pythongh-83383: Always mark the dbm.dumb database as unmodified after open() and sync() (pythonGH-114560)

serhiy-storchaka · fsc-eriker · commit 16baf0eb78eb · 2024-02-14T16:28:44.000+02:00
The directory file for a newly created database is now created
immediately after opening instead of deferring this until synchronizing
or closing.
diff --git a/Lib/dbm/dumb.py b/Lib/dbm/dumb.py
@@ -98,7 +98,8 @@ def _update(self, flag):
         except OSError:
             if flag not in ('c', 'n'):
                 raise
-            self._modified = True
+            with self._io.open(self._dirfile, 'w', encoding="Latin-1") as f:
+                self._chmod(self._dirfile)
         else:
             with f:
                 for line in f:
@@ -134,6 +135,7 @@ def _commit(self):
                 # position; UTF-8, though, does care sometimes.
                 entry = "%r, %r\n" % (key.decode('Latin-1'), pos_and_siz_pair)
                 f.write(entry)
+        self._modified = False
 
     sync = _commit
 
diff --git a/Lib/test/test_dbm_dumb.py b/Lib/test/test_dbm_dumb.py
@@ -246,9 +246,27 @@ def test_missing_data(self):
             _delete_files()
             with self.assertRaises(FileNotFoundError):
                 dumbdbm.open(_fname, value)
+            self.assertFalse(os.path.exists(_fname + '.dat'))
             self.assertFalse(os.path.exists(_fname + '.dir'))
             self.assertFalse(os.path.exists(_fname + '.bak'))
 
+        for value in ('c', 'n'):
+            _delete_files()
+            with dumbdbm.open(_fname, value) as f:
+                self.assertTrue(os.path.exists(_fname + '.dat'))
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+            self.assertFalse(os.path.exists(_fname + '.bak'))
+
+        for value in ('c', 'n'):
+            _delete_files()
+            with dumbdbm.open(_fname, value) as f:
+                f['key'] = 'value'
+                self.assertTrue(os.path.exists(_fname + '.dat'))
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+            self.assertTrue(os.path.exists(_fname + '.bak'))
+
     def test_missing_index(self):
         with dumbdbm.open(_fname, 'n') as f:
             pass
@@ -259,6 +277,60 @@ def test_missing_index(self):
             self.assertFalse(os.path.exists(_fname + '.dir'))
             self.assertFalse(os.path.exists(_fname + '.bak'))
 
+        for value in ('c', 'n'):
+            with dumbdbm.open(_fname, value) as f:
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+            self.assertFalse(os.path.exists(_fname + '.bak'))
+            os.unlink(_fname + '.dir')
+
+        for value in ('c', 'n'):
+            with dumbdbm.open(_fname, value) as f:
+                f['key'] = 'value'
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+            self.assertTrue(os.path.exists(_fname + '.bak'))
+            os.unlink(_fname + '.dir')
+            os.unlink(_fname + '.bak')
+
+    def test_sync_empty_unmodified(self):
+        with dumbdbm.open(_fname, 'n') as f:
+            pass
+        os.unlink(_fname + '.dir')
+        for value in ('c', 'n'):
+            with dumbdbm.open(_fname, value) as f:
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+                f.sync()
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+                os.unlink(_fname + '.dir')
+                f.sync()
+                self.assertFalse(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+            self.assertFalse(os.path.exists(_fname + '.dir'))
+            self.assertFalse(os.path.exists(_fname + '.bak'))
+
+    def test_sync_nonempty_unmodified(self):
+        with dumbdbm.open(_fname, 'n') as f:
+            pass
+        os.unlink(_fname + '.dir')
+        for value in ('c', 'n'):
+            with dumbdbm.open(_fname, value) as f:
+                f['key'] = 'value'
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+                f.sync()
+                self.assertTrue(os.path.exists(_fname + '.dir'))
+                self.assertTrue(os.path.exists(_fname + '.bak'))
+                os.unlink(_fname + '.dir')
+                os.unlink(_fname + '.bak')
+                f.sync()
+                self.assertFalse(os.path.exists(_fname + '.dir'))
+                self.assertFalse(os.path.exists(_fname + '.bak'))
+            self.assertFalse(os.path.exists(_fname + '.dir'))
+            self.assertFalse(os.path.exists(_fname + '.bak'))
+
     def test_invalid_flag(self):
         for flag in ('x', 'rf', None):
             with self.assertRaisesRegex(ValueError,
diff --git a/Misc/NEWS.d/next/Library/2024-01-25-19-22-17.gh-issue-83383.3GwO9v.rst b/Misc/NEWS.d/next/Library/2024-01-25-19-22-17.gh-issue-83383.3GwO9v.rst
@@ -0,0 +1,5 @@
+Synchronization of the :mod:`dbm.dumb` database is now no-op if there was no
+modification since opening or last synchronization.
+The directory file for a newly created empty :mod:`dbm.dumb` database is now
+created immediately after opening instead of deferring this until
+synchronizing or closing.
