Lazy load hashlib.md5 file system cache (#373)

* lazy load md5

* add changelog entry

* extend tests

CHANGES.rst

@@ -1,3 +1,14 @@
+Version 0.13.0
+--------------
+
+Unreleased
+
+- default ``hashlib.md5`` may not be available in FIPS builds. We
+ now do not access it at import time on ``FileSystemCache``so developers
+ have time to change the default.
+ ``hashlib.md5`` will be lazy loaded when a new default is not provided
+
+
 Version 0.12.0
 --------------
 

src/cachelib/file.py

@@ -1,4 +1,5 @@
 import errno
+import hashlib
 import logging
 import os
 import platform
@@ -7,7 +8,6 @@
 import tempfile
 import typing as _t
 from contextlib import contextmanager
-from hashlib import md5
 from pathlib import Path
 from time import sleep
 from time import time
@@ -16,6 +16,14 @@
 from cachelib.serializers import FileSystemSerializer
 
 
+def _lazy_md5(string: bytes = b"") -> _t.Any:
+ """Don't access ``hashlib.md5`` until runtime. FIPS builds may not include
+ md5, in which case the import and use as a default would fail before the
+ developer can configure something else.
+ """
+ return hashlib.md5(string)
+
+
 class FileSystemCache(BaseCache):
  """A cache that stores the items on the file system. This cache depends
  on being the only user of the `cache_dir`. Make absolutely sure that
@@ -38,6 +46,8 @@ class FileSystemCache(BaseCache):
  _fs_transaction_suffix = ".__wz_cache"
  #: keep amount of files in a cache element
  _fs_count_file = "__wz_cache_count"
+ #: default file name hashing method
+ _default_hash_method = staticmethod(_lazy_md5)
 
  serializer = FileSystemSerializer()
 
@@ -47,12 +57,15 @@ def __init__(
  threshold: int = 500,
  default_timeout: int = 300,
  mode: _t.Optional[int] = None,
- hash_method: _t.Any = md5,
+ hash_method: _t.Any = None,
  ):
  BaseCache.__init__(self, default_timeout)
  self._path = cache_dir
  self._threshold = threshold
- self._hash_method = hash_method
+
+ self._hash_method = self._default_hash_method
+ if hash_method is not None:
+ self._hash_method = hash_method
 
  # Mode set by user takes precedence. If no mode has
  # been given, we need to set the correct default based

tests/test_file_system_cache.py

@@ -42,9 +42,21 @@ def __init__(self, *args, **kwargs):
  super().__init__(*args, hash_method=hashlib.sha256, **kwargs)
 
 
+class CustomDefaultHashingMethodCache(FileSystemCache):
+ _default_hash_method = hashlib.sha256
+
+ def __init__(self, *args, **kwargs):
+ super().__init__(*args, **kwargs)
+
+
 @pytest.fixture(
  autouse=True,
- params=[FileSystemCache, CustomSerializerCache, CustomHashingMethodCache],
+ params=[
+ FileSystemCache,
+ CustomSerializerCache,
+ CustomHashingMethodCache,
+ CustomDefaultHashingMethodCache,
+ ],
 )
 def cache_factory(request, tmpdir):
  def _factory(self, *args, **kwargs):