Merge pull request #10815 from pradyunsg/vendoring-upgrade

Author: pradyunsg

MANIFEST.in

@@ -6,6 +6,7 @@ include pyproject.toml
 
 include src/pip/_vendor/README.rst
 include src/pip/_vendor/vendor.txt
+include src/pip/_vendor/pyparsing/diagram/template.jinja2
 recursive-include  src/pip/_vendor *LICENSE*
 recursive-include  src/pip/_vendor *COPYING*
 

news/CacheControl.vendor.rst

@@ -0,0 +1 @@
+Upgrade CacheControl to 0.12.10

news/certifi.vendor.rst

@@ -0,0 +1 @@
+Upgrade certifi to 2021.10.8

news/distlib.vendor.rst

@@ -0,0 +1 @@
+Upgrade distlib to 0.3.4

news/idna.vendor.rst

@@ -0,0 +1 @@
+Upgrade idna to 3.3

news/msgpack.vendor.rst

@@ -0,0 +1 @@
+Upgrade msgpack to 1.0.3

news/packaging.vendor.rst

@@ -0,0 +1 @@
+Upgrade packaging to 21.3

news/platformdirs.vendor.rst

@@ -0,0 +1 @@
+Upgrade platformdirs to 2.4.1

news/pygments.vendor.rst

@@ -1 +1 @@
-Add pygments 2.10.0 as a vendored dependency.
+Add pygments 2.11.2 as a vendored dependency.

news/pyparsing.vendor.rst

@@ -0,0 +1 @@
+Upgrade pyparsing to 3.0.7

news/rich-modifications.vendor.rst

@@ -0,0 +1 @@
+Tree-trim unused portions of vendored rich, to reduce the distribution size.

news/rich.vendor.rst

@@ -1 +1 @@
-Add rich 10.14.0 as a vendored dependency.
+Add rich 11.0.0 as a vendored dependency.

news/typing_extensions.vendor.rst

@@ -1 +1 @@
-Add typing_extensions 3.10.0.2 as a vendored dependency.
+Add typing_extensions 4.0.1 as a vendored dependency.

src/pip/_internal/network/cache.py

@@ -53,7 +53,7 @@ def get(self, key: str) -> Optional[bytes]:
             with open(path, "rb") as f:
                 return f.read()
 
-    def set(self, key: str, value: bytes) -> None:
+    def set(self, key: str, value: bytes, expires: Optional[int] = None) -> None:
         path = self._get_cache_path(key)
         with suppressed_cache_errors():
             ensure_dir(os.path.dirname(path))

src/pip/_vendor/cachecontrol/LICENSE.txt

@@ -1,4 +1,4 @@
-Copyright 2015 Eric Larson
+Copyright 2012-2021  Eric Larson
 
 Licensed under the Apache License, Version 2.0 (the "License");
 you may not use this file except in compliance with the License.
@@ -8,8 +8,6 @@ You may obtain a copy of the License at
 
 Unless required by applicable law or agreed to in writing, software
 distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
-implied.
-
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License.

src/pip/_vendor/cachecontrol/__init__.py

@@ -1,11 +1,18 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 """CacheControl import Interface.
 
 Make it easy to import from cachecontrol without long namespaces.
 """
 __author__ = "Eric Larson"
 __email__ = "[email protected]"
-__version__ = "0.12.6"
+__version__ = "0.12.10"
 
 from .wrapper import CacheControl
 from .adapter import CacheControlAdapter
 from .controller import CacheController
+
+import logging
+logging.getLogger(__name__).addHandler(logging.NullHandler())

src/pip/_vendor/cachecontrol/_cmd.py

@@ -1,3 +1,7 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 import logging
 
 from pip._vendor import requests

src/pip/_vendor/cachecontrol/adapter.py

@@ -1,16 +1,20 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 import types
 import functools
 import zlib
 
 from pip._vendor.requests.adapters import HTTPAdapter
 
-from .controller import CacheController
+from .controller import CacheController, PERMANENT_REDIRECT_STATUSES
 from .cache import DictCache
 from .filewrapper import CallbackFileWrapper
 
 
 class CacheControlAdapter(HTTPAdapter):
-    invalidating_methods = {"PUT", "DELETE"}
+    invalidating_methods = {"PUT", "PATCH", "DELETE"}
 
     def __init__(
         self,
@@ -93,7 +97,7 @@ def build_response(
                 response = cached_response
 
             # We always cache the 301 responses
-            elif response.status == 301:
+            elif int(response.status) in PERMANENT_REDIRECT_STATUSES:
                 self.controller.cache_response(request, response)
             else:
                 # Wrap the response file with a wrapper that will cache the

src/pip/_vendor/cachecontrol/cache.py

@@ -1,3 +1,7 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 """
 The cache object API for implementing caches. The default is a thread
 safe in-memory dictionary.
@@ -10,7 +14,7 @@ class BaseCache(object):
     def get(self, key):
         raise NotImplementedError()
 
-    def set(self, key, value):
+    def set(self, key, value, expires=None):
         raise NotImplementedError()
 
     def delete(self, key):
@@ -29,7 +33,7 @@ def __init__(self, init_dict=None):
     def get(self, key):
         return self.data.get(key, None)
 
-    def set(self, key, value):
+    def set(self, key, value, expires=None):
         with self.lock:
             self.data.update({key: value})
 

src/pip/_vendor/cachecontrol/caches/__init__.py

@@ -1,2 +1,6 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 from .file_cache import FileCache  # noqa
 from .redis_cache import RedisCache  # noqa

src/pip/_vendor/cachecontrol/caches/file_cache.py

@@ -1,3 +1,7 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 import hashlib
 import os
 from textwrap import dedent
@@ -114,7 +118,7 @@ def get(self, key):
         except FileNotFoundError:
             return None
 
-    def set(self, key, value):
+    def set(self, key, value, expires=None):
         name = self._fn(key)
 
         # Make sure the directory exists

src/pip/_vendor/cachecontrol/caches/redis_cache.py

@@ -1,3 +1,7 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 from __future__ import division
 
 from datetime import datetime

src/pip/_vendor/cachecontrol/compat.py

@@ -1,3 +1,7 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 try:
     from urllib.parse import urljoin
 except ImportError:
@@ -9,7 +13,6 @@
 except ImportError:
     import pickle
 
-
 # Handle the case where the requests module has been patched to not have
 # urllib3 bundled as part of its source.
 try:

src/pip/_vendor/cachecontrol/controller.py

@@ -1,3 +1,7 @@
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
 """
 The httplib2 algorithms ported for use with requests.
 """
@@ -17,6 +21,8 @@
 
 URI = re.compile(r"^(([^:/?#]+):)?(//([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?")
 
+PERMANENT_REDIRECT_STATUSES = (301, 308)
+
 
 def parse_uri(uri):
     """Parses a URI using the regex given in Appendix B of RFC 3986.
@@ -37,7 +43,7 @@ def __init__(
         self.cache = DictCache() if cache is None else cache
         self.cache_etags = cache_etags
         self.serializer = serializer or Serializer()
-        self.cacheable_status_codes = status_codes or (200, 203, 300, 301)
+        self.cacheable_status_codes = status_codes or (200, 203, 300, 301, 308)
 
     @classmethod
     def _urlnorm(cls, uri):
@@ -147,17 +153,18 @@ def cached_request(self, request):
             logger.warning("Cache entry deserialization failed, entry ignored")
             return False
 
-        # If we have a cached 301, return it immediately. We don't
-        # need to test our response for other headers b/c it is
+        # If we have a cached permanent redirect, return it immediately. We
+        # don't need to test our response for other headers b/c it is
         # intrinsically "cacheable" as it is Permanent.
+        #
         # See:
         #   https://tools.ietf.org/html/rfc7231#section-6.4.2
         #
         # Client can try to refresh the value by repeating the request
         # with cache busting headers as usual (ie no-cache).
-        if resp.status == 301:
+        if int(resp.status) in PERMANENT_REDIRECT_STATUSES:
             msg = (
-                'Returning cached "301 Moved Permanently" response '
+                "Returning cached permanent redirect response "
                 "(ignoring date and etag information)"
             )
             logger.debug(msg)
@@ -261,6 +268,11 @@ def cache_response(self, request, response, body=None, status_codes=None):
 
         response_headers = CaseInsensitiveDict(response.headers)
 
+        if "date" in response_headers:
+            date = calendar.timegm(parsedate_tz(response_headers["date"]))
+        else:
+            date = 0
+
         # If we've been given a body, our response has a Content-Length, that
         # Content-Length is valid then we can check to see if the body we've
         # been given matches the expected size, and if it doesn't we'll just
@@ -304,35 +316,62 @@ def cache_response(self, request, response, body=None, status_codes=None):
 
         # If we've been given an etag, then keep the response
         if self.cache_etags and "etag" in response_headers:
+            expires_time = 0
+            if response_headers.get("expires"):
+                expires = parsedate_tz(response_headers["expires"])
+                if expires is not None:
+                    expires_time = calendar.timegm(expires) - date
+
+            expires_time = max(expires_time, 14 * 86400)
+
+            logger.debug("etag object cached for {0} seconds".format(expires_time))
             logger.debug("Caching due to etag")
             self.cache.set(
-                cache_url, self.serializer.dumps(request, response, body=body)
+                cache_url,
+                self.serializer.dumps(request, response, body),
+                expires=expires_time,
             )
 
-        # Add to the cache any 301s. We do this before looking that
-        # the Date headers.
-        elif response.status == 301:
-            logger.debug("Caching permanant redirect")
-            self.cache.set(cache_url, self.serializer.dumps(request, response))
+        # Add to the cache any permanent redirects. We do this before looking
+        # that the Date headers.
+        elif int(response.status) in PERMANENT_REDIRECT_STATUSES:
+            logger.debug("Caching permanent redirect")
+            self.cache.set(cache_url, self.serializer.dumps(request, response, b""))
 
         # Add to the cache if the response headers demand it. If there
         # is no date header then we can't do anything about expiring
         # the cache.
         elif "date" in response_headers:
+            date = calendar.timegm(parsedate_tz(response_headers["date"]))
             # cache when there is a max-age > 0
             if "max-age" in cc and cc["max-age"] > 0:
                 logger.debug("Caching b/c date exists and max-age > 0")
+                expires_time = cc["max-age"]
                 self.cache.set(
-                    cache_url, self.serializer.dumps(request, response, body=body)
+                    cache_url,
+                    self.serializer.dumps(request, response, body),
+                    expires=expires_time,
                 )
 
             # If the request can expire, it means we should cache it
             # in the meantime.
             elif "expires" in response_headers:
                 if response_headers["expires"]:
-                    logger.debug("Caching b/c of expires header")
+                    expires = parsedate_tz(response_headers["expires"])
+                    if expires is not None:
+                        expires_time = calendar.timegm(expires) - date
+                    else:
+                        expires_time = None
+
+                    logger.debug(
+                        "Caching b/c of expires header. expires in {0} seconds".format(
+                            expires_time
+                        )
+                    )
                     self.cache.set(
-                        cache_url, self.serializer.dumps(request, response, body=body)
+                        cache_url,
+                        self.serializer.dumps(request, response, body=body),
+                        expires=expires_time,
                     )
 
     def update_cached_response(self, request, response):

src/pip/_vendor/cachecontrol/filewrapper.py

@@ -1,4 +1,9 @@
-from io import BytesIO
+# SPDX-FileCopyrightText: 2015 Eric Larson
+#
+# SPDX-License-Identifier: Apache-2.0
+
+from tempfile import NamedTemporaryFile
+import mmap
 
 
 class CallbackFileWrapper(object):
@@ -11,10 +16,17 @@ class CallbackFileWrapper(object):
 
     This class uses members with a double underscore (__) leading prefix so as
     not to accidentally shadow an attribute.
+
+    The data is stored in a temporary file until it is all available.  As long
+    as the temporary files directory is disk-based (sometimes it's a
+    memory-backed-``tmpfs`` on Linux), data will be unloaded to disk if memory
+    pressure is high.  For small files the disk usually won't be used at all,
+    it'll all be in the filesystem memory cache, so there should be no
+    performance impact.
     """
 
     def __init__(self, fp, callback):
-        self.__buf = BytesIO()
+        self.__buf = NamedTemporaryFile("rb+", delete=True)
         self.__fp = fp
         self.__callback = callback
 
@@ -49,7 +61,19 @@ def __is_fp_closed(self):
 
     def _close(self):
         if self.__callback:
-            self.__callback(self.__buf.getvalue())
+            if self.__buf.tell() == 0:
+                # Empty file:
+                result = b""
+            else:
+                # Return the data without actually loading it into memory,
+                # relying on Python's buffer API and mmap(). mmap() just gives
+                # a view directly into the filesystem's memory cache, so it
+                # doesn't result in duplicate memory use.
+                self.__buf.seek(0, 0)
+                result = memoryview(
+                    mmap.mmap(self.__buf.fileno(), 0, access=mmap.ACCESS_READ)
+                )
+            self.__callback(result)
 
         # We assign this to None here, because otherwise we can get into
         # really tricky problems where the CPython interpreter dead locks
@@ -58,9 +82,16 @@ def _close(self):
         # and allows the garbage collector to do it's thing normally.
         self.__callback = None
 
+        # Closing the temporary file releases memory and frees disk space.
+        # Important when caching big files.
+        self.__buf.close()
+
     def read(self, amt=None):
         data = self.__fp.read(amt)
-        self.__buf.write(data)
+        if data:
+            # We may be dealing with b'', a sign that things are over:
+            # it's passed e.g. after we've already closed self.__buf.
+            self.__buf.write(data)
         if self.__is_fp_closed():
             self._close()