Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/compat.py

   1 from __future__ import unicode_literals
   2
   3 import binascii
   4 import collections
   5 import email
   6 import getpass
   7 import io
   8 import optparse
   9 import os
  10 import re
  11 import shlex
  12 import shutil
  13 import socket
  14 import subprocess
  15 import sys
  16 import itertools
  17 import xml.etree.ElementTree
  18
  19
  20 try:
  21     import urllib.request as compat_urllib_request
  22 except ImportError:  # Python 2
  23     import urllib2 as compat_urllib_request
  24
  25 try:
  26     import urllib.error as compat_urllib_error
  27 except ImportError:  # Python 2
  28     import urllib2 as compat_urllib_error
  29
  30 try:
  31     import urllib.parse as compat_urllib_parse
  32 except ImportError:  # Python 2
  33     import urllib as compat_urllib_parse
  34
  35 try:
  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
  39
  40 try:
  41     import urllib.parse as compat_urlparse
  42 except ImportError:  # Python 2
  43     import urlparse as compat_urlparse
  44
  45 try:
  46     import urllib.response as compat_urllib_response
  47 except ImportError:  # Python 2
  48     import urllib as compat_urllib_response
  49
  50 try:
  51     import http.cookiejar as compat_cookiejar
  52 except ImportError:  # Python 2
  53     import cookielib as compat_cookiejar
  54
  55 try:
  56     import http.cookies as compat_cookies
  57 except ImportError:  # Python 2
  58     import Cookie as compat_cookies
  59
  60 try:
  61     import html.entities as compat_html_entities
  62 except ImportError:  # Python 2
  63     import htmlentitydefs as compat_html_entities
  64
  65 try:
  66     import http.client as compat_http_client
  67 except ImportError:  # Python 2
  68     import httplib as compat_http_client
  69
  70 try:
  71     from urllib.error import HTTPError as compat_HTTPError
  72 except ImportError:  # Python 2
  73     from urllib2 import HTTPError as compat_HTTPError
  74
  75 try:
  76     from urllib.request import urlretrieve as compat_urlretrieve
  77 except ImportError:  # Python 2
  78     from urllib import urlretrieve as compat_urlretrieve
  79
  80
  81 try:
  82     from subprocess import DEVNULL
  83     compat_subprocess_get_DEVNULL = lambda: DEVNULL
  84 except ImportError:
  85     compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w')
  86
  87 try:
  88     import http.server as compat_http_server
  89 except ImportError:
  90     import BaseHTTPServer as compat_http_server
  91
  92 try:
  93     compat_str = unicode  # Python 2
  94 except NameError:
  95     compat_str = str
  96
  97 try:
  98     from urllib.parse import unquote_to_bytes as compat_urllib_parse_unquote_to_bytes
  99     from urllib.parse import unquote as compat_urllib_parse_unquote
 100     from urllib.parse import unquote_plus as compat_urllib_parse_unquote_plus
 101 except ImportError:  # Python 2
 102     _asciire = (compat_urllib_parse._asciire if hasattr(compat_urllib_parse, '_asciire')
 103                 else re.compile('([\x00-\x7f]+)'))
 104
 105     # HACK: The following are the correct unquote_to_bytes, unquote and unquote_plus
 106     # implementations from cpython 3.4.3's stdlib. Python 2's version
 107     # is apparently broken (see https://github.com/rg3/youtube-dl/pull/6244)
 108
 109     def compat_urllib_parse_unquote_to_bytes(string):
 110         """unquote_to_bytes('abc%20def') -> b'abc def'."""
 111         # Note: strings are encoded as UTF-8. This is only an issue if it contains
 112         # unescaped non-ASCII characters, which URIs should not.
 113         if not string:
 114             # Is it a string-like object?
 115             string.split
 116             return b''
 117         if isinstance(string, compat_str):
 118             string = string.encode('utf-8')
 119         bits = string.split(b'%')
 120         if len(bits) == 1:
 121             return string
 122         res = [bits[0]]
 123         append = res.append
 124         for item in bits[1:]:
 125             try:
 126                 append(compat_urllib_parse._hextochr[item[:2]])
 127                 append(item[2:])
 128             except KeyError:
 129                 append(b'%')
 130                 append(item)
 131         return b''.join(res)
 132
 133     def compat_urllib_parse_unquote(string, encoding='utf-8', errors='replace'):
 134         """Replace %xx escapes by their single-character equivalent. The optional
 135         encoding and errors parameters specify how to decode percent-encoded
 136         sequences into Unicode characters, as accepted by the bytes.decode()
 137         method.
 138         By default, percent-encoded sequences are decoded with UTF-8, and invalid
 139         sequences are replaced by a placeholder character.
 140
 141         unquote('abc%20def') -> 'abc def'.
 142         """
 143         if '%' not in string:
 144             string.split
 145             return string
 146         if encoding is None:
 147             encoding = 'utf-8'
 148         if errors is None:
 149             errors = 'replace'
 150         bits = _asciire.split(string)
 151         res = [bits[0]]
 152         append = res.append
 153         for i in range(1, len(bits), 2):
 154             append(compat_urllib_parse_unquote_to_bytes(bits[i]).decode(encoding, errors))
 155             append(bits[i + 1])
 156         return ''.join(res)
 157
 158     def compat_urllib_parse_unquote_plus(string, encoding='utf-8', errors='replace'):
 159         """Like unquote(), but also replace plus signs by spaces, as required for
 160         unquoting HTML form values.
 161
 162         unquote_plus('%7e/abc+def') -> '~/abc def'
 163         """
 164         string = string.replace('+', ' ')
 165         return compat_urllib_parse_unquote(string, encoding, errors)
 166
 167 try:
 168     from urllib.request import DataHandler as compat_urllib_request_DataHandler
 169 except ImportError:  # Python < 3.4
 170     # Ported from CPython 98774:1733b3bd46db, Lib/urllib/request.py
 171     class compat_urllib_request_DataHandler(compat_urllib_request.BaseHandler):
 172         def data_open(self, req):
 173             # data URLs as specified in RFC 2397.
 174             #
 175             # ignores POSTed data
 176             #
 177             # syntax:
 178             # dataurl   := "data:" [ mediatype ] [ ";base64" ] "," data
 179             # mediatype := [ type "/" subtype ] *( ";" parameter )
 180             # data      := *urlchar
 181             # parameter := attribute "=" value
 182             url = req.get_full_url()
 183
 184             scheme, data = url.split(":", 1)
 185             mediatype, data = data.split(",", 1)
 186
 187             # even base64 encoded data URLs might be quoted so unquote in any case:
 188             data = compat_urllib_parse_unquote_to_bytes(data)
 189             if mediatype.endswith(";base64"):
 190                 data = binascii.a2b_base64(data)
 191                 mediatype = mediatype[:-7]
 192
 193             if not mediatype:
 194                 mediatype = "text/plain;charset=US-ASCII"
 195
 196             headers = email.message_from_string(
 197                 "Content-type: %s\nContent-length: %d\n" % (mediatype, len(data)))
 198
 199             return compat_urllib_response.addinfourl(io.BytesIO(data), headers, url)
 200
 201 try:
 202     compat_basestring = basestring  # Python 2
 203 except NameError:
 204     compat_basestring = str
 205
 206 try:
 207     compat_chr = unichr  # Python 2
 208 except NameError:
 209     compat_chr = chr
 210
 211 try:
 212     from xml.etree.ElementTree import ParseError as compat_xml_parse_error
 213 except ImportError:  # Python 2.6
 214     from xml.parsers.expat import ExpatError as compat_xml_parse_error
 215
 216 if sys.version_info[0] >= 3:
 217     compat_etree_fromstring = xml.etree.ElementTree.fromstring
 218 else:
 219     # python 2.x tries to encode unicode strings with ascii (see the
 220     # XMLParser._fixtext method)
 221     etree = xml.etree.ElementTree
 222
 223     try:
 224         _etree_iter = etree.Element.iter
 225     except AttributeError:  # Python <=2.6
 226         def _etree_iter(root):
 227             for el in root.findall('*'):
 228                 yield el
 229                 for sub in _etree_iter(el):
 230                     yield sub
 231
 232     # on 2.6 XML doesn't have a parser argument, function copied from CPython
 233     # 2.7 source
 234     def _XML(text, parser=None):
 235         if not parser:
 236             parser = etree.XMLParser(target=etree.TreeBuilder())
 237         parser.feed(text)
 238         return parser.close()
 239
 240     def _element_factory(*args, **kwargs):
 241         el = etree.Element(*args, **kwargs)
 242         for k, v in el.items():
 243             if isinstance(v, bytes):
 244                 el.set(k, v.decode('utf-8'))
 245         return el
 246
 247     def compat_etree_fromstring(text):
 248         doc = _XML(text, parser=etree.XMLParser(target=etree.TreeBuilder(element_factory=_element_factory)))
 249         for el in _etree_iter(doc):
 250             if el.text is not None and isinstance(el.text, bytes):
 251                 el.text = el.text.decode('utf-8')
 252         return doc
 253
 254 try:
 255     from urllib.parse import parse_qs as compat_parse_qs
 256 except ImportError:  # Python 2
 257     # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib.
 258     # Python 2's version is apparently totally broken
 259
 260     def _parse_qsl(qs, keep_blank_values=False, strict_parsing=False,
 261                    encoding='utf-8', errors='replace'):
 262         qs, _coerce_result = qs, compat_str
 263         pairs = [s2 for s1 in qs.split('&') for s2 in s1.split(';')]
 264         r = []
 265         for name_value in pairs:
 266             if not name_value and not strict_parsing:
 267                 continue
 268             nv = name_value.split('=', 1)
 269             if len(nv) != 2:
 270                 if strict_parsing:
 271                     raise ValueError("bad query field: %r" % (name_value,))
 272                 # Handle case of a control-name with no equal sign
 273                 if keep_blank_values:
 274                     nv.append('')
 275                 else:
 276                     continue
 277             if len(nv[1]) or keep_blank_values:
 278                 name = nv[0].replace('+', ' ')
 279                 name = compat_urllib_parse_unquote(
 280                     name, encoding=encoding, errors=errors)
 281                 name = _coerce_result(name)
 282                 value = nv[1].replace('+', ' ')
 283                 value = compat_urllib_parse_unquote(
 284                     value, encoding=encoding, errors=errors)
 285                 value = _coerce_result(value)
 286                 r.append((name, value))
 287         return r
 288
 289     def compat_parse_qs(qs, keep_blank_values=False, strict_parsing=False,
 290                         encoding='utf-8', errors='replace'):
 291         parsed_result = {}
 292         pairs = _parse_qsl(qs, keep_blank_values, strict_parsing,
 293                            encoding=encoding, errors=errors)
 294         for name, value in pairs:
 295             if name in parsed_result:
 296                 parsed_result[name].append(value)
 297             else:
 298                 parsed_result[name] = [value]
 299         return parsed_result
 300
 301 try:
 302     from shlex import quote as shlex_quote
 303 except ImportError:  # Python < 3.3
 304     def shlex_quote(s):
 305         if re.match(r'^[-_\w./]+$', s):
 306             return s
 307         else:
 308             return "'" + s.replace("'", "'\"'\"'") + "'"
 309
 310
 311 if sys.version_info >= (2, 7, 3):
 312     compat_shlex_split = shlex.split
 313 else:
 314     # Working around shlex issue with unicode strings on some python 2
 315     # versions (see http://bugs.python.org/issue1548891)
 316     def compat_shlex_split(s, comments=False, posix=True):
 317         if isinstance(s, compat_str):
 318             s = s.encode('utf-8')
 319         return shlex.split(s, comments, posix)
 320
 321
 322 def compat_ord(c):
 323     if type(c) is int:
 324         return c
 325     else:
 326         return ord(c)
 327
 328
 329 if sys.version_info >= (3, 0):
 330     compat_getenv = os.getenv
 331     compat_expanduser = os.path.expanduser
 332 else:
 333     # Environment variables should be decoded with filesystem encoding.
 334     # Otherwise it will fail if any non-ASCII characters present (see #3854 #3217 #2918)
 335
 336     def compat_getenv(key, default=None):
 337         from .utils import get_filesystem_encoding
 338         env = os.getenv(key, default)
 339         if env:
 340             env = env.decode(get_filesystem_encoding())
 341         return env
 342
 343     # HACK: The default implementations of os.path.expanduser from cpython do not decode
 344     # environment variables with filesystem encoding. We will work around this by
 345     # providing adjusted implementations.
 346     # The following are os.path.expanduser implementations from cpython 2.7.8 stdlib
 347     # for different platforms with correct environment variables decoding.
 348
 349     if os.name == 'posix':
 350         def compat_expanduser(path):
 351             """Expand ~ and ~user constructions.  If user or $HOME is unknown,
 352             do nothing."""
 353             if not path.startswith('~'):
 354                 return path
 355             i = path.find('/', 1)
 356             if i < 0:
 357                 i = len(path)
 358             if i == 1:
 359                 if 'HOME' not in os.environ:
 360                     import pwd
 361                     userhome = pwd.getpwuid(os.getuid()).pw_dir
 362                 else:
 363                     userhome = compat_getenv('HOME')
 364             else:
 365                 import pwd
 366                 try:
 367                     pwent = pwd.getpwnam(path[1:i])
 368                 except KeyError:
 369                     return path
 370                 userhome = pwent.pw_dir
 371             userhome = userhome.rstrip('/')
 372             return (userhome + path[i:]) or '/'
 373     elif os.name == 'nt' or os.name == 'ce':
 374         def compat_expanduser(path):
 375             """Expand ~ and ~user constructs.
 376
 377             If user or $HOME is unknown, do nothing."""
 378             if path[:1] != '~':
 379                 return path
 380             i, n = 1, len(path)
 381             while i < n and path[i] not in '/\\':
 382                 i = i + 1
 383
 384             if 'HOME' in os.environ:
 385                 userhome = compat_getenv('HOME')
 386             elif 'USERPROFILE' in os.environ:
 387                 userhome = compat_getenv('USERPROFILE')
 388             elif 'HOMEPATH' not in os.environ:
 389                 return path
 390             else:
 391                 try:
 392                     drive = compat_getenv('HOMEDRIVE')
 393                 except KeyError:
 394                     drive = ''
 395                 userhome = os.path.join(drive, compat_getenv('HOMEPATH'))
 396
 397             if i != 1:  # ~user
 398                 userhome = os.path.join(os.path.dirname(userhome), path[1:i])
 399
 400             return userhome + path[i:]
 401     else:
 402         compat_expanduser = os.path.expanduser
 403
 404
 405 if sys.version_info < (3, 0):
 406     def compat_print(s):
 407         from .utils import preferredencoding
 408         print(s.encode(preferredencoding(), 'xmlcharrefreplace'))
 409 else:
 410     def compat_print(s):
 411         assert isinstance(s, compat_str)
 412         print(s)
 413
 414
 415 try:
 416     subprocess_check_output = subprocess.check_output
 417 except AttributeError:
 418     def subprocess_check_output(*args, **kwargs):
 419         assert 'input' not in kwargs
 420         p = subprocess.Popen(*args, stdout=subprocess.PIPE, **kwargs)
 421         output, _ = p.communicate()
 422         ret = p.poll()
 423         if ret:
 424             raise subprocess.CalledProcessError(ret, p.args, output=output)
 425         return output
 426
 427 if sys.version_info < (3, 0) and sys.platform == 'win32':
 428     def compat_getpass(prompt, *args, **kwargs):
 429         if isinstance(prompt, compat_str):
 430             from .utils import preferredencoding
 431             prompt = prompt.encode(preferredencoding())
 432         return getpass.getpass(prompt, *args, **kwargs)
 433 else:
 434     compat_getpass = getpass.getpass
 435
 436 # Old 2.6 and 2.7 releases require kwargs to be bytes
 437 try:
 438     def _testfunc(x):
 439         pass
 440     _testfunc(**{'x': 0})
 441 except TypeError:
 442     def compat_kwargs(kwargs):
 443         return dict((bytes(k), v) for k, v in kwargs.items())
 444 else:
 445     compat_kwargs = lambda kwargs: kwargs
 446
 447
 448 if sys.version_info < (2, 7):
 449     def compat_socket_create_connection(address, timeout, source_address=None):
 450         host, port = address
 451         err = None
 452         for res in socket.getaddrinfo(host, port, 0, socket.SOCK_STREAM):
 453             af, socktype, proto, canonname, sa = res
 454             sock = None
 455             try:
 456                 sock = socket.socket(af, socktype, proto)
 457                 sock.settimeout(timeout)
 458                 if source_address:
 459                     sock.bind(source_address)
 460                 sock.connect(sa)
 461                 return sock
 462             except socket.error as _:
 463                 err = _
 464                 if sock is not None:
 465                     sock.close()
 466         if err is not None:
 467             raise err
 468         else:
 469             raise socket.error("getaddrinfo returns an empty list")
 470 else:
 471     compat_socket_create_connection = socket.create_connection
 472
 473
 474 # Fix https://github.com/rg3/youtube-dl/issues/4223
 475 # See http://bugs.python.org/issue9161 for what is broken
 476 def workaround_optparse_bug9161():
 477     op = optparse.OptionParser()
 478     og = optparse.OptionGroup(op, 'foo')
 479     try:
 480         og.add_option('-t')
 481     except TypeError:
 482         real_add_option = optparse.OptionGroup.add_option
 483
 484         def _compat_add_option(self, *args, **kwargs):
 485             enc = lambda v: (
 486                 v.encode('ascii', 'replace') if isinstance(v, compat_str)
 487                 else v)
 488             bargs = [enc(a) for a in args]
 489             bkwargs = dict(
 490                 (k, enc(v)) for k, v in kwargs.items())
 491             return real_add_option(self, *bargs, **bkwargs)
 492         optparse.OptionGroup.add_option = _compat_add_option
 493
 494 if hasattr(shutil, 'get_terminal_size'):  # Python >= 3.3
 495     compat_get_terminal_size = shutil.get_terminal_size
 496 else:
 497     _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines'])
 498
 499     def compat_get_terminal_size(fallback=(80, 24)):
 500         columns = compat_getenv('COLUMNS')
 501         if columns:
 502             columns = int(columns)
 503         else:
 504             columns = None
 505         lines = compat_getenv('LINES')
 506         if lines:
 507             lines = int(lines)
 508         else:
 509             lines = None
 510
 511         if columns is None or lines is None or columns <= 0 or lines <= 0:
 512             try:
 513                 sp = subprocess.Popen(
 514                     ['stty', 'size'],
 515                     stdout=subprocess.PIPE, stderr=subprocess.PIPE)
 516                 out, err = sp.communicate()
 517                 _lines, _columns = map(int, out.split())
 518             except Exception:
 519                 _columns, _lines = _terminal_size(*fallback)
 520
 521             if columns is None or columns <= 0:
 522                 columns = _columns
 523             if lines is None or lines <= 0:
 524                 lines = _lines
 525         return _terminal_size(columns, lines)
 526
 527 try:
 528     itertools.count(start=0, step=1)
 529     compat_itertools_count = itertools.count
 530 except TypeError:  # Python 2.6
 531     def compat_itertools_count(start=0, step=1):
 532         n = start
 533         while True:
 534             yield n
 535             n += step
 536
 537 if sys.version_info >= (3, 0):
 538     from tokenize import tokenize as compat_tokenize_tokenize
 539 else:
 540     from tokenize import generate_tokens as compat_tokenize_tokenize
 541
 542 __all__ = [
 543     'compat_HTTPError',
 544     'compat_basestring',
 545     'compat_chr',
 546     'compat_cookiejar',
 547     'compat_cookies',
 548     'compat_etree_fromstring',
 549     'compat_expanduser',
 550     'compat_get_terminal_size',
 551     'compat_getenv',
 552     'compat_getpass',
 553     'compat_html_entities',
 554     'compat_http_client',
 555     'compat_http_server',
 556     'compat_itertools_count',
 557     'compat_kwargs',
 558     'compat_ord',
 559     'compat_parse_qs',
 560     'compat_print',
 561     'compat_shlex_split',
 562     'compat_socket_create_connection',
 563     'compat_str',
 564     'compat_subprocess_get_DEVNULL',
 565     'compat_tokenize_tokenize',
 566     'compat_urllib_error',
 567     'compat_urllib_parse',
 568     'compat_urllib_parse_unquote',
 569     'compat_urllib_parse_unquote_plus',
 570     'compat_urllib_parse_unquote_to_bytes',
 571     'compat_urllib_parse_urlparse',
 572     'compat_urllib_request',
 573     'compat_urllib_request_DataHandler',
 574     'compat_urllib_response',
 575     'compat_urlparse',
 576     'compat_urlretrieve',
 577     'compat_xml_parse_error',
 578     'shlex_quote',
 579     'subprocess_check_output',
 580     'workaround_optparse_bug9161',
 581 ]