Issue #18149: Add filecmp.clear_cache() to manually clear the filecmp cache.

Patch by Mark Levitt

Doc/library/filecmp.rst

@@ -27,6 +27,10 @@ The :mod:`filecmp` module defines the following functions:
    Note that no external programs are called from this function, giving it
    portability and efficiency.
 
+   This function uses a cache for past comparisons and the results,
+   with a cache invalidation mechanism relying on stale signatures
+   or by explicitly calling :func:`clear_cache`.
+
 
 .. function:: cmpfiles(dir1, dir2, common, shallow=True)
 
@@ -48,6 +52,15 @@ The :mod:`filecmp` module defines the following functions:
    one of the three returned lists.
 
 
+.. function:: clear_cache()
+
+   .. versionadded:: 3.4
+
+   Clear the filecmp cache. This may be useful if a file is compared so quickly
+   after it is modified that it is within the mtime resolution of
+   the underlying filesystem.
+
+
 .. _dircmp-objects:
 
 The :class:`dircmp` class

Lib/filecmp.py

@@ -6,6 +6,7 @@ Classes:
 Functions:
     cmp(f1, f2, shallow=True) -> int
     cmpfiles(a, b, common) -> ([], [], [])
+    clear_cache()
 
 """
 
@@ -13,7 +14,7 @@ import os
 import stat
 from itertools import filterfalse
 
-__all__ = ['cmp', 'dircmp', 'cmpfiles', 'DEFAULT_IGNORES']
+__all__ = ['clear_cache', 'cmp', 'dircmp', 'cmpfiles', 'DEFAULT_IGNORES']
 
 _cache = {}
 BUFSIZE = 8*1024
@@ -21,6 +22,9 @@ BUFSIZE = 8*1024
 DEFAULT_IGNORES = [
     'RCS', 'CVS', 'tags', '.git', '.hg', '.bzr', '_darcs', '__pycache__']
 
+def clear_cache():
+    """Clear the filecmp cache."""
+    _cache.clear()
 
 def cmp(f1, f2, shallow=True):
     """Compare two files.
@@ -39,7 +43,8 @@ def cmp(f1, f2, shallow=True):
     True if the files are the same, False otherwise.
 
     This function uses a cache for past comparisons and the results,
-    with a cache invalidation mechanism relying on stale signatures.
+    with a cache invalidation mechanism relying on stale signatures
+    or by explicitly calling clear_cache().
 
     """
 
@@ -56,7 +61,7 @@ def cmp(f1, f2, shallow=True):
     if outcome is None:
         outcome = _do_cmp(f1, f2)
         if len(_cache) > 100:      # limit the maximum size of the cache
-            _cache.clear()
+            clear_cache()
         _cache[f1, f2, s1, s2] = outcome
     return outcome
 

Lib/test/test_filecmp.py

@@ -39,6 +39,13 @@ class FileCompareTestCase(unittest.TestCase):
         self.assertFalse(filecmp.cmp(self.name, self.dir),
                     "File and directory compare as equal")
 
+    def test_cache_clear(self):
+        first_compare = filecmp.cmp(self.name, self.name_same, shallow=False)
+        second_compare = filecmp.cmp(self.name, self.name_diff, shallow=False)
+        filecmp.clear_cache()
+        self.assertTrue(len(filecmp._cache) == 0,
+                        "Cache not cleared after calling clear_cache")
+
 class DirCompareTestCase(unittest.TestCase):
     def setUp(self):
         tmpdir = tempfile.gettempdir()

Misc/ACKS

@@ -738,6 +738,7 @@ Benno Leslie
 Christopher Tur Lesniewski-Laas
 Alain Leufroy
 Mark Levinson
+Mark Levitt
 William Lewis
 Akira Li
 Xuanji Li

Misc/NEWS

@@ -123,6 +123,9 @@ Core and Builtins
 Library
 -------
 
+- Issue #18149: Add filecmp.clear_cache() to manually clear the filecmp cache.
+  Patch by Mark Levitt
+
 - Issue #18193: Add importlib.reload().
 
 - Issue #18157: Stop using imp.load_module() in pydoc.