Adding upstream version 2024.03.10.upstream/2024.03.10

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

10 files changed, 414 insertions, 0 deletions

diff --git a/yt_dlp/compat/__init__.py b/yt_dlp/compat/__init__.py
new file mode 100644
index 0000000..5ad5c70
--- /dev/null
+++ b/yt_dlp/compat/__init__.py
@@ -0,0 +1,79 @@
+import os
+import sys
+import xml.etree.ElementTree as etree
+
+from .compat_utils import passthrough_module
+
+passthrough_module(__name__, '._deprecated')
+del passthrough_module
+
+
+# HTMLParseError has been deprecated in Python 3.3 and removed in
+# Python 3.5. Introducing dummy exception for Python >3.5 for compatible
+# and uniform cross-version exception handling
+class compat_HTMLParseError(ValueError):
+    pass
+
+
+class _TreeBuilder(etree.TreeBuilder):
+    def doctype(self, name, pubid, system):
+        pass
+
+
+def compat_etree_fromstring(text):
+    return etree.XML(text, parser=etree.XMLParser(target=_TreeBuilder()))
+
+
+compat_os_name = os._name if os.name == 'java' else os.name
+
+
+if compat_os_name == 'nt':
+    def compat_shlex_quote(s):
+        import re
+        return s if re.match(r'^[-_\w./]+$', s) else s.replace('"', '""').join('""')
+else:
+    from shlex import quote as compat_shlex_quote  # noqa: F401
+
+
+def compat_ord(c):
+    return c if isinstance(c, int) else ord(c)
+
+
+if compat_os_name == 'nt' and sys.version_info < (3, 8):
+    # os.path.realpath on Windows does not follow symbolic links
+    # prior to Python 3.8 (see https://bugs.python.org/issue9949)
+    def compat_realpath(path):
+        while os.path.islink(path):
+            path = os.path.abspath(os.readlink(path))
+        return os.path.realpath(path)
+else:
+    compat_realpath = os.path.realpath
+
+
+# Python 3.8+ does not honor %HOME% on windows, but this breaks compatibility with youtube-dl
+# See https://github.com/yt-dlp/yt-dlp/issues/792
+# https://docs.python.org/3/library/os.path.html#os.path.expanduser
+if compat_os_name in ('nt', 'ce'):
+    def compat_expanduser(path):
+        HOME = os.environ.get('HOME')
+        if not HOME:
+            return os.path.expanduser(path)
+        elif not path.startswith('~'):
+            return path
+        i = path.replace('\\', '/', 1).find('/')  # ~user
+        if i < 0:
+            i = len(path)
+        userhome = os.path.join(os.path.dirname(HOME), path[1:i]) if i > 1 else HOME
+        return userhome + path[i:]
+else:
+    compat_expanduser = os.path.expanduser
+
+
+def urllib_req_to_req(urllib_request):
+    """Convert urllib Request to a networking Request"""
+    from ..networking import Request
+    from ..utils.networking import HTTPHeaderDict
+    return Request(
+        urllib_request.get_full_url(), data=urllib_request.data, method=urllib_request.get_method(),
+        headers=HTTPHeaderDict(urllib_request.headers, urllib_request.unredirected_hdrs),
+        extensions={'timeout': urllib_request.timeout} if hasattr(urllib_request, 'timeout') else None)
diff --git a/yt_dlp/compat/_deprecated.py b/yt_dlp/compat/_deprecated.py
new file mode 100644
index 0000000..607bae9
--- /dev/null
+++ b/yt_dlp/compat/_deprecated.py
@@ -0,0 +1,23 @@
+"""Deprecated - New code should avoid these"""
+import warnings
+
+from .compat_utils import passthrough_module
+
+# XXX: Implement this the same way as other DeprecationWarnings without circular import
+passthrough_module(__name__, '.._legacy', callback=lambda attr: warnings.warn(
+    DeprecationWarning(f'{__name__}.{attr} is deprecated'), stacklevel=6))
+del passthrough_module
+
+import base64
+import urllib.error
+import urllib.parse
+
+compat_str = str
+
+compat_b64decode = base64.b64decode
+
+compat_urlparse = urllib.parse
+compat_parse_qs = urllib.parse.parse_qs
+compat_urllib_parse_unquote = urllib.parse.unquote
+compat_urllib_parse_urlencode = urllib.parse.urlencode
+compat_urllib_parse_urlparse = urllib.parse.urlparse
diff --git a/yt_dlp/compat/_legacy.py b/yt_dlp/compat/_legacy.py
new file mode 100644
index 0000000..7ea5d08
--- /dev/null
+++ b/yt_dlp/compat/_legacy.py
@@ -0,0 +1,108 @@
+""" Do not use! """
+
+import base64
+import collections
+import ctypes
+import getpass
+import html.entities
+import html.parser
+import http.client
+import http.cookiejar
+import http.cookies
+import http.server
+import itertools
+import os
+import shlex
+import shutil
+import socket
+import struct
+import subprocess
+import tokenize
+import urllib.error
+import urllib.parse
+import urllib.request
+import xml.etree.ElementTree as etree
+
+# isort: split
+import asyncio  # noqa: F401
+import re  # noqa: F401
+from asyncio import run as compat_asyncio_run  # noqa: F401
+from re import Pattern as compat_Pattern  # noqa: F401
+from re import match as compat_Match  # noqa: F401
+
+from . import compat_expanduser, compat_HTMLParseError, compat_realpath
+from .compat_utils import passthrough_module
+from ..dependencies import brotli as compat_brotli  # noqa: F401
+from ..dependencies import websockets as compat_websockets  # noqa: F401
+from ..dependencies.Cryptodome import AES as compat_pycrypto_AES  # noqa: F401
+from ..networking.exceptions import HTTPError as compat_HTTPError  # noqa: F401
+
+passthrough_module(__name__, '...utils', ('WINDOWS_VT_MODE', 'windows_enable_vt_mode'))
+
+
+# compat_ctypes_WINFUNCTYPE = ctypes.WINFUNCTYPE
+# will not work since ctypes.WINFUNCTYPE does not exist in UNIX machines
+def compat_ctypes_WINFUNCTYPE(*args, **kwargs):
+    return ctypes.WINFUNCTYPE(*args, **kwargs)
+
+
+def compat_setenv(key, value, env=os.environ):
+    env[key] = value
+
+
+compat_base64_b64decode = base64.b64decode
+compat_basestring = str
+compat_casefold = str.casefold
+compat_chr = chr
+compat_collections_abc = collections.abc
+compat_cookiejar = compat_http_cookiejar = http.cookiejar
+compat_cookiejar_Cookie = compat_http_cookiejar_Cookie = http.cookiejar.Cookie
+compat_cookies = compat_http_cookies = http.cookies
+compat_cookies_SimpleCookie = compat_http_cookies_SimpleCookie = http.cookies.SimpleCookie
+compat_etree_Element = compat_xml_etree_ElementTree_Element = etree.Element
+compat_etree_register_namespace = compat_xml_etree_register_namespace = etree.register_namespace
+compat_filter = filter
+compat_get_terminal_size = shutil.get_terminal_size
+compat_getenv = os.getenv
+compat_getpass = compat_getpass_getpass = getpass.getpass
+compat_html_entities = html.entities
+compat_html_entities_html5 = html.entities.html5
+compat_html_parser_HTMLParseError = compat_HTMLParseError
+compat_HTMLParser = compat_html_parser_HTMLParser = html.parser.HTMLParser
+compat_http_client = http.client
+compat_http_server = http.server
+compat_input = input
+compat_integer_types = (int, )
+compat_itertools_count = itertools.count
+compat_kwargs = lambda kwargs: kwargs
+compat_map = map
+compat_numeric_types = (int, float, complex)
+compat_os_path_expanduser = compat_expanduser
+compat_os_path_realpath = compat_realpath
+compat_print = print
+compat_shlex_split = shlex.split
+compat_socket_create_connection = socket.create_connection
+compat_Struct = struct.Struct
+compat_struct_pack = struct.pack
+compat_struct_unpack = struct.unpack
+compat_subprocess_get_DEVNULL = lambda: subprocess.DEVNULL
+compat_tokenize_tokenize = tokenize.tokenize
+compat_urllib_error = urllib.error
+compat_urllib_HTTPError = compat_HTTPError
+compat_urllib_parse = urllib.parse
+compat_urllib_parse_parse_qs = urllib.parse.parse_qs
+compat_urllib_parse_quote = urllib.parse.quote
+compat_urllib_parse_quote_plus = urllib.parse.quote_plus
+compat_urllib_parse_unquote_plus = urllib.parse.unquote_plus
+compat_urllib_parse_unquote_to_bytes = urllib.parse.unquote_to_bytes
+compat_urllib_parse_urlunparse = urllib.parse.urlunparse
+compat_urllib_request = urllib.request
+compat_urllib_request_DataHandler = urllib.request.DataHandler
+compat_urllib_response = urllib.response
+compat_urlretrieve = compat_urllib_request_urlretrieve = urllib.request.urlretrieve
+compat_xml_parse_error = compat_xml_etree_ElementTree_ParseError = etree.ParseError
+compat_xpath = lambda xpath: xpath
+compat_zip = zip
+workaround_optparse_bug9161 = lambda: None
+
+legacy = []
diff --git a/yt_dlp/compat/compat_utils.py b/yt_dlp/compat/compat_utils.py
new file mode 100644
index 0000000..d62b7d0
--- /dev/null
+++ b/yt_dlp/compat/compat_utils.py
@@ -0,0 +1,83 @@
+import collections
+import contextlib
+import functools
+import importlib
+import sys
+import types
+
+_NO_ATTRIBUTE = object()
+
+_Package = collections.namedtuple('Package', ('name', 'version'))
+
+
+def get_package_info(module):
+    return _Package(
+        name=getattr(module, '_yt_dlp__identifier', module.__name__),
+        version=str(next(filter(None, (
+            getattr(module, attr, None)
+            for attr in ('_yt_dlp__version', '__version__', 'version_string', 'version')
+        )), None)))
+
+
+def _is_package(module):
+    return '__path__' in vars(module)
+
+
+def _is_dunder(name):
+    return name.startswith('__') and name.endswith('__')
+
+
+class EnhancedModule(types.ModuleType):
+    def __bool__(self):
+        return vars(self).get('__bool__', lambda: True)()
+
+    def __getattribute__(self, attr):
+        try:
+            ret = super().__getattribute__(attr)
+        except AttributeError:
+            if _is_dunder(attr):
+                raise
+            getter = getattr(self, '__getattr__', None)
+            if not getter:
+                raise
+            ret = getter(attr)
+        return ret.fget() if isinstance(ret, property) else ret
+
+
+def passthrough_module(parent, child, allowed_attributes=(..., ), *, callback=lambda _: None):
+    """Passthrough parent module into a child module, creating the parent if necessary"""
+    def __getattr__(attr):
+        if _is_package(parent):
+            with contextlib.suppress(ModuleNotFoundError):
+                return importlib.import_module(f'.{attr}', parent.__name__)
+
+        ret = from_child(attr)
+        if ret is _NO_ATTRIBUTE:
+            raise AttributeError(f'module {parent.__name__} has no attribute {attr}')
+        callback(attr)
+        return ret
+
+    @functools.lru_cache(maxsize=None)
+    def from_child(attr):
+        nonlocal child
+        if attr not in allowed_attributes:
+            if ... not in allowed_attributes or _is_dunder(attr):
+                return _NO_ATTRIBUTE
+
+        if isinstance(child, str):
+            child = importlib.import_module(child, parent.__name__)
+
+        if _is_package(child):
+            with contextlib.suppress(ImportError):
+                return passthrough_module(f'{parent.__name__}.{attr}',
+                                          importlib.import_module(f'.{attr}', child.__name__))
+
+        with contextlib.suppress(AttributeError):
+            return getattr(child, attr)
+
+        return _NO_ATTRIBUTE
+
+    parent = sys.modules.get(parent, types.ModuleType(parent))
+    parent.__class__ = EnhancedModule
+    parent.__getattr__ = __getattr__
+    return parent
diff --git a/yt_dlp/compat/functools.py b/yt_dlp/compat/functools.py
new file mode 100644
index 0000000..36c9836
--- /dev/null
+++ b/yt_dlp/compat/functools.py
@@ -0,0 +1,12 @@
+# flake8: noqa: F405
+from functools import *  # noqa: F403
+
+from .compat_utils import passthrough_module
+
+passthrough_module(__name__, 'functools')
+del passthrough_module
+
+try:
+    cache  # >= 3.9
+except NameError:
+    cache = lru_cache(maxsize=None)
diff --git a/yt_dlp/compat/imghdr.py b/yt_dlp/compat/imghdr.py
new file mode 100644
index 0000000..5d64ab0
--- /dev/null
+++ b/yt_dlp/compat/imghdr.py
@@ -0,0 +1,16 @@
+tests = {
+    'webp': lambda h: h[0:4] == b'RIFF' and h[8:] == b'WEBP',
+    'png': lambda h: h[:8] == b'\211PNG\r\n\032\n',
+    'jpeg': lambda h: h[6:10] in (b'JFIF', b'Exif'),
+    'gif': lambda h: h[:6] in (b'GIF87a', b'GIF89a'),
+}
+
+
+def what(file=None, h=None):
+    """Detect format of image (Currently supports jpeg, png, webp, gif only)
+    Ref: https://github.com/python/cpython/blob/3.10/Lib/imghdr.py
+    """
+    if h is None:
+        with open(file, 'rb') as f:
+            h = f.read(12)
+    return next((type_ for type_, test in tests.items() if test(h)), None)
diff --git a/yt_dlp/compat/shutil.py b/yt_dlp/compat/shutil.py
new file mode 100644
index 0000000..23239d5
--- /dev/null
+++ b/yt_dlp/compat/shutil.py
@@ -0,0 +1,30 @@
+# flake8: noqa: F405
+from shutil import *  # noqa: F403
+
+from .compat_utils import passthrough_module
+
+passthrough_module(__name__, 'shutil')
+del passthrough_module
+
+
+import sys
+
+if sys.platform.startswith('freebsd'):
+    import errno
+    import os
+    import shutil
+
+    # Workaround for PermissionError when using restricted ACL mode on FreeBSD
+    def copy2(src, dst, *args, **kwargs):
+        if os.path.isdir(dst):
+            dst = os.path.join(dst, os.path.basename(src))
+        shutil.copyfile(src, dst, *args, **kwargs)
+        try:
+            shutil.copystat(src, dst, *args, **kwargs)
+        except PermissionError as e:
+            if e.errno != getattr(errno, 'EPERM', None):
+                raise
+        return dst
+
+    def move(*args, copy_function=copy2, **kwargs):
+        return shutil.move(*args, copy_function=copy_function, **kwargs)
diff --git a/yt_dlp/compat/types.py b/yt_dlp/compat/types.py
new file mode 100644
index 0000000..4aa3b0e
--- /dev/null
+++ b/yt_dlp/compat/types.py
@@ -0,0 +1,13 @@
+# flake8: noqa: F405
+from types import *  # noqa: F403
+
+from .compat_utils import passthrough_module
+
+passthrough_module(__name__, 'types')
+del passthrough_module
+
+try:
+    # NB: pypy has builtin NoneType, so checking NameError won't work
+    from types import NoneType  # >= 3.10
+except ImportError:
+    NoneType = type(None)
diff --git a/yt_dlp/compat/urllib/__init__.py b/yt_dlp/compat/urllib/__init__.py
new file mode 100644
index 0000000..9084b3c
--- /dev/null
+++ b/yt_dlp/compat/urllib/__init__.py
@@ -0,0 +1,10 @@
+# flake8: noqa: F405
+from urllib import *  # noqa: F403
+
+del request  # noqa: F821
+from . import request  # noqa: F401
+
+from ..compat_utils import passthrough_module
+
+passthrough_module(__name__, 'urllib')
+del passthrough_module
diff --git a/yt_dlp/compat/urllib/request.py b/yt_dlp/compat/urllib/request.py
new file mode 100644
index 0000000..ad9fa83
--- /dev/null
+++ b/yt_dlp/compat/urllib/request.py
@@ -0,0 +1,40 @@
+# flake8: noqa: F405
+from urllib.request import *  # noqa: F403
+
+from ..compat_utils import passthrough_module
+
+passthrough_module(__name__, 'urllib.request')
+del passthrough_module
+
+
+from .. import compat_os_name
+
+if compat_os_name == 'nt':
+    # On older Python versions, proxies are extracted from Windows registry erroneously. [1]
+    # If the https proxy in the registry does not have a scheme, urllib will incorrectly add https:// to it. [2]
+    # It is unlikely that the user has actually set it to be https, so we should be fine to safely downgrade
+    # it to http on these older Python versions to avoid issues
+    # This also applies for ftp proxy type, as ftp:// proxy scheme is not supported.
+    # 1: https://github.com/python/cpython/issues/86793
+    # 2: https://github.com/python/cpython/blob/51f1ae5ceb0673316c4e4b0175384e892e33cc6e/Lib/urllib/request.py#L2683-L2698
+    import sys
+    from urllib.request import getproxies_environment, getproxies_registry
+
+    def getproxies_registry_patched():
+        proxies = getproxies_registry()
+        if (
+            sys.version_info >= (3, 10, 5)  # https://docs.python.org/3.10/whatsnew/changelog.html#python-3-10-5-final
+            or (3, 9, 13) <= sys.version_info < (3, 10)  # https://docs.python.org/3.9/whatsnew/changelog.html#python-3-9-13-final
+        ):
+            return proxies
+
+        for scheme in ('https', 'ftp'):
+            if scheme in proxies and proxies[scheme].startswith(f'{scheme}://'):
+                proxies[scheme] = 'http' + proxies[scheme][len(scheme):]
+
+        return proxies
+
+    def getproxies():
+        return getproxies_environment() or getproxies_registry_patched()
+
+del compat_os_name