[3.13] gh-120158: Fix inconsistent monitoring state when setting events too frequently (gh-141845) (gh-141880)

If we overflowed the global version counter (i.e., after 2*24 calls to
`_PyMonitoring_SetEvents`), we bailed out after setting global monitoring
events but before instrumenting code objects, which led to assertion errors
later on.

Also add a `time.sleep()` to `test_free_threading.test_monitoring` to avoid
overflowing the global version counter.
(cherry picked from commit e457d60)

Author: colesbury

Lib/test/test_free_threading/test_monitoring.py

@@ -72,6 +72,9 @@ def test_instrumentation(self):
                 break
 
             self.during_threads()
+            # Sleep to avoid setting monitoring events too rapidly and
+            # overflowing the global version counter
+            time.sleep(0.0001)
 
         self.after_test()
 

Misc/NEWS.d/next/Core and Builtins/2025-11-22-10-43-26.gh-issue-120158.41_rXd.rst

@@ -0,0 +1,2 @@
+Fix inconsistent state when enabling or disabling monitoring events too many
+times.

Python/instrumentation.c

@@ -1989,12 +1989,12 @@ _PyMonitoring_SetEvents(int tool_id, _PyMonitoringEventSet events)
     if (existing_events == events) {
         return 0;
     }
-    set_events(&interp->monitors, tool_id, events);
     uint32_t new_version = global_version(interp) + MONITORING_VERSION_INCREMENT;
     if (new_version == 0) {
         PyErr_Format(PyExc_OverflowError, "events set too many times");
         return -1;
     }
+    set_events(&interp->monitors, tool_id, events);
     set_global_version(tstate, new_version);
 #ifdef _Py_TIER2
     _Py_Executors_InvalidateAll(interp, 1);