Test for efficacy of fast sequence locking macro on str.join, both tests segfault or die with assertion failures >95% of the time without locking macro, pass reliably with locking macro

MojoVampire · MojoVampire · commit 2b4fe199c77d · 2024-05-22T16:09:28.000Z
diff --git a/Include/internal/pycore_critical_section.h b/Include/internal/pycore_critical_section.h
@@ -114,7 +114,7 @@ extern "C" {
 // For performance, the argument *to* PySequence_Fast is provided to the
 // macro, not the *result* of PySequence_Fast (which would require an extra
 // test to determine if the lock must be held)
-# define Py_BEGIN_CRITICAL_SECTION_SEQUENCE_FAST(original)            	\
+# define Py_BEGIN_CRITICAL_SECTION_SEQUENCE_FAST(original)              \
     {                                                                   \
         PyObject *_orig_seq = _PyObject_CAST(original);                 \
         const bool _should_lock_cs = PyList_CheckExact(_orig_seq);      \
diff --git a/Lib/test/test_free_threading/test_str.py b/Lib/test/test_free_threading/test_str.py
@@ -0,0 +1,74 @@
+import sys
+import unittest
+
+from itertools import cycle
+from threading import Thread
+from unittest import TestCase
+
+from test.support import threading_helper
+
+@threading_helper.requires_working_threading()
+class TestStr(TestCase):
+    def test_racing_join_extend(self):
+        '''Test joining a string being extended by another thread'''
+        l = []
+        OBJECT_COUNT = 100_000
+
+        def writer_func():
+            l.extend(map(str, range(OBJECT_COUNT, OBJECT_COUNT*2)))
+
+        def reader_func():
+            while True:
+                count = len(l)
+                ''.join(l)
+                if count == OBJECT_COUNT:
+                    break
+
+        writer = Thread(target=writer_func)
+        readers = []
+        for x in range(30):
+            reader = Thread(target=reader_func)
+            readers.append(reader)
+            reader.start()
+
+        writer.start()
+        writer.join()
+        for reader in readers:
+            reader.join()
+
+    def test_racing_join_replace(self):
+        '''
+        Test joining a string of characters being replaced with ephemeral
+        strings by another thread.
+        '''
+        l = [*'abcdefg']
+        MAX_ORDINAL = 10_000
+
+        def writer_func():
+            for i, c in zip(cycle(range(len(l))),
+                            map(chr, range(128, MAX_ORDINAL))):
+                l[i] = c
+            del l[:]  # Empty list to tell readers to exit
+
+        def reader_func():
+            while True:
+                empty = not l
+                ''.join(l)
+                if empty:
+                    break
+
+        writer = Thread(target=writer_func)
+        readers = []
+        for x in range(30):
+            reader = Thread(target=reader_func)
+            readers.append(reader)
+            reader.start()
+
+        writer.start()
+        writer.join()
+        for reader in readers:
+            reader.join()
+
+
+if __name__ == "__main__":
+    unittest.main()
