관리-도구
편집 파일: serialize.py
# SPDX-FileCopyrightText: 2015 Eric Larson # # SPDX-License-Identifier: Apache-2.0 from __future__ import annotations import io from typing import IO, TYPE_CHECKING, Any, Mapping, cast from pip._vendor import msgpack from pip._vendor.requests.structures import CaseInsensitiveDict from pip._vendor.urllib3 import HTTPResponse if TYPE_CHECKING: from pip._vendor.requests import PreparedRequest class Serializer: serde_version = "4" def dumps( self, request: PreparedRequest, response: HTTPResponse, body: bytes | None = None, ) -> bytes: response_headers: CaseInsensitiveDict[str] = CaseInsensitiveDict( response.headers ) if body is None: # When a body isn't passed in, we'll read the response. We # also update the response with a new file handler to be # sure it acts as though it was never read. body = response.read(decode_content=False) response._fp = io.BytesIO(body) # type: ignore[attr-defined] response.length_remaining = len(body) data = { "response": { "body": body, # Empty bytestring if body is stored separately "headers": {str(k): str(v) for k, v in response.headers.items()}, # type: ignore[no-untyped-call] "status": response.status, "version": response.version, "reason": str(response.reason), "decode_content": response.decode_content, } } # Construct our vary headers data["vary"] = {} if "vary" in response_headers: varied_headers = response_headers["vary"].split(",") for header in varied_headers: header = str(header).strip() header_value = request.headers.get(header, None) if header_value is not None: header_value = str(header_value) data["vary"][header] = header_value return b",".join([f"cc={self.serde_version}".encode(), self.serialize(data)]) def serialize(self, data: dict[str, Any]) -> bytes: return cast(bytes, msgpack.dumps(data, use_bin_type=True)) def loads( self, request: PreparedRequest, data: bytes, body_file: IO[bytes] | None = None, ) -> HTTPResponse | None: # Short circuit if we've been given an empty set of data if not data: return None # Determine what version of the serializer the data was serialized # with try: ver, data = data.split(b",", 1) except ValueError: ver = b"cc=0" # Make sure that our "ver" is actually a version and isn't a false # positive from a , being in the data stream. if ver[:3] != b"cc=": data = ver + data ver = b"cc=0" # Get the version number out of the cc=N verstr = ver.split(b"=", 1)[-1].decode("ascii") # Dispatch to the actual load method for the given version try: return getattr(self, f"_loads_v{verstr}")(request, data, body_file) # type: ignore[no-any-return] except AttributeError: # This is a version we don't have a loads function for, so we'll # just treat it as a miss and return None return None def prepare_response( self, request: PreparedRequest, cached: Mapping[str, Any], body_file: IO[bytes] | None = None, ) -> HTTPResponse | None: """Verify our vary headers match and construct a real urllib3 HTTPResponse object. """ # Special case the '*' Vary value as it means we cannot actually # determine if the cached response is suitable for this request. # This case is also handled in the controller code when creating # a cache entry, but is left here for backwards compatibility. if "*" in cached.get("vary", {}): return None # Ensure that the Vary headers for the cached response match our # request for header, value in cached.get("vary", {}).items(): if request.headers.get(header, None) != value: return None body_raw = cached["response"].pop("body") headers: CaseInsensitiveDict[str] = CaseInsensitiveDict( data=cached["response"]["headers"] ) if headers.get("transfer-encoding", "") == "chunked": headers.pop("transfer-encoding") cached["response"]["headers"] = headers try: body: IO[bytes] if body_file is None: body = io.BytesIO(body_raw) else: body = body_file except TypeError: # This can happen if cachecontrol serialized to v1 format (pickle) # using Python 2. A Python 2 str(byte string) will be unpickled as # a Python 3 str (unicode string), which will cause the above to # fail with: # # TypeError: 'str' does not support the buffer interface body = io.BytesIO(body_raw.encode("utf8")) # Discard any `strict` parameter serialized by older version of cachecontrol. cached["response"].pop("strict", None) return HTTPResponse(body=body, preload_content=False, **cached["response"]) def _loads_v0( self, request: PreparedRequest, data: bytes, body_file: IO[bytes] | None = None, ) -> None: # The original legacy cache data. This doesn't contain enough # information to construct everything we need, so we'll treat this as # a miss. return None def _loads_v1( self, request: PreparedRequest, data: bytes, body_file: IO[bytes] | None = None, ) -> HTTPResponse | None: # The "v1" pickled cache format. This is no longer supported # for security reasons, so we treat it as a miss. return None def _loads_v2( self, request: PreparedRequest, data: bytes, body_file: IO[bytes] | None = None, ) -> HTTPResponse | None: # The "v2" compressed base64 cache format. # This has been removed due to age and poor size/performance # characteristics, so we treat it as a miss. return None def _loads_v3( self, request: PreparedRequest, data: bytes, body_file: IO[bytes] | None = None, ) -> None: # Due to Python 2 encoding issues, it's impossible to know for sure # exactly how to load v3 entries, thus we'll treat these as a miss so # that they get rewritten out as v4 entries. return None def _loads_v4( self, request: PreparedRequest, data: bytes, body_file: IO[bytes] | None = None, ) -> HTTPResponse | None: try: cached = msgpack.loads(data, raw=False) except ValueError: return None return self.prepare_response(request, cached, body_file)