Move get_cache_path_for_link to WheelCache

pip/cache.py

@@ -2,6 +2,7 @@
 """
 
 import errno
+import hashlib
 import logging
 import os
 
@@ -10,7 +11,6 @@
 import pip.index
 from pip.compat import expanduser
 from pip.download import path_to_url
-from pip.utils.cache import get_cache_path_for_link
 from pip.wheel import InvalidWheelFilename, Wheel
 
 logger = logging.getLogger(__name__)
@@ -29,6 +29,46 @@ def __init__(self, cache_dir, format_control):
         self._cache_dir = expanduser(cache_dir) if cache_dir else None
         self._format_control = format_control
 
+    def get_cache_path_for_link(self, link):
+        """
+        Return a directory to store cached wheels in for link.
+
+        Because there are M wheels for any one sdist, we provide a directory
+        to cache them in, and then consult that directory when looking up
+        cache hits.
+
+        We only insert things into the cache if they have plausible version
+        numbers, so that we don't contaminate the cache with things that were
+        not unique. E.g. ./package might have dozens of installs done for it
+        and build a version of 0.0...and if we built and cached a wheel, we'd
+        end up using the same wheel even if the source has been edited.
+
+        :param link: The link of the sdist for which this will cache wheels.
+        """
+
+        # We want to generate an url to use as our cache key, we don't want to
+        # just re-use the URL because it might have other items in the fragment
+        # and we don't care about those.
+        key_parts = [link.url_without_fragment]
+        if link.hash_name is not None and link.hash is not None:
+            key_parts.append("=".join([link.hash_name, link.hash]))
+        key_url = "#".join(key_parts)
+
+        # Encode our key url with sha224, we'll use this because it has similar
+        # security properties to sha256, but with a shorter total output (and
+        # thus less secure). However the differences don't make a lot of
+        # difference for our use case here.
+        hashed = hashlib.sha224(key_url.encode()).hexdigest()
+
+        # We want to nest the directories some to prevent having a ton of top
+        # level directories where we might run out of sub directories on some
+        # FS.
+        parts = [hashed[:2], hashed[2:4], hashed[4:6], hashed[6:]]
+
+        # Inside of the base location for cached wheels, expand our parts and
+        # join them all together.
+        return os.path.join(self._cache_dir, "wheels", *parts)
+
     def cached_wheel(self, link, package_name):
         not_cached = (
             not self._cache_dir or
@@ -47,7 +87,7 @@ def cached_wheel(self, link, package_name):
         )
         if "binary" not in formats:
             return link
-        root = get_cache_path_for_link(self._cache_dir, link)
+        root = self.get_cache_path_for_link(link)
         try:
             wheel_names = os.listdir(root)
         except OSError as err:

pip/wheel.py

@@ -31,7 +31,6 @@
 )
 from pip.locations import PIP_DELETE_MARKER_FILENAME, distutils_scheme
 from pip.utils import call_subprocess, captured_stdout, ensure_dir, read_chunks
-from pip.utils.cache import get_cache_path_for_link
 from pip.utils.logging import indent_log
 from pip.utils.setuptools_build import SETUPTOOLS_SHIM
 from pip.utils.temp_dir import TempDirectory
@@ -780,8 +779,8 @@ def build(self, session, autobuilding=False):
                 if autobuilding:
                     python_tag = pep425tags.implementation_tag
                     # NOTE: Should move out a method on the cache directly.
-                    output_dir = get_cache_path_for_link(
-                        self.wheel_cache._cache_dir, req.link
+                    output_dir = self.wheel_cache.get_cache_path_for_link(
+                        req.link
                     )
                     try:
                         ensure_dir(output_dir)