urllib_parse_urlparse',
629 'compat_urllib_request',
630 'compat_urllib_request_DataHandler',
631 'compat_urllib_response',
633 'compat_urlretrieve',
634 'compat_xml_parse_error',
637 'subprocess_check_output',
638 'workaround_optparse_bug9161',