bpo-36818: Add PyInterpreterState.runtime field. (gh-13129)
https://bugs.python.org/issue36818
diff --git a/Include/cpython/pystate.h b/Include/cpython/pystate.h
index 94b0809..74e7fc9 100644
--- a/Include/cpython/pystate.h
+++ b/Include/cpython/pystate.h
@@ -110,9 +110,9 @@
* if the thread holds the last reference to the lock, decref'ing the
* lock will delete the lock, and that may trigger arbitrary Python code
* if there's a weakref, with a callback, to the lock. But by this time
- * _PyRuntime.gilstate.tstate_current is already NULL, so only the simplest
- * of C code can be allowed to run (in particular it must not be possible to
- * release the GIL).
+ * _PyRuntimeState.gilstate.tstate_current is already NULL, so only the
+ * simplest of C code can be allowed to run (in particular it must not be
+ * possible to release the GIL).
* So instead of holding the lock directly, the tstate holds a weakref to
* the lock: that's the value of on_delete_data below. Decref'ing a
* weakref is harmless.