Prevent corruption of memory cache

.github/workflows/enforce-labels.yml

@@ -7,7 +7,7 @@ jobs:
   enforce-label:
     runs-on: ubuntu-latest
     steps:
-    - uses: yogevbd/enforce-label-action@2.1.0
+    - uses: yogevbd/enforce-label-action@2.2.1
       with:
         REQUIRED_LABELS_ANY: "pr:change,pr:deprecation,pr:fix,pr:new-feature,pr:removal"
         REQUIRED_LABELS_ANY_DESCRIPTION: "Select at least one label with a 'pr:' prefix for this pull request" 

src/pymor/core/cache.py

@@ -63,6 +63,7 @@
 
 import atexit
 from collections import OrderedDict
+from copy import deepcopy
 import functools
 import getpass
 import hashlib
@@ -150,22 +151,15 @@ def get(self, key):
         if value is self.NO_VALUE:
             return False, None
         else:
-            from pymor.vectorarrays.interface import VectorArray
-            if isinstance(value, VectorArray):
-                value = value.copy()
-            return True, value
+            return True, deepcopy(value)
 
     def set(self, key, value):
         if key in self._cache:
             getLogger('pymor.core.cache.MemoryRegion').warning('Key already present in cache region, ignoring.')
             return
         if len(self._cache) == self.max_keys:
             self._cache.popitem(last=False)
-
-        import numpy as np
-        if isinstance(value, np.ndarray):
-            value.setflags(write=False)
-        self._cache[key] = value
+        self._cache[key] = deepcopy(value)
 
     def clear(self):
         self._cache = OrderedDict()

src/pymortests/cache.py

@@ -7,7 +7,11 @@
 from uuid import uuid4
 from datetime import datetime, timedelta
 
+import numpy as np
+
 from pymor.core import cache
+from pymor.models.basic import StationaryModel
+from pymor.operators.numpy import NumpyMatrixOperator
 from pymortests.base import runmodule
 
 
@@ -86,5 +90,21 @@ def test_region_api():
             assert backend.get('mykey') == (True, 1)
 
 
+def test_memory_region_safety():
+
+    op = NumpyMatrixOperator(np.eye(1))
+    rhs = op.range.make_array(np.array([1]))
+    m = StationaryModel(op, rhs)
+    m.enable_caching('memory')
+
+    U = m.solve()
+    del U[:]
+    U = m.solve()
+    assert len(U) == 1
+    del U[:]
+    U = m.solve()
+    assert len(U) == 1
+
+
 if __name__ == "__main__":
     runmodule(filename=__file__)