Wrote down the invariants of some common objects whose structure is

exposed in header files.  Fixed a few comments in these headers.

As we might have expected, writing down invariants systematically exposed a
(minor) bug.  In this case, function objects have a writeable func_code
attribute, which could be set to code objects with the wrong number of
free variables.  Calling the resulting function segfaulted the interpreter.
Added a corresponding test.

1 files changed, 8 insertions, 0 deletions

diff --git a/Include/setobject.h b/Include/setobject.h
index abbd847..cc2d683 100644
--- a/Include/setobject.h
+++ b/Include/setobject.h
@@ -16,6 +16,14 @@ typedef struct {
 	PyObject *data;
 	long hash;	/* only used by frozenset objects */
 	PyObject *weakreflist; /* List of weak references */
+
+	/* Invariants:
+	 *     data is a dictionary whose values are all True.
+	 *     data points to the same dict for the whole life of the set.
+	 * For frozensets only:
+	 *     data is immutable.
+	 *     hash is the hash of the frozenset or -1 if not computed yet.
+	 */
 } PySetObject;
 
 PyAPI_DATA(PyTypeObject) PySet_Type;