diff options
Diffstat (limited to '.venv/lib/python3.12/site-packages/urllib3/http2')
3 files changed, 496 insertions, 0 deletions
diff --git a/.venv/lib/python3.12/site-packages/urllib3/http2/__init__.py b/.venv/lib/python3.12/site-packages/urllib3/http2/__init__.py new file mode 100644 index 00000000..133e1d8f --- /dev/null +++ b/.venv/lib/python3.12/site-packages/urllib3/http2/__init__.py @@ -0,0 +1,53 @@ +from __future__ import annotations + +from importlib.metadata import version + +__all__ = [ + "inject_into_urllib3", + "extract_from_urllib3", +] + +import typing + +orig_HTTPSConnection: typing.Any = None + + +def inject_into_urllib3() -> None: + # First check if h2 version is valid + h2_version = version("h2") + if not h2_version.startswith("4."): + raise ImportError( + "urllib3 v2 supports h2 version 4.x.x, currently " + f"the 'h2' module is compiled with {h2_version!r}. " + "See: https://github.com/urllib3/urllib3/issues/3290" + ) + + # Import here to avoid circular dependencies. + from .. import connection as urllib3_connection + from .. import util as urllib3_util + from ..connectionpool import HTTPSConnectionPool + from ..util import ssl_ as urllib3_util_ssl + from .connection import HTTP2Connection + + global orig_HTTPSConnection + orig_HTTPSConnection = urllib3_connection.HTTPSConnection + + HTTPSConnectionPool.ConnectionCls = HTTP2Connection + urllib3_connection.HTTPSConnection = HTTP2Connection # type: ignore[misc] + + # TODO: Offer 'http/1.1' as well, but for testing purposes this is handy. + urllib3_util.ALPN_PROTOCOLS = ["h2"] + urllib3_util_ssl.ALPN_PROTOCOLS = ["h2"] + + +def extract_from_urllib3() -> None: + from .. import connection as urllib3_connection + from .. import util as urllib3_util + from ..connectionpool import HTTPSConnectionPool + from ..util import ssl_ as urllib3_util_ssl + + HTTPSConnectionPool.ConnectionCls = orig_HTTPSConnection + urllib3_connection.HTTPSConnection = orig_HTTPSConnection # type: ignore[misc] + + urllib3_util.ALPN_PROTOCOLS = ["http/1.1"] + urllib3_util_ssl.ALPN_PROTOCOLS = ["http/1.1"] diff --git a/.venv/lib/python3.12/site-packages/urllib3/http2/connection.py b/.venv/lib/python3.12/site-packages/urllib3/http2/connection.py new file mode 100644 index 00000000..f4861452 --- /dev/null +++ b/.venv/lib/python3.12/site-packages/urllib3/http2/connection.py @@ -0,0 +1,356 @@ +from __future__ import annotations + +import logging +import re +import threading +import types +import typing + +import h2.config # type: ignore[import-untyped] +import h2.connection # type: ignore[import-untyped] +import h2.events # type: ignore[import-untyped] + +from .._base_connection import _TYPE_BODY +from .._collections import HTTPHeaderDict +from ..connection import HTTPSConnection, _get_default_user_agent +from ..exceptions import ConnectionError +from ..response import BaseHTTPResponse + +orig_HTTPSConnection = HTTPSConnection + +T = typing.TypeVar("T") + +log = logging.getLogger(__name__) + +RE_IS_LEGAL_HEADER_NAME = re.compile(rb"^[!#$%&'*+\-.^_`|~0-9a-z]+$") +RE_IS_ILLEGAL_HEADER_VALUE = re.compile(rb"[\0\x00\x0a\x0d\r\n]|^[ \r\n\t]|[ \r\n\t]$") + + +def _is_legal_header_name(name: bytes) -> bool: + """ + "An implementation that validates fields according to the definitions in Sections + 5.1 and 5.5 of [HTTP] only needs an additional check that field names do not + include uppercase characters." (https://httpwg.org/specs/rfc9113.html#n-field-validity) + + `http.client._is_legal_header_name` does not validate the field name according to the + HTTP 1.1 spec, so we do that here, in addition to checking for uppercase characters. + + This does not allow for the `:` character in the header name, so should not + be used to validate pseudo-headers. + """ + return bool(RE_IS_LEGAL_HEADER_NAME.match(name)) + + +def _is_illegal_header_value(value: bytes) -> bool: + """ + "A field value MUST NOT contain the zero value (ASCII NUL, 0x00), line feed + (ASCII LF, 0x0a), or carriage return (ASCII CR, 0x0d) at any position. A field + value MUST NOT start or end with an ASCII whitespace character (ASCII SP or HTAB, + 0x20 or 0x09)." (https://httpwg.org/specs/rfc9113.html#n-field-validity) + """ + return bool(RE_IS_ILLEGAL_HEADER_VALUE.search(value)) + + +class _LockedObject(typing.Generic[T]): + """ + A wrapper class that hides a specific object behind a lock. + The goal here is to provide a simple way to protect access to an object + that cannot safely be simultaneously accessed from multiple threads. The + intended use of this class is simple: take hold of it with a context + manager, which returns the protected object. + """ + + __slots__ = ( + "lock", + "_obj", + ) + + def __init__(self, obj: T): + self.lock = threading.RLock() + self._obj = obj + + def __enter__(self) -> T: + self.lock.acquire() + return self._obj + + def __exit__( + self, + exc_type: type[BaseException] | None, + exc_val: BaseException | None, + exc_tb: types.TracebackType | None, + ) -> None: + self.lock.release() + + +class HTTP2Connection(HTTPSConnection): + def __init__( + self, host: str, port: int | None = None, **kwargs: typing.Any + ) -> None: + self._h2_conn = self._new_h2_conn() + self._h2_stream: int | None = None + self._headers: list[tuple[bytes, bytes]] = [] + + if "proxy" in kwargs or "proxy_config" in kwargs: # Defensive: + raise NotImplementedError("Proxies aren't supported with HTTP/2") + + super().__init__(host, port, **kwargs) + + if self._tunnel_host is not None: + raise NotImplementedError("Tunneling isn't supported with HTTP/2") + + def _new_h2_conn(self) -> _LockedObject[h2.connection.H2Connection]: + config = h2.config.H2Configuration(client_side=True) + return _LockedObject(h2.connection.H2Connection(config=config)) + + def connect(self) -> None: + super().connect() + with self._h2_conn as conn: + conn.initiate_connection() + if data_to_send := conn.data_to_send(): + self.sock.sendall(data_to_send) + + def putrequest( # type: ignore[override] + self, + method: str, + url: str, + **kwargs: typing.Any, + ) -> None: + """putrequest + This deviates from the HTTPConnection method signature since we never need to override + sending accept-encoding headers or the host header. + """ + if "skip_host" in kwargs: + raise NotImplementedError("`skip_host` isn't supported") + if "skip_accept_encoding" in kwargs: + raise NotImplementedError("`skip_accept_encoding` isn't supported") + + self._request_url = url or "/" + self._validate_path(url) # type: ignore[attr-defined] + + if ":" in self.host: + authority = f"[{self.host}]:{self.port or 443}" + else: + authority = f"{self.host}:{self.port or 443}" + + self._headers.append((b":scheme", b"https")) + self._headers.append((b":method", method.encode())) + self._headers.append((b":authority", authority.encode())) + self._headers.append((b":path", url.encode())) + + with self._h2_conn as conn: + self._h2_stream = conn.get_next_available_stream_id() + + def putheader(self, header: str | bytes, *values: str | bytes) -> None: + # TODO SKIPPABLE_HEADERS from urllib3 are ignored. + header = header.encode() if isinstance(header, str) else header + header = header.lower() # A lot of upstream code uses capitalized headers. + if not _is_legal_header_name(header): + raise ValueError(f"Illegal header name {str(header)}") + + for value in values: + value = value.encode() if isinstance(value, str) else value + if _is_illegal_header_value(value): + raise ValueError(f"Illegal header value {str(value)}") + self._headers.append((header, value)) + + def endheaders(self, message_body: typing.Any = None) -> None: # type: ignore[override] + if self._h2_stream is None: + raise ConnectionError("Must call `putrequest` first.") + + with self._h2_conn as conn: + conn.send_headers( + stream_id=self._h2_stream, + headers=self._headers, + end_stream=(message_body is None), + ) + if data_to_send := conn.data_to_send(): + self.sock.sendall(data_to_send) + self._headers = [] # Reset headers for the next request. + + def send(self, data: typing.Any) -> None: + """Send data to the server. + `data` can be: `str`, `bytes`, an iterable, or file-like objects + that support a .read() method. + """ + if self._h2_stream is None: + raise ConnectionError("Must call `putrequest` first.") + + with self._h2_conn as conn: + if data_to_send := conn.data_to_send(): + self.sock.sendall(data_to_send) + + if hasattr(data, "read"): # file-like objects + while True: + chunk = data.read(self.blocksize) + if not chunk: + break + if isinstance(chunk, str): + chunk = chunk.encode() # pragma: no cover + conn.send_data(self._h2_stream, chunk, end_stream=False) + if data_to_send := conn.data_to_send(): + self.sock.sendall(data_to_send) + conn.end_stream(self._h2_stream) + return + + if isinstance(data, str): # str -> bytes + data = data.encode() + + try: + if isinstance(data, bytes): + conn.send_data(self._h2_stream, data, end_stream=True) + if data_to_send := conn.data_to_send(): + self.sock.sendall(data_to_send) + else: + for chunk in data: + conn.send_data(self._h2_stream, chunk, end_stream=False) + if data_to_send := conn.data_to_send(): + self.sock.sendall(data_to_send) + conn.end_stream(self._h2_stream) + except TypeError: + raise TypeError( + "`data` should be str, bytes, iterable, or file. got %r" + % type(data) + ) + + def set_tunnel( + self, + host: str, + port: int | None = None, + headers: typing.Mapping[str, str] | None = None, + scheme: str = "http", + ) -> None: + raise NotImplementedError( + "HTTP/2 does not support setting up a tunnel through a proxy" + ) + + def getresponse( # type: ignore[override] + self, + ) -> HTTP2Response: + status = None + data = bytearray() + with self._h2_conn as conn: + end_stream = False + while not end_stream: + # TODO: Arbitrary read value. + if received_data := self.sock.recv(65535): + events = conn.receive_data(received_data) + for event in events: + if isinstance(event, h2.events.ResponseReceived): + headers = HTTPHeaderDict() + for header, value in event.headers: + if header == b":status": + status = int(value.decode()) + else: + headers.add( + header.decode("ascii"), value.decode("ascii") + ) + + elif isinstance(event, h2.events.DataReceived): + data += event.data + conn.acknowledge_received_data( + event.flow_controlled_length, event.stream_id + ) + + elif isinstance(event, h2.events.StreamEnded): + end_stream = True + + if data_to_send := conn.data_to_send(): + self.sock.sendall(data_to_send) + + assert status is not None + return HTTP2Response( + status=status, + headers=headers, + request_url=self._request_url, + data=bytes(data), + ) + + def request( # type: ignore[override] + self, + method: str, + url: str, + body: _TYPE_BODY | None = None, + headers: typing.Mapping[str, str] | None = None, + *, + preload_content: bool = True, + decode_content: bool = True, + enforce_content_length: bool = True, + **kwargs: typing.Any, + ) -> None: + """Send an HTTP/2 request""" + if "chunked" in kwargs: + # TODO this is often present from upstream. + # raise NotImplementedError("`chunked` isn't supported with HTTP/2") + pass + + if self.sock is not None: + self.sock.settimeout(self.timeout) + + self.putrequest(method, url) + + headers = headers or {} + for k, v in headers.items(): + if k.lower() == "transfer-encoding" and v == "chunked": + continue + else: + self.putheader(k, v) + + if b"user-agent" not in dict(self._headers): + self.putheader(b"user-agent", _get_default_user_agent()) + + if body: + self.endheaders(message_body=body) + self.send(body) + else: + self.endheaders() + + def close(self) -> None: + with self._h2_conn as conn: + try: + conn.close_connection() + if data := conn.data_to_send(): + self.sock.sendall(data) + except Exception: + pass + + # Reset all our HTTP/2 connection state. + self._h2_conn = self._new_h2_conn() + self._h2_stream = None + self._headers = [] + + super().close() + + +class HTTP2Response(BaseHTTPResponse): + # TODO: This is a woefully incomplete response object, but works for non-streaming. + def __init__( + self, + status: int, + headers: HTTPHeaderDict, + request_url: str, + data: bytes, + decode_content: bool = False, # TODO: support decoding + ) -> None: + super().__init__( + status=status, + headers=headers, + # Following CPython, we map HTTP versions to major * 10 + minor integers + version=20, + version_string="HTTP/2", + # No reason phrase in HTTP/2 + reason=None, + decode_content=decode_content, + request_url=request_url, + ) + self._data = data + self.length_remaining = 0 + + @property + def data(self) -> bytes: + return self._data + + def get_redirect_location(self) -> None: + return None + + def close(self) -> None: + pass diff --git a/.venv/lib/python3.12/site-packages/urllib3/http2/probe.py b/.venv/lib/python3.12/site-packages/urllib3/http2/probe.py new file mode 100644 index 00000000..9ea90076 --- /dev/null +++ b/.venv/lib/python3.12/site-packages/urllib3/http2/probe.py @@ -0,0 +1,87 @@ +from __future__ import annotations + +import threading + + +class _HTTP2ProbeCache: + __slots__ = ( + "_lock", + "_cache_locks", + "_cache_values", + ) + + def __init__(self) -> None: + self._lock = threading.Lock() + self._cache_locks: dict[tuple[str, int], threading.RLock] = {} + self._cache_values: dict[tuple[str, int], bool | None] = {} + + def acquire_and_get(self, host: str, port: int) -> bool | None: + # By the end of this block we know that + # _cache_[values,locks] is available. + value = None + with self._lock: + key = (host, port) + try: + value = self._cache_values[key] + # If it's a known value we return right away. + if value is not None: + return value + except KeyError: + self._cache_locks[key] = threading.RLock() + self._cache_values[key] = None + + # If the value is unknown, we acquire the lock to signal + # to the requesting thread that the probe is in progress + # or that the current thread needs to return their findings. + key_lock = self._cache_locks[key] + key_lock.acquire() + try: + # If the by the time we get the lock the value has been + # updated we want to return the updated value. + value = self._cache_values[key] + + # In case an exception like KeyboardInterrupt is raised here. + except BaseException as e: # Defensive: + assert not isinstance(e, KeyError) # KeyError shouldn't be possible. + key_lock.release() + raise + + return value + + def set_and_release( + self, host: str, port: int, supports_http2: bool | None + ) -> None: + key = (host, port) + key_lock = self._cache_locks[key] + with key_lock: # Uses an RLock, so can be locked again from same thread. + if supports_http2 is None and self._cache_values[key] is not None: + raise ValueError( + "Cannot reset HTTP/2 support for origin after value has been set." + ) # Defensive: not expected in normal usage + + self._cache_values[key] = supports_http2 + key_lock.release() + + def _values(self) -> dict[tuple[str, int], bool | None]: + """This function is for testing purposes only. Gets the current state of the probe cache""" + with self._lock: + return {k: v for k, v in self._cache_values.items()} + + def _reset(self) -> None: + """This function is for testing purposes only. Reset the cache values""" + with self._lock: + self._cache_locks = {} + self._cache_values = {} + + +_HTTP2_PROBE_CACHE = _HTTP2ProbeCache() + +set_and_release = _HTTP2_PROBE_CACHE.set_and_release +acquire_and_get = _HTTP2_PROBE_CACHE.acquire_and_get +_values = _HTTP2_PROBE_CACHE._values +_reset = _HTTP2_PROBE_CACHE._reset + +__all__ = [ + "set_and_release", + "acquire_and_get", +] |