1 from __future__ import unicode_literals
12 import urllib.request as compat_urllib_request
13 except ImportError: # Python 2
14 import urllib2 as compat_urllib_request
17 import urllib.error as compat_urllib_error
18 except ImportError: # Python 2
19 import urllib2 as compat_urllib_error
22 import urllib.parse as compat_urllib_parse
23 except ImportError: # Python 2
24 import urllib as compat_urllib_parse
27 from urllib.parse import urlparse as compat_urllib_parse_urlparse
28 except ImportError: # Python 2
29 from urlparse import urlparse as compat_urllib_parse_urlparse
32 import urllib.parse as compat_urlparse
33 except ImportError: # Python 2
34 import urlparse as compat_urlparse
37 import http.cookiejar as compat_cookiejar
38 except ImportError: # Python 2
39 import cookielib as compat_cookiejar
42 import html.entities as compat_html_entities
43 except ImportError: # Python 2
44 import htmlentitydefs as compat_html_entities
47 import html.parser as compat_html_parser
48 except ImportError: # Python 2
49 import HTMLParser as compat_html_parser
52 import http.client as compat_http_client
53 except ImportError: # Python 2
54 import httplib as compat_http_client
57 from urllib.error import HTTPError as compat_HTTPError
58 except ImportError: # Python 2
59 from urllib2 import HTTPError as compat_HTTPError
62 from urllib.request import urlretrieve as compat_urlretrieve
63 except ImportError: # Python 2
64 from urllib import urlretrieve as compat_urlretrieve
68 from subprocess import DEVNULL
69 compat_subprocess_get_DEVNULL = lambda: DEVNULL
71 compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w')
74 from urllib.parse import unquote as compat_urllib_parse_unquote
76 def compat_urllib_parse_unquote(string, encoding='utf-8', errors='replace'):
79 res = string.split('%')
86 # pct_sequence: contiguous sequence of percent-encoded bytes, decoded
93 pct_sequence += item[:2].decode('hex')
96 # This segment was just a single percent-encoded character.
97 # May be part of a sequence of code units, so delay decoding.
98 # (Stored in pct_sequence).
102 # Encountered non-percent-encoded characters. Flush the current
104 string += pct_sequence.decode(encoding, errors) + rest
107 # Flush the final pct_sequence
108 string += pct_sequence.decode(encoding, errors)
113 from urllib.parse import parse_qs as compat_parse_qs
114 except ImportError: # Python 2
115 # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib.
116 # Python 2's version is apparently totally broken
118 def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
119 encoding='utf-8', errors='replace'):
120 qs, _coerce_result = qs, unicode
121 pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')]
123 for name_value in pairs:
124 if not name_value and not strict_parsing:
126 nv = name_value.split('=', 1)
129 raise ValueError("bad query field: %r" % (name_value,))
130 # Handle case of a control-name with no equal sign
131 if keep_blank_values:
135 if len(nv[1]) or keep_blank_values:
136 name = nv[0].replace('+', ' ')
137 name = compat_urllib_parse_unquote(
138 name, encoding=encoding, errors=errors)
139 name = _coerce_result(name)
140 value = nv[1].replace('+', ' ')
141 value = compat_urllib_parse_unquote(
142 value, encoding=encoding, errors=errors)
143 value = _coerce_result(value)
144 r.append((name, value))
147 def compat_parse_qs(qs, keep_blank_values=False, strict_parsing=False,
148 encoding='utf-8', errors='replace'):
150 pairs = _parse_qsl(qs, keep_blank_values, strict_parsing,
151 encoding=encoding, errors=errors)
152 for name, value in pairs:
153 if name in parsed_result:
154 parsed_result[name].append(value)
156 parsed_result[name] = [value]
160 compat_str = unicode # Python 2
165 compat_chr = unichr # Python 2
170 from xml.etree.ElementTree import ParseError as compat_xml_parse_error
171 except ImportError: # Python 2.6
172 from xml.parsers.expat import ExpatError as compat_xml_parse_error
175 from shlex import quote as shlex_quote
176 except ImportError: # Python < 3.3
178 if re.match(r'^[-_\w./]+$', s):
181 return "'" + s.replace("'", "'\"'\"'") + "'"
191 if sys.version_info >= (3, 0):
192 compat_getenv = os.getenv
193 compat_expanduser = os.path.expanduser
195 # Environment variables should be decoded with filesystem encoding.
196 # Otherwise it will fail if any non-ASCII characters present (see #3854 #3217 #2918)
198 def compat_getenv(key, default=None):
199 from .utils import get_filesystem_encoding
200 env = os.getenv(key, default)
202 env = env.decode(get_filesystem_encoding())
205 # HACK: The default implementations of os.path.expanduser from cpython do not decode
206 # environment variables with filesystem encoding. We will work around this by
207 # providing adjusted implementations.
208 # The following are os.path.expanduser implementations from cpython 2.7.8 stdlib
209 # for different platforms with correct environment variables decoding.
211 if os.name == 'posix':
212 def compat_expanduser(path):
213 """Expand ~ and ~user constructions. If user or $HOME is unknown,
215 if not path.startswith('~'):
217 i = path.find('/', 1)
221 if 'HOME' not in os.environ:
223 userhome = pwd.getpwuid(os.getuid()).pw_dir
225 userhome = compat_getenv('HOME')
229 pwent = pwd.getpwnam(path[1:i])
232 userhome = pwent.pw_dir
233 userhome = userhome.rstrip('/')
234 return (userhome + path[i:]) or '/'
235 elif os.name == 'nt' or os.name == 'ce':
236 def compat_expanduser(path):
237 """Expand ~ and ~user constructs.
239 If user or $HOME is unknown, do nothing."""
243 while i < n and path[i] not in '/\\':
246 if 'HOME' in os.environ:
247 userhome = compat_getenv('HOME')
248 elif 'USERPROFILE' in os.environ:
249 userhome = compat_getenv('USERPROFILE')
250 elif not 'HOMEPATH' in os.environ:
254 drive = compat_getenv('HOMEDRIVE')
257 userhome = os.path.join(drive, compat_getenv('HOMEPATH'))
260 userhome = os.path.join(os.path.dirname(userhome), path[1:i])
262 return userhome + path[i:]
264 compat_expanduser = os.path.expanduser
267 if sys.version_info < (3, 0):
269 from .utils import preferredencoding
270 print(s.encode(preferredencoding(), 'xmlcharrefreplace'))
273 assert isinstance(s, compat_str)
278 subprocess_check_output = subprocess.check_output
279 except AttributeError:
280 def subprocess_check_output(*args, **kwargs):
281 assert 'input' not in kwargs
282 p = subprocess.Popen(*args, stdout=subprocess.PIPE, **kwargs)
283 output, _ = p.communicate()
286 raise subprocess.CalledProcessError(ret, p.args, output=output)
289 if sys.version_info < (3, 0) and sys.platform == 'win32':
290 def compat_getpass(prompt, *args, **kwargs):
291 if isinstance(prompt, compat_str):
292 from .utils import preferredencoding
293 prompt = prompt.encode(preferredencoding())
294 return getpass.getpass(prompt, *args, **kwargs)
296 compat_getpass = getpass.getpass
298 # Old 2.6 and 2.7 releases require kwargs to be bytes
300 (lambda x: x)(**{'x': 0})
302 def compat_kwargs(kwargs):
303 return dict((bytes(k), v) for k, v in kwargs.items())
305 compat_kwargs = lambda kwargs: kwargs
308 # Fix https://github.com/rg3/youtube-dl/issues/4223
309 # See http://bugs.python.org/issue9161 for what is broken
310 def workaround_optparse_bug9161():
311 op = optparse.OptionParser()
312 og = optparse.OptionGroup(op, 'foo')
316 real_add_option = optparse.OptionGroup.add_option
318 def _compat_add_option(self, *args, **kwargs):
320 v.encode('ascii', 'replace') if isinstance(v, compat_str)
322 bargs = [enc(a) for a in args]
324 (k, enc(v)) for k, v in kwargs.items())
325 return real_add_option(self, *bargs, **bkwargs)
326 optparse.OptionGroup.add_option = _compat_add_option
336 'compat_html_entities',
337 'compat_html_parser',
338 'compat_http_client',
344 'compat_subprocess_get_DEVNULL',
345 'compat_urllib_error',
346 'compat_urllib_parse',
347 'compat_urllib_parse_unquote',
348 'compat_urllib_parse_urlparse',
349 'compat_urllib_request',
351 'compat_urlretrieve',
352 'compat_xml_parse_error',
354 'subprocess_check_output',
355 'workaround_optparse_bug9161',