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
248 if sys.version_info[0] >= 3:
249 compat_etree_fromstring = xml.etree.ElementTree.fromstring
251 # python 2.x tries to encode unicode strings with ascii (see the
252 # XMLParser._fixtext method)
253 etree = xml.etree.ElementTree
256 _etree_iter = etree.Element.iter
257 except AttributeError: # Python <=2.6
258 def _etree_iter(root):
259 for el in root.findall('*'):
261 for sub in _etree_iter(el):
264 # on 2.6 XML doesn't have a parser argument, function copied from CPython
266 def _XML(text, parser=None):
268 parser = etree.XMLParser(target=etree.TreeBuilder())
270 return parser.close()
272 def _element_factory(*args, **kwargs):
273 el = etree.Element(*args, **kwargs)
274 for k, v in el.items():
275 if isinstance(v, bytes):
276 el.set(k, v.decode('utf-8'))
279 def compat_etree_fromstring(text):
280 doc = _XML(text, parser=etree.XMLParser(target=etree.TreeBuilder(element_factory=_element_factory)))
281 for el in _etree_iter(doc):
282 if el.text is not None and isinstance(el.text, bytes):
283 el.text = el.text.decode('utf-8')
286 if sys.version_info < (2, 7):
287 # Here comes the crazy part: In 2.6, if the xpath is a unicode,
288 # .//node does not match if a node is a direct child of . !
289 def compat_xpath(xpath):
290 if isinstance(xpath, compat_str):
291 xpath = xpath.encode('ascii')
294 compat_xpath = lambda xpath: xpath
297 from urllib.parse import parse_qs as compat_parse_qs
298 except ImportError: # Python 2
299 # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib.
300 # Python 2's version is apparently totally broken
302 def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
303 encoding='utf-8', errors='replace'):
304 qs, _coerce_result = qs, compat_str
305 pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')]
307 for name_value in pairs:
308 if not name_value and not strict_parsing:
310 nv = name_value.split('=', 1)
313 raise ValueError('bad query field: %r' % (name_value,))
314 # Handle case of a control-name with no equal sign
315 if keep_blank_values:
319 if len(nv[1]) or keep_blank_values:
320 name = nv[0].replace('+', ' ')
321 name = compat_urllib_parse_unquote(
322 name, encoding=encoding, errors=errors)
323 name = _coerce_result(name)
324 value = nv[1].replace('+', ' ')
325 value = compat_urllib_parse_unquote(
326 value, encoding=encoding, errors=errors)
327 value = _coerce_result(value)
328 r.append((name, value))
331 def compat_parse_qs(qs, keep_blank_values=False, strict_parsing=False,
332 encoding='utf-8', errors='replace'):
334 pairs = _parse_qsl(qs, keep_blank_values, strict_parsing,
335 encoding=encoding, errors=errors)
336 for name, value in pairs:
337 if name in parsed_result:
338 parsed_result[name].append(value)
340 parsed_result[name] = [value]
344 from shlex import quote as compat_shlex_quote
345 except ImportError: # Python < 3.3
346 def compat_shlex_quote(s):
347 if re.match(r'^[-_\w./]+$', s):
350 return "'" + s.replace("'", "'\"'\"'") + "'"
353 if sys.version_info >= (2, 7, 3):
354 compat_shlex_split = shlex.split
356 # Working around shlex issue with unicode strings on some python 2
357 # versions (see http://bugs.python.org/issue1548891)
358 def compat_shlex_split(s, comments=False, posix=True):
359 if isinstance(s, compat_str):
360 s = s.encode('utf-8')
361 return shlex.split(s, comments, posix)
371 compat_os_name = os._name if os.name == 'java' else os.name
374 if sys.version_info >= (3, 0):
375 compat_getenv = os.getenv
376 compat_expanduser = os.path.expanduser
378 def compat_setenv(key, value, env=os.environ):
381 # Environment variables should be decoded with filesystem encoding.
382 # Otherwise it will fail if any non-ASCII characters present (see #3854 #3217 #2918)
384 def compat_getenv(key, default=None):
385 from .utils import get_filesystem_encoding
386 env = os.getenv(key, default)
388 env = env.decode(get_filesystem_encoding())
391 def compat_setenv(key, value, env=os.environ):
393 from .utils import get_filesystem_encoding
394 return v.encode(get_filesystem_encoding()) if isinstance(v, compat_str) else v
395 env[encode(key)] = encode(value)
397 # HACK: The default implementations of os.path.expanduser from cpython do not decode
398 # environment variables with filesystem encoding. We will work around this by
399 # providing adjusted implementations.
400 # The following are os.path.expanduser implementations from cpython 2.7.8 stdlib
401 # for different platforms with correct environment variables decoding.
403 if compat_os_name == 'posix':
404 def compat_expanduser(path):
405 """Expand ~ and ~user constructions. If user or $HOME is unknown,
407 if not path.startswith('~'):
409 i = path.find('/', 1)
413 if 'HOME' not in os.environ:
415 userhome = pwd.getpwuid(os.getuid()).pw_dir
417 userhome = compat_getenv('HOME')
421 pwent = pwd.getpwnam(path[1:i])
424 userhome = pwent.pw_dir
425 userhome = userhome.rstrip('/')
426 return (userhome + path[i:]) or '/'
427 elif compat_os_name == 'nt' or compat_os_name == 'ce':
428 def compat_expanduser(path):
429 """Expand ~ and ~user constructs.
431 If user or $HOME is unknown, do nothing."""
435 while i < n and path[i] not in '/\\':
438 if 'HOME' in os.environ:
439 userhome = compat_getenv('HOME')
440 elif 'USERPROFILE' in os.environ:
441 userhome = compat_getenv('USERPROFILE')
442 elif 'HOMEPATH' not in os.environ:
446 drive = compat_getenv('HOMEDRIVE')
449 userhome = os.path.join(drive, compat_getenv('HOMEPATH'))
452 userhome = os.path.join(os.path.dirname(userhome), path[1:i])
454 return userhome + path[i:]
456 compat_expanduser = os.path.expanduser
459 if sys.version_info < (3, 0):
461 from .utils import preferredencoding
462 print(s.encode(preferredencoding(), 'xmlcharrefreplace'))
465 assert isinstance(s, compat_str)
469 if sys.version_info < (3, 0) and sys.platform == 'win32':
470 def compat_getpass(prompt, *args, **kwargs):
471 if isinstance(prompt, compat_str):
472 from .utils import preferredencoding
473 prompt = prompt.encode(preferredencoding())
474 return getpass.getpass(prompt, *args, **kwargs)
476 compat_getpass = getpass.getpass
478 # Python < 2.6.5 require kwargs to be bytes
482 _testfunc(**{'x': 0})
484 def compat_kwargs(kwargs):
485 return dict((bytes(k), v) for k, v in kwargs.items())
487 compat_kwargs = lambda kwargs: kwargs
490 if sys.version_info < (2, 7):
491 def compat_socket_create_connection(address, timeout, source_address=None):
494 for res in socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM):
495 af, socktype, proto, canonname, sa = res
498 sock = socket.socket(af, socktype, proto)
499 sock.settimeout(timeout)
501 sock.bind(source_address)
504 except socket.error as _:
511 raise socket.error('getaddrinfo returns an empty list')
513 compat_socket_create_connection = socket.create_connection
516 # Fix https://github.com/rg3/youtube-dl/issues/4223
517 # See http://bugs.python.org/issue9161 for what is broken
518 def workaround_optparse_bug9161():
519 op = optparse.OptionParser()
520 og = optparse.OptionGroup(op, 'foo')
524 real_add_option = optparse.OptionGroup.add_option
526 def _compat_add_option(self, *args, **kwargs):
528 v.encode('ascii', 'replace') if isinstance(v, compat_str)
530 bargs = [enc(a) for a in args]
532 (k, enc(v)) for k, v in kwargs.items())
533 return real_add_option(self, *bargs, **bkwargs)
534 optparse.OptionGroup.add_option = _compat_add_option
536 if hasattr(shutil, 'get_terminal_size'): # Python >= 3.3
537 compat_get_terminal_size = shutil.get_terminal_size
539 _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines'])
541 def compat_get_terminal_size(fallback=(80, 24)):
542 columns = compat_getenv('COLUMNS')
544 columns = int(columns)
547 lines = compat_getenv('LINES')
553 if columns is None or lines is None or columns <= 0 or lines <= 0:
555 sp = subprocess.Popen(
557 stdout=subprocess.PIPE, stderr=subprocess.PIPE)
558 out, err = sp.communicate()
559 _lines, _columns = map(int, out.split())
561 _columns, _lines = _terminal_size(*fallback)
563 if columns is None or columns <= 0:
565 if lines is None or lines <= 0:
567 return _terminal_size(columns, lines)
570 itertools.count(start=0, step=1)
571 compat_itertools_count = itertools.count
572 except TypeError: # Python 2.6
573 def compat_itertools_count(start=0, step=1):
579 if sys.version_info >= (3, 0):
580 from tokenize import tokenize as compat_tokenize_tokenize
582 from tokenize import generate_tokens as compat_tokenize_tokenize
588 # In Python 2.6 and 2.7.x < 2.7.7, struct requires a bytes argument
589 # See https://bugs.python.org/issue19099
590 def compat_struct_pack(spec, *args):
591 if isinstance(spec, compat_str):
592 spec = spec.encode('ascii')
593 return struct.pack(spec, *args)
595 def compat_struct_unpack(spec, *args):
596 if isinstance(spec, compat_str):
597 spec = spec.encode('ascii')
598 return struct.unpack(spec, *args)
600 compat_struct_pack = struct.pack
601 compat_struct_unpack = struct.unpack
611 'compat_etree_fromstring',
613 'compat_get_terminal_size',
616 'compat_html_entities',
617 'compat_http_client',
618 'compat_http_server',
619 'compat_itertools_count',
626 'compat_shlex_quote',
627 'compat_shlex_split',
628 'compat_socket_create_connection',
630 'compat_struct_pack',
631 'compat_struct_unpack',
632 'compat_subprocess_get_DEVNULL',
633 'compat_tokenize_tokenize',
634 'compat_urllib_error',
635 'compat_urllib_parse',
636 'compat_urllib_parse_unquote',
637 'compat_urllib_parse_unquote_plus',
638 'compat_urllib_parse_unquote_to_bytes',
639 'compat_urllib_parse_urlencode',
640 'compat_urllib_parse_urlparse',
641 'compat_urllib_request',
642 'compat_urllib_request_DataHandler',
643 'compat_urllib_response',
645 'compat_urlretrieve',
646 'compat_xml_parse_error',
648 'workaround_optparse_bug9161',