1 from __future__ import unicode_literals
18 import xml.etree.ElementTree
22 import urllib.request as compat_urllib_request
23 except ImportError: # Python 2
24 import urllib2 as compat_urllib_request
27 import urllib.error as compat_urllib_error
28 except ImportError: # Python 2
29 import urllib2 as compat_urllib_error
32 import urllib.parse as compat_urllib_parse
33 except ImportError: # Python 2
34 import urllib as compat_urllib_parse
37 from urllib.parse import urlparse as compat_urllib_parse_urlparse
38 except ImportError: # Python 2
39 from urlparse import urlparse as compat_urllib_parse_urlparse
42 import urllib.parse as compat_urlparse
43 except ImportError: # Python 2
44 import urlparse as compat_urlparse
47 import urllib.response as compat_urllib_response
48 except ImportError: # Python 2
49 import urllib as compat_urllib_response
52 import http.cookiejar as compat_cookiejar
53 except ImportError: # Python 2
54 import cookielib as compat_cookiejar
57 import http.cookies as compat_cookies
58 except ImportError: # Python 2
59 import Cookie as compat_cookies
62 import html.entities as compat_html_entities
63 except ImportError: # Python 2
64 import htmlentitydefs as compat_html_entities
67 import http.client as compat_http_client
68 except ImportError: # Python 2
69 import httplib as compat_http_client
72 from urllib.error import HTTPError as compat_HTTPError
73 except ImportError: # Python 2
74 from urllib2 import HTTPError as compat_HTTPError
77 from urllib.request import urlretrieve as compat_urlretrieve
78 except ImportError: # Python 2
79 from urllib import urlretrieve as compat_urlretrieve
82 from html.parser import HTMLParser as compat_HTMLParser
83 except ImportError: # Python 2
84 from HTMLParser import HTMLParser as compat_HTMLParser
88 from subprocess import DEVNULL
89 compat_subprocess_get_DEVNULL = lambda: DEVNULL
91 compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w')
94 import http.server as compat_http_server
96 import BaseHTTPServer as compat_http_server
99 compat_str = unicode # Python 2
104 from urllib.parse import unquote_to_bytes as compat_urllib_parse_unquote_to_bytes
105 from urllib.parse import unquote as compat_urllib_parse_unquote
106 from urllib.parse import unquote_plus as compat_urllib_parse_unquote_plus
107 except ImportError: # Python 2
108 _asciire = (compat_urllib_parse._asciire if hasattr(compat_urllib_parse, '_asciire')
109 else re.compile('([\x00-\x7f]+)'))
111 # HACK: The following are the correct unquote_to_bytes, unquote and unquote_plus
112 # implementations from cpython 3.4.3's stdlib. Python 2's version
113 # is apparently broken (see https://github.com/rg3/youtube-dl/pull/6244)
115 def compat_urllib_parse_unquote_to_bytes(string):
116 """unquote_to_bytes('abc%20def') -> b'abc def'."""
117 # Note: strings are encoded as UTF-8. This is only an issue if it contains
118 # unescaped non-ASCII characters, which URIs should not.
120 # Is it a string-like object?
123 if isinstance(string, compat_str):
124 string = string.encode('utf-8')
125 bits = string.split(b'%')
130 for item in bits[1:]:
132 append(compat_urllib_parse._hextochr[item[:2]])
139 def compat_urllib_parse_unquote(string, encoding='utf-8', errors='replace'):
140 """Replace %xx escapes by their single-character equivalent. The optional
141 encoding and errors parameters specify how to decode percent-encoded
142 sequences into Unicode characters, as accepted by the bytes.decode()
144 By default, percent-encoded sequences are decoded with UTF-8, and invalid
145 sequences are replaced by a placeholder character.
147 unquote('abc%20def') -> 'abc def'.
149 if '%' not in string:
156 bits = _asciire.split(string)
159 for i in range(1, len(bits), 2):
160 append(compat_urllib_parse_unquote_to_bytes(bits[i]).decode(encoding, errors))
164 def compat_urllib_parse_unquote_plus(string, encoding='utf-8', errors='replace'):
165 """Like unquote(), but also replace plus signs by spaces, as required for
166 unquoting HTML form values.
168 unquote_plus('%7e/abc+def') -> '~/abc def'
170 string = string.replace('+', ' ')
171 return compat_urllib_parse_unquote(string, encoding, errors)
174 from urllib.parse import urlencode as compat_urllib_parse_urlencode
175 except ImportError: # Python 2
176 # Python 2 will choke in urlencode on mixture of byte and unicode strings.
177 # Possible solutions are to either port it from python 3 with all
178 # the friends or manually ensure input query contains only byte strings.
179 # We will stick with latter thus recursively encoding the whole query.
180 def compat_urllib_parse_urlencode(query, doseq=0, encoding='utf-8'):
182 if isinstance(e, dict):
184 elif isinstance(e, (list, tuple,)):
185 list_e = encode_list(e)
186 e = tuple(list_e) if isinstance(e, tuple) else list_e
187 elif isinstance(e, compat_str):
188 e = e.encode(encoding)
192 return dict((encode_elem(k), encode_elem(v)) for k, v in d.items())
195 return [encode_elem(e) for e in l]
197 return compat_urllib_parse.urlencode(encode_elem(query), doseq=doseq)
200 from urllib.request import DataHandler as compat_urllib_request_DataHandler
201 except ImportError: # Python < 3.4
202 # Ported from CPython 98774:1733b3bd46db, Lib/urllib/request.py
203 class compat_urllib_request_DataHandler(compat_urllib_request.BaseHandler):
204 def data_open(self, req):
205 # data URLs as specified in RFC 2397.
207 # ignores POSTed data
210 # dataurl := "data:" [ mediatype ] [ ";base64" ] "," data
211 # mediatype := [ type "/" subtype ] *( ";" parameter )
213 # parameter := attribute "=" value
214 url = req.get_full_url()
216 scheme, data = url.split(':', 1)
217 mediatype, data = data.split(',', 1)
219 # even base64 encoded data URLs might be quoted so unquote in any case:
220 data = compat_urllib_parse_unquote_to_bytes(data)
221 if mediatype.endswith(';base64'):
222 data = binascii.a2b_base64(data)
223 mediatype = mediatype[:-7]
226 mediatype = 'text/plain;charset=US-ASCII'
228 headers = email.message_from_string(
229 'Content-type: %s\nContent-length: %d\n' % (mediatype, len(data)))
231 return compat_urllib_response.addinfourl(io.BytesIO(data), headers, url)
234 compat_basestring = basestring # Python 2
236 compat_basestring = str
239 compat_chr = unichr # Python 2
244 from xml.etree.ElementTree import ParseError as compat_xml_parse_error
245 except ImportError: # Python 2.6
246 from xml.parsers.expat import ExpatError as compat_xml_parse_error
249 etree = xml.etree.ElementTree
252 class _TreeBuilder(etree.TreeBuilder):
253 def doctype(self, name, pubid, system):
256 if sys.version_info[0] >= 3:
257 def compat_etree_fromstring(text):
258 return etree.XML(text, parser=etree.XMLParser(target=_TreeBuilder()))
260 # python 2.x tries to encode unicode strings with ascii (see the
261 # XMLParser._fixtext method)
263 _etree_iter = etree.Element.iter
264 except AttributeError: # Python <=2.6
265 def _etree_iter(root):
266 for el in root.findall('*'):
268 for sub in _etree_iter(el):
271 # on 2.6 XML doesn't have a parser argument, function copied from CPython
273 def _XML(text, parser=None):
275 parser = etree.XMLParser(target=_TreeBuilder())
277 return parser.close()
279 def _element_factory(*args, **kwargs):
280 el = etree.Element(*args, **kwargs)
281 for k, v in el.items():
282 if isinstance(v, bytes):
283 el.set(k, v.decode('utf-8'))
286 def compat_etree_fromstring(text):
287 doc = _XML(text, parser=etree.XMLParser(target=_TreeBuilder(element_factory=_element_factory)))
288 for el in _etree_iter(doc):
289 if el.text is not None and isinstance(el.text, bytes):
290 el.text = el.text.decode('utf-8')
293 if sys.version_info < (2, 7):
294 # Here comes the crazy part: In 2.6, if the xpath is a unicode,
295 # .//node does not match if a node is a direct child of . !
296 def compat_xpath(xpath):
297 if isinstance(xpath, compat_str):
298 xpath = xpath.encode('ascii')
301 compat_xpath = lambda xpath: xpath
304 from urllib.parse import parse_qs as compat_parse_qs
305 except ImportError: # Python 2
306 # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib.
307 # Python 2's version is apparently totally broken
309 def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
310 encoding='utf-8', errors='replace'):
311 qs, _coerce_result = qs, compat_str
312 pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')]
314 for name_value in pairs:
315 if not name_value and not strict_parsing:
317 nv = name_value.split('=', 1)
320 raise ValueError('bad query field: %r' % (name_value,))
321 # Handle case of a control-name with no equal sign
322 if keep_blank_values:
326 if len(nv[1]) or keep_blank_values:
327 name = nv[0].replace('+', ' ')
328 name = compat_urllib_parse_unquote(
329 name, encoding=encoding, errors=errors)
330 name = _coerce_result(name)
331 value = nv[1].replace('+', ' ')
332 value = compat_urllib_parse_unquote(
333 value, encoding=encoding, errors=errors)
334 value = _coerce_result(value)
335 r.append((name, value))
338 def compat_parse_qs(qs, keep_blank_values=False, strict_parsing=False,
339 encoding='utf-8', errors='replace'):
341 pairs = _parse_qsl(qs, keep_blank_values, strict_parsing,
342 encoding=encoding, errors=errors)
343 for name, value in pairs:
344 if name in parsed_result:
345 parsed_result[name].append(value)
347 parsed_result[name] = [value]
351 from shlex import quote as compat_shlex_quote
352 except ImportError: # Python < 3.3
353 def compat_shlex_quote(s):
354 if re.match(r'^[-_\w./]+$', s):
357 return "'" + s.replace("'", "'\"'\"'") + "'"
360 if sys.version_info >= (2, 7, 3):
361 compat_shlex_split = shlex.split
363 # Working around shlex issue with unicode strings on some python 2
364 # versions (see http://bugs.python.org/issue1548891)
365 def compat_shlex_split(s, comments=False, posix=True):
366 if isinstance(s, compat_str):
367 s = s.encode('utf-8')
368 return shlex.split(s, comments, posix)
378 compat_os_name = os._name if os.name == 'java' else os.name
381 if sys.version_info >= (3, 0):
382 compat_getenv = os.getenv
383 compat_expanduser = os.path.expanduser
385 def compat_setenv(key, value, env=os.environ):
388 # Environment variables should be decoded with filesystem encoding.
389 # Otherwise it will fail if any non-ASCII characters present (see #3854 #3217 #2918)
391 def compat_getenv(key, default=None):
392 from .utils import get_filesystem_encoding
393 env = os.getenv(key, default)
395 env = env.decode(get_filesystem_encoding())
398 def compat_setenv(key, value, env=os.environ):
400 from .utils import get_filesystem_encoding
401 return v.encode(get_filesystem_encoding()) if isinstance(v, compat_str) else v
402 env[encode(key)] = encode(value)
404 # HACK: The default implementations of os.path.expanduser from cpython do not decode
405 # environment variables with filesystem encoding. We will work around this by
406 # providing adjusted implementations.
407 # The following are os.path.expanduser implementations from cpython 2.7.8 stdlib
408 # for different platforms with correct environment variables decoding.
410 if compat_os_name == 'posix':
411 def compat_expanduser(path):
412 """Expand ~ and ~user constructions. If user or $HOME is unknown,
414 if not path.startswith('~'):
416 i = path.find('/', 1)
420 if 'HOME' not in os.environ:
422 userhome = pwd.getpwuid(os.getuid()).pw_dir
424 userhome = compat_getenv('HOME')
428 pwent = pwd.getpwnam(path[1:i])
431 userhome = pwent.pw_dir
432 userhome = userhome.rstrip('/')
433 return (userhome + path[i:]) or '/'
434 elif compat_os_name == 'nt' or compat_os_name == 'ce':
435 def compat_expanduser(path):
436 """Expand ~ and ~user constructs.
438 If user or $HOME is unknown, do nothing."""
442 while i < n and path[i] not in '/\\':
445 if 'HOME' in os.environ:
446 userhome = compat_getenv('HOME')
447 elif 'USERPROFILE' in os.environ:
448 userhome = compat_getenv('USERPROFILE')
449 elif 'HOMEPATH' not in os.environ:
453 drive = compat_getenv('HOMEDRIVE')
456 userhome = os.path.join(drive, compat_getenv('HOMEPATH'))
459 userhome = os.path.join(os.path.dirname(userhome), path[1:i])
461 return userhome + path[i:]
463 compat_expanduser = os.path.expanduser
466 if sys.version_info < (3, 0):
468 from .utils import preferredencoding
469 print(s.encode(preferredencoding(), 'xmlcharrefreplace'))
472 assert isinstance(s, compat_str)
476 if sys.version_info < (3, 0) and sys.platform == 'win32':
477 def compat_getpass(prompt, *args, **kwargs):
478 if isinstance(prompt, compat_str):
479 from .utils import preferredencoding
480 prompt = prompt.encode(preferredencoding())
481 return getpass.getpass(prompt, *args, **kwargs)
483 compat_getpass = getpass.getpass
485 # Python < 2.6.5 require kwargs to be bytes
489 _testfunc(**{'x': 0})
491 def compat_kwargs(kwargs):
492 return dict((bytes(k), v) for k, v in kwargs.items())
494 compat_kwargs = lambda kwargs: kwargs
497 if sys.version_info < (2, 7):
498 def compat_socket_create_connection(address, timeout, source_address=None):
501 for res in socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM):
502 af, socktype, proto, canonname, sa = res
505 sock = socket.socket(af, socktype, proto)
506 sock.settimeout(timeout)
508 sock.bind(source_address)
511 except socket.error as _:
518 raise socket.error('getaddrinfo returns an empty list')
520 compat_socket_create_connection = socket.create_connection
523 # Fix https://github.com/rg3/youtube-dl/issues/4223
524 # See http://bugs.python.org/issue9161 for what is broken
525 def workaround_optparse_bug9161():
526 op = optparse.OptionParser()
527 og = optparse.OptionGroup(op, 'foo')
531 real_add_option = optparse.OptionGroup.add_option
533 def _compat_add_option(self, *args, **kwargs):
535 v.encode('ascii', 'replace') if isinstance(v, compat_str)
537 bargs = [enc(a) for a in args]
539 (k, enc(v)) for k, v in kwargs.items())
540 return real_add_option(self, *bargs, **bkwargs)
541 optparse.OptionGroup.add_option = _compat_add_option
543 if hasattr(shutil, 'get_terminal_size'): # Python >= 3.3
544 compat_get_terminal_size = shutil.get_terminal_size
546 _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines'])
548 def compat_get_terminal_size(fallback=(80, 24)):
549 columns = compat_getenv('COLUMNS')
551 columns = int(columns)
554 lines = compat_getenv('LINES')
560 if columns is None or lines is None or columns <= 0 or lines <= 0:
562 sp = subprocess.Popen(
564 stdout=subprocess.PIPE, stderr=subprocess.PIPE)
565 out, err = sp.communicate()
566 _lines, _columns = map(int, out.split())
568 _columns, _lines = _terminal_size(*fallback)
570 if columns is None or columns <= 0:
572 if lines is None or lines <= 0:
574 return _terminal_size(columns, lines)
577 itertools.count(start=0, step=1)
578 compat_itertools_count = itertools.count
579 except TypeError: # Python 2.6
580 def compat_itertools_count(start=0, step=1):
586 if sys.version_info >= (3, 0):
587 from tokenize import tokenize as compat_tokenize_tokenize
589 from tokenize import generate_tokens as compat_tokenize_tokenize
595 # In Python 2.6 and 2.7.x < 2.7.7, struct requires a bytes argument
596 # See https://bugs.python.org/issue19099
597 def compat_struct_pack(spec, *args):
598 if isinstance(spec, compat_str):
599 spec = spec.encode('ascii')
600 return struct.pack(spec, *args)
602 def compat_struct_unpack(spec, *args):
603 if isinstance(spec, compat_str):
604 spec = spec.encode('ascii')
605 return struct.unpack(spec, *args)
607 compat_struct_pack = struct.pack
608 compat_struct_unpack = struct.unpack
618 'compat_etree_fromstring',
620 'compat_get_terminal_size',
623 'compat_html_entities',
624 'compat_http_client',
625 'compat_http_server',
626 'compat_itertools_count',
633 'compat_shlex_quote',
634 'compat_shlex_split',
635 'compat_socket_create_connection',
637 'compat_struct_pack',
638 'compat_struct_unpack',
639 'compat_subprocess_get_DEVNULL',
640 'compat_tokenize_tokenize',
641 'compat_urllib_error',
642 'compat_urllib_parse',
643 'compat_urllib_parse_unquote',
644 'compat_urllib_parse_unquote_plus',
645 'compat_urllib_parse_unquote_to_bytes',
646 'compat_urllib_parse_urlencode',
647 'compat_urllib_parse_urlparse',
648 'compat_urllib_request',
649 'compat_urllib_request_DataHandler',
650 'compat_urllib_response',
652 'compat_urlretrieve',
653 'compat_xml_parse_error',
655 'workaround_optparse_bug9161',