jpayne@7: from __future__ import annotations
jpayne@7: 
jpayne@7: import datetime
jpayne@7: import logging
jpayne@7: import os
jpayne@7: import re
jpayne@7: import socket
jpayne@7: import sys
jpayne@7: import typing
jpayne@7: import warnings
jpayne@7: from http.client import HTTPConnection as _HTTPConnection
jpayne@7: from http.client import HTTPException as HTTPException  # noqa: F401
jpayne@7: from http.client import ResponseNotReady
jpayne@7: from socket import timeout as SocketTimeout
jpayne@7: 
jpayne@7: if typing.TYPE_CHECKING:
jpayne@7:     from typing import Literal
jpayne@7: 
jpayne@7:     from .response import HTTPResponse
jpayne@7:     from .util.ssl_ import _TYPE_PEER_CERT_RET_DICT
jpayne@7:     from .util.ssltransport import SSLTransport
jpayne@7: 
jpayne@7: from ._collections import HTTPHeaderDict
jpayne@7: from .util.response import assert_header_parsing
jpayne@7: from .util.timeout import _DEFAULT_TIMEOUT, _TYPE_TIMEOUT, Timeout
jpayne@7: from .util.util import to_str
jpayne@7: from .util.wait import wait_for_read
jpayne@7: 
jpayne@7: try:  # Compiled with SSL?
jpayne@7:     import ssl
jpayne@7: 
jpayne@7:     BaseSSLError = ssl.SSLError
jpayne@7: except (ImportError, AttributeError):
jpayne@7:     ssl = None  # type: ignore[assignment]
jpayne@7: 
jpayne@7:     class BaseSSLError(BaseException):  # type: ignore[no-redef]
jpayne@7:         pass
jpayne@7: 
jpayne@7: 
jpayne@7: from ._base_connection import _TYPE_BODY
jpayne@7: from ._base_connection import ProxyConfig as ProxyConfig
jpayne@7: from ._base_connection import _ResponseOptions as _ResponseOptions
jpayne@7: from ._version import __version__
jpayne@7: from .exceptions import (
jpayne@7:     ConnectTimeoutError,
jpayne@7:     HeaderParsingError,
jpayne@7:     NameResolutionError,
jpayne@7:     NewConnectionError,
jpayne@7:     ProxyError,
jpayne@7:     SystemTimeWarning,
jpayne@7: )
jpayne@7: from .util import SKIP_HEADER, SKIPPABLE_HEADERS, connection, ssl_
jpayne@7: from .util.request import body_to_chunks
jpayne@7: from .util.ssl_ import assert_fingerprint as _assert_fingerprint
jpayne@7: from .util.ssl_ import (
jpayne@7:     create_urllib3_context,
jpayne@7:     is_ipaddress,
jpayne@7:     resolve_cert_reqs,
jpayne@7:     resolve_ssl_version,
jpayne@7:     ssl_wrap_socket,
jpayne@7: )
jpayne@7: from .util.ssl_match_hostname import CertificateError, match_hostname
jpayne@7: from .util.url import Url
jpayne@7: 
jpayne@7: # Not a no-op, we're adding this to the namespace so it can be imported.
jpayne@7: ConnectionError = ConnectionError
jpayne@7: BrokenPipeError = BrokenPipeError
jpayne@7: 
jpayne@7: 
jpayne@7: log = logging.getLogger(__name__)
jpayne@7: 
jpayne@7: port_by_scheme = {"http": 80, "https": 443}
jpayne@7: 
jpayne@7: # When it comes time to update this value as a part of regular maintenance
jpayne@7: # (ie test_recent_date is failing) update it to ~6 months before the current date.
jpayne@7: RECENT_DATE = datetime.date(2023, 6, 1)
jpayne@7: 
jpayne@7: _CONTAINS_CONTROL_CHAR_RE = re.compile(r"[^-!#$%&'*+.^_`|~0-9a-zA-Z]")
jpayne@7: 
jpayne@7: _HAS_SYS_AUDIT = hasattr(sys, "audit")
jpayne@7: 
jpayne@7: 
jpayne@7: class HTTPConnection(_HTTPConnection):
jpayne@7:     """
jpayne@7:     Based on :class:`http.client.HTTPConnection` but provides an extra constructor
jpayne@7:     backwards-compatibility layer between older and newer Pythons.
jpayne@7: 
jpayne@7:     Additional keyword parameters are used to configure attributes of the connection.
jpayne@7:     Accepted parameters include:
jpayne@7: 
jpayne@7:     - ``source_address``: Set the source address for the current connection.
jpayne@7:     - ``socket_options``: Set specific options on the underlying socket. If not specified, then
jpayne@7:       defaults are loaded from ``HTTPConnection.default_socket_options`` which includes disabling
jpayne@7:       Nagle's algorithm (sets TCP_NODELAY to 1) unless the connection is behind a proxy.
jpayne@7: 
jpayne@7:       For example, if you wish to enable TCP Keep Alive in addition to the defaults,
jpayne@7:       you might pass:
jpayne@7: 
jpayne@7:       .. code-block:: python
jpayne@7: 
jpayne@7:          HTTPConnection.default_socket_options + [
jpayne@7:              (socket.SOL_SOCKET, socket.SO_KEEPALIVE, 1),
jpayne@7:          ]
jpayne@7: 
jpayne@7:       Or you may want to disable the defaults by passing an empty list (e.g., ``[]``).
jpayne@7:     """
jpayne@7: 
jpayne@7:     default_port: typing.ClassVar[int] = port_by_scheme["http"]  # type: ignore[misc]
jpayne@7: 
jpayne@7:     #: Disable Nagle's algorithm by default.
jpayne@7:     #: ``[(socket.IPPROTO_TCP, socket.TCP_NODELAY, 1)]``
jpayne@7:     default_socket_options: typing.ClassVar[connection._TYPE_SOCKET_OPTIONS] = [
jpayne@7:         (socket.IPPROTO_TCP, socket.TCP_NODELAY, 1)
jpayne@7:     ]
jpayne@7: 
jpayne@7:     #: Whether this connection verifies the host's certificate.
jpayne@7:     is_verified: bool = False
jpayne@7: 
jpayne@7:     #: Whether this proxy connection verified the proxy host's certificate.
jpayne@7:     # If no proxy is currently connected to the value will be ``None``.
jpayne@7:     proxy_is_verified: bool | None = None
jpayne@7: 
jpayne@7:     blocksize: int
jpayne@7:     source_address: tuple[str, int] | None
jpayne@7:     socket_options: connection._TYPE_SOCKET_OPTIONS | None
jpayne@7: 
jpayne@7:     _has_connected_to_proxy: bool
jpayne@7:     _response_options: _ResponseOptions | None
jpayne@7:     _tunnel_host: str | None
jpayne@7:     _tunnel_port: int | None
jpayne@7:     _tunnel_scheme: str | None
jpayne@7: 
jpayne@7:     def __init__(
jpayne@7:         self,
jpayne@7:         host: str,
jpayne@7:         port: int | None = None,
jpayne@7:         *,
jpayne@7:         timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
jpayne@7:         source_address: tuple[str, int] | None = None,
jpayne@7:         blocksize: int = 16384,
jpayne@7:         socket_options: None
jpayne@7:         | (connection._TYPE_SOCKET_OPTIONS) = default_socket_options,
jpayne@7:         proxy: Url | None = None,
jpayne@7:         proxy_config: ProxyConfig | None = None,
jpayne@7:     ) -> None:
jpayne@7:         super().__init__(
jpayne@7:             host=host,
jpayne@7:             port=port,
jpayne@7:             timeout=Timeout.resolve_default_timeout(timeout),
jpayne@7:             source_address=source_address,
jpayne@7:             blocksize=blocksize,
jpayne@7:         )
jpayne@7:         self.socket_options = socket_options
jpayne@7:         self.proxy = proxy
jpayne@7:         self.proxy_config = proxy_config
jpayne@7: 
jpayne@7:         self._has_connected_to_proxy = False
jpayne@7:         self._response_options = None
jpayne@7:         self._tunnel_host: str | None = None
jpayne@7:         self._tunnel_port: int | None = None
jpayne@7:         self._tunnel_scheme: str | None = None
jpayne@7: 
jpayne@7:     @property
jpayne@7:     def host(self) -> str:
jpayne@7:         """
jpayne@7:         Getter method to remove any trailing dots that indicate the hostname is an FQDN.
jpayne@7: 
jpayne@7:         In general, SSL certificates don't include the trailing dot indicating a
jpayne@7:         fully-qualified domain name, and thus, they don't validate properly when
jpayne@7:         checked against a domain name that includes the dot. In addition, some
jpayne@7:         servers may not expect to receive the trailing dot when provided.
jpayne@7: 
jpayne@7:         However, the hostname with trailing dot is critical to DNS resolution; doing a
jpayne@7:         lookup with the trailing dot will properly only resolve the appropriate FQDN,
jpayne@7:         whereas a lookup without a trailing dot will search the system's search domain
jpayne@7:         list. Thus, it's important to keep the original host around for use only in
jpayne@7:         those cases where it's appropriate (i.e., when doing DNS lookup to establish the
jpayne@7:         actual TCP connection across which we're going to send HTTP requests).
jpayne@7:         """
jpayne@7:         return self._dns_host.rstrip(".")
jpayne@7: 
jpayne@7:     @host.setter
jpayne@7:     def host(self, value: str) -> None:
jpayne@7:         """
jpayne@7:         Setter for the `host` property.
jpayne@7: 
jpayne@7:         We assume that only urllib3 uses the _dns_host attribute; httplib itself
jpayne@7:         only uses `host`, and it seems reasonable that other libraries follow suit.
jpayne@7:         """
jpayne@7:         self._dns_host = value
jpayne@7: 
jpayne@7:     def _new_conn(self) -> socket.socket:
jpayne@7:         """Establish a socket connection and set nodelay settings on it.
jpayne@7: 
jpayne@7:         :return: New socket connection.
jpayne@7:         """
jpayne@7:         try:
jpayne@7:             sock = connection.create_connection(
jpayne@7:                 (self._dns_host, self.port),
jpayne@7:                 self.timeout,
jpayne@7:                 source_address=self.source_address,
jpayne@7:                 socket_options=self.socket_options,
jpayne@7:             )
jpayne@7:         except socket.gaierror as e:
jpayne@7:             raise NameResolutionError(self.host, self, e) from e
jpayne@7:         except SocketTimeout as e:
jpayne@7:             raise ConnectTimeoutError(
jpayne@7:                 self,
jpayne@7:                 f"Connection to {self.host} timed out. (connect timeout={self.timeout})",
jpayne@7:             ) from e
jpayne@7: 
jpayne@7:         except OSError as e:
jpayne@7:             raise NewConnectionError(
jpayne@7:                 self, f"Failed to establish a new connection: {e}"
jpayne@7:             ) from e
jpayne@7: 
jpayne@7:         # Audit hooks are only available in Python 3.8+
jpayne@7:         if _HAS_SYS_AUDIT:
jpayne@7:             sys.audit("http.client.connect", self, self.host, self.port)
jpayne@7: 
jpayne@7:         return sock
jpayne@7: 
jpayne@7:     def set_tunnel(
jpayne@7:         self,
jpayne@7:         host: str,
jpayne@7:         port: int | None = None,
jpayne@7:         headers: typing.Mapping[str, str] | None = None,
jpayne@7:         scheme: str = "http",
jpayne@7:     ) -> None:
jpayne@7:         if scheme not in ("http", "https"):
jpayne@7:             raise ValueError(
jpayne@7:                 f"Invalid proxy scheme for tunneling: {scheme!r}, must be either 'http' or 'https'"
jpayne@7:             )
jpayne@7:         super().set_tunnel(host, port=port, headers=headers)
jpayne@7:         self._tunnel_scheme = scheme
jpayne@7: 
jpayne@7:     def connect(self) -> None:
jpayne@7:         self.sock = self._new_conn()
jpayne@7:         if self._tunnel_host:
jpayne@7:             # If we're tunneling it means we're connected to our proxy.
jpayne@7:             self._has_connected_to_proxy = True
jpayne@7: 
jpayne@7:             # TODO: Fix tunnel so it doesn't depend on self.sock state.
jpayne@7:             self._tunnel()  # type: ignore[attr-defined]
jpayne@7: 
jpayne@7:         # If there's a proxy to be connected to we are fully connected.
jpayne@7:         # This is set twice (once above and here) due to forwarding proxies
jpayne@7:         # not using tunnelling.
jpayne@7:         self._has_connected_to_proxy = bool(self.proxy)
jpayne@7: 
jpayne@7:         if self._has_connected_to_proxy:
jpayne@7:             self.proxy_is_verified = False
jpayne@7: 
jpayne@7:     @property
jpayne@7:     def is_closed(self) -> bool:
jpayne@7:         return self.sock is None
jpayne@7: 
jpayne@7:     @property
jpayne@7:     def is_connected(self) -> bool:
jpayne@7:         if self.sock is None:
jpayne@7:             return False
jpayne@7:         return not wait_for_read(self.sock, timeout=0.0)
jpayne@7: 
jpayne@7:     @property
jpayne@7:     def has_connected_to_proxy(self) -> bool:
jpayne@7:         return self._has_connected_to_proxy
jpayne@7: 
jpayne@7:     @property
jpayne@7:     def proxy_is_forwarding(self) -> bool:
jpayne@7:         """
jpayne@7:         Return True if a forwarding proxy is configured, else return False
jpayne@7:         """
jpayne@7:         return bool(self.proxy) and self._tunnel_host is None
jpayne@7: 
jpayne@7:     def close(self) -> None:
jpayne@7:         try:
jpayne@7:             super().close()
jpayne@7:         finally:
jpayne@7:             # Reset all stateful properties so connection
jpayne@7:             # can be re-used without leaking prior configs.
jpayne@7:             self.sock = None
jpayne@7:             self.is_verified = False
jpayne@7:             self.proxy_is_verified = None
jpayne@7:             self._has_connected_to_proxy = False
jpayne@7:             self._response_options = None
jpayne@7:             self._tunnel_host = None
jpayne@7:             self._tunnel_port = None
jpayne@7:             self._tunnel_scheme = None
jpayne@7: 
jpayne@7:     def putrequest(
jpayne@7:         self,
jpayne@7:         method: str,
jpayne@7:         url: str,
jpayne@7:         skip_host: bool = False,
jpayne@7:         skip_accept_encoding: bool = False,
jpayne@7:     ) -> None:
jpayne@7:         """"""
jpayne@7:         # Empty docstring because the indentation of CPython's implementation
jpayne@7:         # is broken but we don't want this method in our documentation.
jpayne@7:         match = _CONTAINS_CONTROL_CHAR_RE.search(method)
jpayne@7:         if match:
jpayne@7:             raise ValueError(
jpayne@7:                 f"Method cannot contain non-token characters {method!r} (found at least {match.group()!r})"
jpayne@7:             )
jpayne@7: 
jpayne@7:         return super().putrequest(
jpayne@7:             method, url, skip_host=skip_host, skip_accept_encoding=skip_accept_encoding
jpayne@7:         )
jpayne@7: 
jpayne@7:     def putheader(self, header: str, *values: str) -> None:  # type: ignore[override]
jpayne@7:         """"""
jpayne@7:         if not any(isinstance(v, str) and v == SKIP_HEADER for v in values):
jpayne@7:             super().putheader(header, *values)
jpayne@7:         elif to_str(header.lower()) not in SKIPPABLE_HEADERS:
jpayne@7:             skippable_headers = "', '".join(
jpayne@7:                 [str.title(header) for header in sorted(SKIPPABLE_HEADERS)]
jpayne@7:             )
jpayne@7:             raise ValueError(
jpayne@7:                 f"urllib3.util.SKIP_HEADER only supports '{skippable_headers}'"
jpayne@7:             )
jpayne@7: 
jpayne@7:     # `request` method's signature intentionally violates LSP.
jpayne@7:     # urllib3's API is different from `http.client.HTTPConnection` and the subclassing is only incidental.
jpayne@7:     def request(  # type: ignore[override]
jpayne@7:         self,
jpayne@7:         method: str,
jpayne@7:         url: str,
jpayne@7:         body: _TYPE_BODY | None = None,
jpayne@7:         headers: typing.Mapping[str, str] | None = None,
jpayne@7:         *,
jpayne@7:         chunked: bool = False,
jpayne@7:         preload_content: bool = True,
jpayne@7:         decode_content: bool = True,
jpayne@7:         enforce_content_length: bool = True,
jpayne@7:     ) -> None:
jpayne@7:         # Update the inner socket's timeout value to send the request.
jpayne@7:         # This only triggers if the connection is re-used.
jpayne@7:         if self.sock is not None:
jpayne@7:             self.sock.settimeout(self.timeout)
jpayne@7: 
jpayne@7:         # Store these values to be fed into the HTTPResponse
jpayne@7:         # object later. TODO: Remove this in favor of a real
jpayne@7:         # HTTP lifecycle mechanism.
jpayne@7: 
jpayne@7:         # We have to store these before we call .request()
jpayne@7:         # because sometimes we can still salvage a response
jpayne@7:         # off the wire even if we aren't able to completely
jpayne@7:         # send the request body.
jpayne@7:         self._response_options = _ResponseOptions(
jpayne@7:             request_method=method,
jpayne@7:             request_url=url,
jpayne@7:             preload_content=preload_content,
jpayne@7:             decode_content=decode_content,
jpayne@7:             enforce_content_length=enforce_content_length,
jpayne@7:         )
jpayne@7: 
jpayne@7:         if headers is None:
jpayne@7:             headers = {}
jpayne@7:         header_keys = frozenset(to_str(k.lower()) for k in headers)
jpayne@7:         skip_accept_encoding = "accept-encoding" in header_keys
jpayne@7:         skip_host = "host" in header_keys
jpayne@7:         self.putrequest(
jpayne@7:             method, url, skip_accept_encoding=skip_accept_encoding, skip_host=skip_host
jpayne@7:         )
jpayne@7: 
jpayne@7:         # Transform the body into an iterable of sendall()-able chunks
jpayne@7:         # and detect if an explicit Content-Length is doable.
jpayne@7:         chunks_and_cl = body_to_chunks(body, method=method, blocksize=self.blocksize)
jpayne@7:         chunks = chunks_and_cl.chunks
jpayne@7:         content_length = chunks_and_cl.content_length
jpayne@7: 
jpayne@7:         # When chunked is explicit set to 'True' we respect that.
jpayne@7:         if chunked:
jpayne@7:             if "transfer-encoding" not in header_keys:
jpayne@7:                 self.putheader("Transfer-Encoding", "chunked")
jpayne@7:         else:
jpayne@7:             # Detect whether a framing mechanism is already in use. If so
jpayne@7:             # we respect that value, otherwise we pick chunked vs content-length
jpayne@7:             # depending on the type of 'body'.
jpayne@7:             if "content-length" in header_keys:
jpayne@7:                 chunked = False
jpayne@7:             elif "transfer-encoding" in header_keys:
jpayne@7:                 chunked = True
jpayne@7: 
jpayne@7:             # Otherwise we go off the recommendation of 'body_to_chunks()'.
jpayne@7:             else:
jpayne@7:                 chunked = False
jpayne@7:                 if content_length is None:
jpayne@7:                     if chunks is not None:
jpayne@7:                         chunked = True
jpayne@7:                         self.putheader("Transfer-Encoding", "chunked")
jpayne@7:                 else:
jpayne@7:                     self.putheader("Content-Length", str(content_length))
jpayne@7: 
jpayne@7:         # Now that framing headers are out of the way we send all the other headers.
jpayne@7:         if "user-agent" not in header_keys:
jpayne@7:             self.putheader("User-Agent", _get_default_user_agent())
jpayne@7:         for header, value in headers.items():
jpayne@7:             self.putheader(header, value)
jpayne@7:         self.endheaders()
jpayne@7: 
jpayne@7:         # If we're given a body we start sending that in chunks.
jpayne@7:         if chunks is not None:
jpayne@7:             for chunk in chunks:
jpayne@7:                 # Sending empty chunks isn't allowed for TE: chunked
jpayne@7:                 # as it indicates the end of the body.
jpayne@7:                 if not chunk:
jpayne@7:                     continue
jpayne@7:                 if isinstance(chunk, str):
jpayne@7:                     chunk = chunk.encode("utf-8")
jpayne@7:                 if chunked:
jpayne@7:                     self.send(b"%x\r\n%b\r\n" % (len(chunk), chunk))
jpayne@7:                 else:
jpayne@7:                     self.send(chunk)
jpayne@7: 
jpayne@7:         # Regardless of whether we have a body or not, if we're in
jpayne@7:         # chunked mode we want to send an explicit empty chunk.
jpayne@7:         if chunked:
jpayne@7:             self.send(b"0\r\n\r\n")
jpayne@7: 
jpayne@7:     def request_chunked(
jpayne@7:         self,
jpayne@7:         method: str,
jpayne@7:         url: str,
jpayne@7:         body: _TYPE_BODY | None = None,
jpayne@7:         headers: typing.Mapping[str, str] | None = None,
jpayne@7:     ) -> None:
jpayne@7:         """
jpayne@7:         Alternative to the common request method, which sends the
jpayne@7:         body with chunked encoding and not as one block
jpayne@7:         """
jpayne@7:         warnings.warn(
jpayne@7:             "HTTPConnection.request_chunked() is deprecated and will be removed "
jpayne@7:             "in urllib3 v2.1.0. Instead use HTTPConnection.request(..., chunked=True).",
jpayne@7:             category=DeprecationWarning,
jpayne@7:             stacklevel=2,
jpayne@7:         )
jpayne@7:         self.request(method, url, body=body, headers=headers, chunked=True)
jpayne@7: 
jpayne@7:     def getresponse(  # type: ignore[override]
jpayne@7:         self,
jpayne@7:     ) -> HTTPResponse:
jpayne@7:         """
jpayne@7:         Get the response from the server.
jpayne@7: 
jpayne@7:         If the HTTPConnection is in the correct state, returns an instance of HTTPResponse or of whatever object is returned by the response_class variable.
jpayne@7: 
jpayne@7:         If a request has not been sent or if a previous response has not be handled, ResponseNotReady is raised. If the HTTP response indicates that the connection should be closed, then it will be closed before the response is returned. When the connection is closed, the underlying socket is closed.
jpayne@7:         """
jpayne@7:         # Raise the same error as http.client.HTTPConnection
jpayne@7:         if self._response_options is None:
jpayne@7:             raise ResponseNotReady()
jpayne@7: 
jpayne@7:         # Reset this attribute for being used again.
jpayne@7:         resp_options = self._response_options
jpayne@7:         self._response_options = None
jpayne@7: 
jpayne@7:         # Since the connection's timeout value may have been updated
jpayne@7:         # we need to set the timeout on the socket.
jpayne@7:         self.sock.settimeout(self.timeout)
jpayne@7: 
jpayne@7:         # This is needed here to avoid circular import errors
jpayne@7:         from .response import HTTPResponse
jpayne@7: 
jpayne@7:         # Get the response from http.client.HTTPConnection
jpayne@7:         httplib_response = super().getresponse()
jpayne@7: 
jpayne@7:         try:
jpayne@7:             assert_header_parsing(httplib_response.msg)
jpayne@7:         except (HeaderParsingError, TypeError) as hpe:
jpayne@7:             log.warning(
jpayne@7:                 "Failed to parse headers (url=%s): %s",
jpayne@7:                 _url_from_connection(self, resp_options.request_url),
jpayne@7:                 hpe,
jpayne@7:                 exc_info=True,
jpayne@7:             )
jpayne@7: 
jpayne@7:         headers = HTTPHeaderDict(httplib_response.msg.items())
jpayne@7: 
jpayne@7:         response = HTTPResponse(
jpayne@7:             body=httplib_response,
jpayne@7:             headers=headers,
jpayne@7:             status=httplib_response.status,
jpayne@7:             version=httplib_response.version,
jpayne@7:             reason=httplib_response.reason,
jpayne@7:             preload_content=resp_options.preload_content,
jpayne@7:             decode_content=resp_options.decode_content,
jpayne@7:             original_response=httplib_response,
jpayne@7:             enforce_content_length=resp_options.enforce_content_length,
jpayne@7:             request_method=resp_options.request_method,
jpayne@7:             request_url=resp_options.request_url,
jpayne@7:         )
jpayne@7:         return response
jpayne@7: 
jpayne@7: 
jpayne@7: class HTTPSConnection(HTTPConnection):
jpayne@7:     """
jpayne@7:     Many of the parameters to this constructor are passed to the underlying SSL
jpayne@7:     socket by means of :py:func:`urllib3.util.ssl_wrap_socket`.
jpayne@7:     """
jpayne@7: 
jpayne@7:     default_port = port_by_scheme["https"]  # type: ignore[misc]
jpayne@7: 
jpayne@7:     cert_reqs: int | str | None = None
jpayne@7:     ca_certs: str | None = None
jpayne@7:     ca_cert_dir: str | None = None
jpayne@7:     ca_cert_data: None | str | bytes = None
jpayne@7:     ssl_version: int | str | None = None
jpayne@7:     ssl_minimum_version: int | None = None
jpayne@7:     ssl_maximum_version: int | None = None
jpayne@7:     assert_fingerprint: str | None = None
jpayne@7: 
jpayne@7:     def __init__(
jpayne@7:         self,
jpayne@7:         host: str,
jpayne@7:         port: int | None = None,
jpayne@7:         *,
jpayne@7:         timeout: _TYPE_TIMEOUT = _DEFAULT_TIMEOUT,
jpayne@7:         source_address: tuple[str, int] | None = None,
jpayne@7:         blocksize: int = 16384,
jpayne@7:         socket_options: None
jpayne@7:         | (connection._TYPE_SOCKET_OPTIONS) = HTTPConnection.default_socket_options,
jpayne@7:         proxy: Url | None = None,
jpayne@7:         proxy_config: ProxyConfig | None = None,
jpayne@7:         cert_reqs: int | str | None = None,
jpayne@7:         assert_hostname: None | str | Literal[False] = None,
jpayne@7:         assert_fingerprint: str | None = None,
jpayne@7:         server_hostname: str | None = None,
jpayne@7:         ssl_context: ssl.SSLContext | None = None,
jpayne@7:         ca_certs: str | None = None,
jpayne@7:         ca_cert_dir: str | None = None,
jpayne@7:         ca_cert_data: None | str | bytes = None,
jpayne@7:         ssl_minimum_version: int | None = None,
jpayne@7:         ssl_maximum_version: int | None = None,
jpayne@7:         ssl_version: int | str | None = None,  # Deprecated
jpayne@7:         cert_file: str | None = None,
jpayne@7:         key_file: str | None = None,
jpayne@7:         key_password: str | None = None,
jpayne@7:     ) -> None:
jpayne@7:         super().__init__(
jpayne@7:             host,
jpayne@7:             port=port,
jpayne@7:             timeout=timeout,
jpayne@7:             source_address=source_address,
jpayne@7:             blocksize=blocksize,
jpayne@7:             socket_options=socket_options,
jpayne@7:             proxy=proxy,
jpayne@7:             proxy_config=proxy_config,
jpayne@7:         )
jpayne@7: 
jpayne@7:         self.key_file = key_file
jpayne@7:         self.cert_file = cert_file
jpayne@7:         self.key_password = key_password
jpayne@7:         self.ssl_context = ssl_context
jpayne@7:         self.server_hostname = server_hostname
jpayne@7:         self.assert_hostname = assert_hostname
jpayne@7:         self.assert_fingerprint = assert_fingerprint
jpayne@7:         self.ssl_version = ssl_version
jpayne@7:         self.ssl_minimum_version = ssl_minimum_version
jpayne@7:         self.ssl_maximum_version = ssl_maximum_version
jpayne@7:         self.ca_certs = ca_certs and os.path.expanduser(ca_certs)
jpayne@7:         self.ca_cert_dir = ca_cert_dir and os.path.expanduser(ca_cert_dir)
jpayne@7:         self.ca_cert_data = ca_cert_data
jpayne@7: 
jpayne@7:         # cert_reqs depends on ssl_context so calculate last.
jpayne@7:         if cert_reqs is None:
jpayne@7:             if self.ssl_context is not None:
jpayne@7:                 cert_reqs = self.ssl_context.verify_mode
jpayne@7:             else:
jpayne@7:                 cert_reqs = resolve_cert_reqs(None)
jpayne@7:         self.cert_reqs = cert_reqs
jpayne@7: 
jpayne@7:     def set_cert(
jpayne@7:         self,
jpayne@7:         key_file: str | None = None,
jpayne@7:         cert_file: str | None = None,
jpayne@7:         cert_reqs: int | str | None = None,
jpayne@7:         key_password: str | None = None,
jpayne@7:         ca_certs: str | None = None,
jpayne@7:         assert_hostname: None | str | Literal[False] = None,
jpayne@7:         assert_fingerprint: str | None = None,
jpayne@7:         ca_cert_dir: str | None = None,
jpayne@7:         ca_cert_data: None | str | bytes = None,
jpayne@7:     ) -> None:
jpayne@7:         """
jpayne@7:         This method should only be called once, before the connection is used.
jpayne@7:         """
jpayne@7:         warnings.warn(
jpayne@7:             "HTTPSConnection.set_cert() is deprecated and will be removed "
jpayne@7:             "in urllib3 v2.1.0. Instead provide the parameters to the "
jpayne@7:             "HTTPSConnection constructor.",
jpayne@7:             category=DeprecationWarning,
jpayne@7:             stacklevel=2,
jpayne@7:         )
jpayne@7: 
jpayne@7:         # If cert_reqs is not provided we'll assume CERT_REQUIRED unless we also
jpayne@7:         # have an SSLContext object in which case we'll use its verify_mode.
jpayne@7:         if cert_reqs is None:
jpayne@7:             if self.ssl_context is not None:
jpayne@7:                 cert_reqs = self.ssl_context.verify_mode
jpayne@7:             else:
jpayne@7:                 cert_reqs = resolve_cert_reqs(None)
jpayne@7: 
jpayne@7:         self.key_file = key_file
jpayne@7:         self.cert_file = cert_file
jpayne@7:         self.cert_reqs = cert_reqs
jpayne@7:         self.key_password = key_password
jpayne@7:         self.assert_hostname = assert_hostname
jpayne@7:         self.assert_fingerprint = assert_fingerprint
jpayne@7:         self.ca_certs = ca_certs and os.path.expanduser(ca_certs)
jpayne@7:         self.ca_cert_dir = ca_cert_dir and os.path.expanduser(ca_cert_dir)
jpayne@7:         self.ca_cert_data = ca_cert_data
jpayne@7: 
jpayne@7:     def connect(self) -> None:
jpayne@7:         sock: socket.socket | ssl.SSLSocket
jpayne@7:         self.sock = sock = self._new_conn()
jpayne@7:         server_hostname: str = self.host
jpayne@7:         tls_in_tls = False
jpayne@7: 
jpayne@7:         # Do we need to establish a tunnel?
jpayne@7:         if self._tunnel_host is not None:
jpayne@7:             # We're tunneling to an HTTPS origin so need to do TLS-in-TLS.
jpayne@7:             if self._tunnel_scheme == "https":
jpayne@7:                 # _connect_tls_proxy will verify and assign proxy_is_verified
jpayne@7:                 self.sock = sock = self._connect_tls_proxy(self.host, sock)
jpayne@7:                 tls_in_tls = True
jpayne@7:             elif self._tunnel_scheme == "http":
jpayne@7:                 self.proxy_is_verified = False
jpayne@7: 
jpayne@7:             # If we're tunneling it means we're connected to our proxy.
jpayne@7:             self._has_connected_to_proxy = True
jpayne@7: 
jpayne@7:             self._tunnel()  # type: ignore[attr-defined]
jpayne@7:             # Override the host with the one we're requesting data from.
jpayne@7:             server_hostname = self._tunnel_host
jpayne@7: 
jpayne@7:         if self.server_hostname is not None:
jpayne@7:             server_hostname = self.server_hostname
jpayne@7: 
jpayne@7:         is_time_off = datetime.date.today() < RECENT_DATE
jpayne@7:         if is_time_off:
jpayne@7:             warnings.warn(
jpayne@7:                 (
jpayne@7:                     f"System time is way off (before {RECENT_DATE}). This will probably "
jpayne@7:                     "lead to SSL verification errors"
jpayne@7:                 ),
jpayne@7:                 SystemTimeWarning,
jpayne@7:             )
jpayne@7: 
jpayne@7:         # Remove trailing '.' from fqdn hostnames to allow certificate validation
jpayne@7:         server_hostname_rm_dot = server_hostname.rstrip(".")
jpayne@7: 
jpayne@7:         sock_and_verified = _ssl_wrap_socket_and_match_hostname(
jpayne@7:             sock=sock,
jpayne@7:             cert_reqs=self.cert_reqs,
jpayne@7:             ssl_version=self.ssl_version,
jpayne@7:             ssl_minimum_version=self.ssl_minimum_version,
jpayne@7:             ssl_maximum_version=self.ssl_maximum_version,
jpayne@7:             ca_certs=self.ca_certs,
jpayne@7:             ca_cert_dir=self.ca_cert_dir,
jpayne@7:             ca_cert_data=self.ca_cert_data,
jpayne@7:             cert_file=self.cert_file,
jpayne@7:             key_file=self.key_file,
jpayne@7:             key_password=self.key_password,
jpayne@7:             server_hostname=server_hostname_rm_dot,
jpayne@7:             ssl_context=self.ssl_context,
jpayne@7:             tls_in_tls=tls_in_tls,
jpayne@7:             assert_hostname=self.assert_hostname,
jpayne@7:             assert_fingerprint=self.assert_fingerprint,
jpayne@7:         )
jpayne@7:         self.sock = sock_and_verified.socket
jpayne@7: 
jpayne@7:         # Forwarding proxies can never have a verified target since
jpayne@7:         # the proxy is the one doing the verification. Should instead
jpayne@7:         # use a CONNECT tunnel in order to verify the target.
jpayne@7:         # See: https://github.com/urllib3/urllib3/issues/3267.
jpayne@7:         if self.proxy_is_forwarding:
jpayne@7:             self.is_verified = False
jpayne@7:         else:
jpayne@7:             self.is_verified = sock_and_verified.is_verified
jpayne@7: 
jpayne@7:         # If there's a proxy to be connected to we are fully connected.
jpayne@7:         # This is set twice (once above and here) due to forwarding proxies
jpayne@7:         # not using tunnelling.
jpayne@7:         self._has_connected_to_proxy = bool(self.proxy)
jpayne@7: 
jpayne@7:         # Set `self.proxy_is_verified` unless it's already set while
jpayne@7:         # establishing a tunnel.
jpayne@7:         if self._has_connected_to_proxy and self.proxy_is_verified is None:
jpayne@7:             self.proxy_is_verified = sock_and_verified.is_verified
jpayne@7: 
jpayne@7:     def _connect_tls_proxy(self, hostname: str, sock: socket.socket) -> ssl.SSLSocket:
jpayne@7:         """
jpayne@7:         Establish a TLS connection to the proxy using the provided SSL context.
jpayne@7:         """
jpayne@7:         # `_connect_tls_proxy` is called when self._tunnel_host is truthy.
jpayne@7:         proxy_config = typing.cast(ProxyConfig, self.proxy_config)
jpayne@7:         ssl_context = proxy_config.ssl_context
jpayne@7:         sock_and_verified = _ssl_wrap_socket_and_match_hostname(
jpayne@7:             sock,
jpayne@7:             cert_reqs=self.cert_reqs,
jpayne@7:             ssl_version=self.ssl_version,
jpayne@7:             ssl_minimum_version=self.ssl_minimum_version,
jpayne@7:             ssl_maximum_version=self.ssl_maximum_version,
jpayne@7:             ca_certs=self.ca_certs,
jpayne@7:             ca_cert_dir=self.ca_cert_dir,
jpayne@7:             ca_cert_data=self.ca_cert_data,
jpayne@7:             server_hostname=hostname,
jpayne@7:             ssl_context=ssl_context,
jpayne@7:             assert_hostname=proxy_config.assert_hostname,
jpayne@7:             assert_fingerprint=proxy_config.assert_fingerprint,
jpayne@7:             # Features that aren't implemented for proxies yet:
jpayne@7:             cert_file=None,
jpayne@7:             key_file=None,
jpayne@7:             key_password=None,
jpayne@7:             tls_in_tls=False,
jpayne@7:         )
jpayne@7:         self.proxy_is_verified = sock_and_verified.is_verified
jpayne@7:         return sock_and_verified.socket  # type: ignore[return-value]
jpayne@7: 
jpayne@7: 
jpayne@7: class _WrappedAndVerifiedSocket(typing.NamedTuple):
jpayne@7:     """
jpayne@7:     Wrapped socket and whether the connection is
jpayne@7:     verified after the TLS handshake
jpayne@7:     """
jpayne@7: 
jpayne@7:     socket: ssl.SSLSocket | SSLTransport
jpayne@7:     is_verified: bool
jpayne@7: 
jpayne@7: 
jpayne@7: def _ssl_wrap_socket_and_match_hostname(
jpayne@7:     sock: socket.socket,
jpayne@7:     *,
jpayne@7:     cert_reqs: None | str | int,
jpayne@7:     ssl_version: None | str | int,
jpayne@7:     ssl_minimum_version: int | None,
jpayne@7:     ssl_maximum_version: int | None,
jpayne@7:     cert_file: str | None,
jpayne@7:     key_file: str | None,
jpayne@7:     key_password: str | None,
jpayne@7:     ca_certs: str | None,
jpayne@7:     ca_cert_dir: str | None,
jpayne@7:     ca_cert_data: None | str | bytes,
jpayne@7:     assert_hostname: None | str | Literal[False],
jpayne@7:     assert_fingerprint: str | None,
jpayne@7:     server_hostname: str | None,
jpayne@7:     ssl_context: ssl.SSLContext | None,
jpayne@7:     tls_in_tls: bool = False,
jpayne@7: ) -> _WrappedAndVerifiedSocket:
jpayne@7:     """Logic for constructing an SSLContext from all TLS parameters, passing
jpayne@7:     that down into ssl_wrap_socket, and then doing certificate verification
jpayne@7:     either via hostname or fingerprint. This function exists to guarantee
jpayne@7:     that both proxies and targets have the same behavior when connecting via TLS.
jpayne@7:     """
jpayne@7:     default_ssl_context = False
jpayne@7:     if ssl_context is None:
jpayne@7:         default_ssl_context = True
jpayne@7:         context = create_urllib3_context(
jpayne@7:             ssl_version=resolve_ssl_version(ssl_version),
jpayne@7:             ssl_minimum_version=ssl_minimum_version,
jpayne@7:             ssl_maximum_version=ssl_maximum_version,
jpayne@7:             cert_reqs=resolve_cert_reqs(cert_reqs),
jpayne@7:         )
jpayne@7:     else:
jpayne@7:         context = ssl_context
jpayne@7: 
jpayne@7:     context.verify_mode = resolve_cert_reqs(cert_reqs)
jpayne@7: 
jpayne@7:     # In some cases, we want to verify hostnames ourselves
jpayne@7:     if (
jpayne@7:         # `ssl` can't verify fingerprints or alternate hostnames
jpayne@7:         assert_fingerprint
jpayne@7:         or assert_hostname
jpayne@7:         # assert_hostname can be set to False to disable hostname checking
jpayne@7:         or assert_hostname is False
jpayne@7:         # We still support OpenSSL 1.0.2, which prevents us from verifying
jpayne@7:         # hostnames easily: https://github.com/pyca/pyopenssl/pull/933
jpayne@7:         or ssl_.IS_PYOPENSSL
jpayne@7:         or not ssl_.HAS_NEVER_CHECK_COMMON_NAME
jpayne@7:     ):
jpayne@7:         context.check_hostname = False
jpayne@7: 
jpayne@7:     # Try to load OS default certs if none are given. We need to do the hasattr() check
jpayne@7:     # for custom pyOpenSSL SSLContext objects because they don't support
jpayne@7:     # load_default_certs().
jpayne@7:     if (
jpayne@7:         not ca_certs
jpayne@7:         and not ca_cert_dir
jpayne@7:         and not ca_cert_data
jpayne@7:         and default_ssl_context
jpayne@7:         and hasattr(context, "load_default_certs")
jpayne@7:     ):
jpayne@7:         context.load_default_certs()
jpayne@7: 
jpayne@7:     # Ensure that IPv6 addresses are in the proper format and don't have a
jpayne@7:     # scope ID. Python's SSL module fails to recognize scoped IPv6 addresses
jpayne@7:     # and interprets them as DNS hostnames.
jpayne@7:     if server_hostname is not None:
jpayne@7:         normalized = server_hostname.strip("[]")
jpayne@7:         if "%" in normalized:
jpayne@7:             normalized = normalized[: normalized.rfind("%")]
jpayne@7:         if is_ipaddress(normalized):
jpayne@7:             server_hostname = normalized
jpayne@7: 
jpayne@7:     ssl_sock = ssl_wrap_socket(
jpayne@7:         sock=sock,
jpayne@7:         keyfile=key_file,
jpayne@7:         certfile=cert_file,
jpayne@7:         key_password=key_password,
jpayne@7:         ca_certs=ca_certs,
jpayne@7:         ca_cert_dir=ca_cert_dir,
jpayne@7:         ca_cert_data=ca_cert_data,
jpayne@7:         server_hostname=server_hostname,
jpayne@7:         ssl_context=context,
jpayne@7:         tls_in_tls=tls_in_tls,
jpayne@7:     )
jpayne@7: 
jpayne@7:     try:
jpayne@7:         if assert_fingerprint:
jpayne@7:             _assert_fingerprint(
jpayne@7:                 ssl_sock.getpeercert(binary_form=True), assert_fingerprint
jpayne@7:             )
jpayne@7:         elif (
jpayne@7:             context.verify_mode != ssl.CERT_NONE
jpayne@7:             and not context.check_hostname
jpayne@7:             and assert_hostname is not False
jpayne@7:         ):
jpayne@7:             cert: _TYPE_PEER_CERT_RET_DICT = ssl_sock.getpeercert()  # type: ignore[assignment]
jpayne@7: 
jpayne@7:             # Need to signal to our match_hostname whether to use 'commonName' or not.
jpayne@7:             # If we're using our own constructed SSLContext we explicitly set 'False'
jpayne@7:             # because PyPy hard-codes 'True' from SSLContext.hostname_checks_common_name.
jpayne@7:             if default_ssl_context:
jpayne@7:                 hostname_checks_common_name = False
jpayne@7:             else:
jpayne@7:                 hostname_checks_common_name = (
jpayne@7:                     getattr(context, "hostname_checks_common_name", False) or False
jpayne@7:                 )
jpayne@7: 
jpayne@7:             _match_hostname(
jpayne@7:                 cert,
jpayne@7:                 assert_hostname or server_hostname,  # type: ignore[arg-type]
jpayne@7:                 hostname_checks_common_name,
jpayne@7:             )
jpayne@7: 
jpayne@7:         return _WrappedAndVerifiedSocket(
jpayne@7:             socket=ssl_sock,
jpayne@7:             is_verified=context.verify_mode == ssl.CERT_REQUIRED
jpayne@7:             or bool(assert_fingerprint),
jpayne@7:         )
jpayne@7:     except BaseException:
jpayne@7:         ssl_sock.close()
jpayne@7:         raise
jpayne@7: 
jpayne@7: 
jpayne@7: def _match_hostname(
jpayne@7:     cert: _TYPE_PEER_CERT_RET_DICT | None,
jpayne@7:     asserted_hostname: str,
jpayne@7:     hostname_checks_common_name: bool = False,
jpayne@7: ) -> None:
jpayne@7:     # Our upstream implementation of ssl.match_hostname()
jpayne@7:     # only applies this normalization to IP addresses so it doesn't
jpayne@7:     # match DNS SANs so we do the same thing!
jpayne@7:     stripped_hostname = asserted_hostname.strip("[]")
jpayne@7:     if is_ipaddress(stripped_hostname):
jpayne@7:         asserted_hostname = stripped_hostname
jpayne@7: 
jpayne@7:     try:
jpayne@7:         match_hostname(cert, asserted_hostname, hostname_checks_common_name)
jpayne@7:     except CertificateError as e:
jpayne@7:         log.warning(
jpayne@7:             "Certificate did not match expected hostname: %s. Certificate: %s",
jpayne@7:             asserted_hostname,
jpayne@7:             cert,
jpayne@7:         )
jpayne@7:         # Add cert to exception and reraise so client code can inspect
jpayne@7:         # the cert when catching the exception, if they want to
jpayne@7:         e._peer_cert = cert  # type: ignore[attr-defined]
jpayne@7:         raise
jpayne@7: 
jpayne@7: 
jpayne@7: def _wrap_proxy_error(err: Exception, proxy_scheme: str | None) -> ProxyError:
jpayne@7:     # Look for the phrase 'wrong version number', if found
jpayne@7:     # then we should warn the user that we're very sure that
jpayne@7:     # this proxy is HTTP-only and they have a configuration issue.
jpayne@7:     error_normalized = " ".join(re.split("[^a-z]", str(err).lower()))
jpayne@7:     is_likely_http_proxy = (
jpayne@7:         "wrong version number" in error_normalized
jpayne@7:         or "unknown protocol" in error_normalized
jpayne@7:         or "record layer failure" in error_normalized
jpayne@7:     )
jpayne@7:     http_proxy_warning = (
jpayne@7:         ". Your proxy appears to only use HTTP and not HTTPS, "
jpayne@7:         "try changing your proxy URL to be HTTP. See: "
jpayne@7:         "https://urllib3.readthedocs.io/en/latest/advanced-usage.html"
jpayne@7:         "#https-proxy-error-http-proxy"
jpayne@7:     )
jpayne@7:     new_err = ProxyError(
jpayne@7:         f"Unable to connect to proxy"
jpayne@7:         f"{http_proxy_warning if is_likely_http_proxy and proxy_scheme == 'https' else ''}",
jpayne@7:         err,
jpayne@7:     )
jpayne@7:     new_err.__cause__ = err
jpayne@7:     return new_err
jpayne@7: 
jpayne@7: 
jpayne@7: def _get_default_user_agent() -> str:
jpayne@7:     return f"python-urllib3/{__version__}"
jpayne@7: 
jpayne@7: 
jpayne@7: class DummyConnection:
jpayne@7:     """Used to detect a failed ConnectionCls import."""
jpayne@7: 
jpayne@7: 
jpayne@7: if not ssl:
jpayne@7:     HTTPSConnection = DummyConnection  # type: ignore[misc, assignment] # noqa: F811
jpayne@7: 
jpayne@7: 
jpayne@7: VerifiedHTTPSConnection = HTTPSConnection
jpayne@7: 
jpayne@7: 
jpayne@7: def _url_from_connection(
jpayne@7:     conn: HTTPConnection | HTTPSConnection, path: str | None = None
jpayne@7: ) -> str:
jpayne@7:     """Returns the URL from a given connection. This is mainly used for testing and logging."""
jpayne@7: 
jpayne@7:     scheme = "https" if isinstance(conn, HTTPSConnection) else "http"
jpayne@7: 
jpayne@7:     return Url(scheme=scheme, host=conn.host, port=conn.port, path=path).url