[issue17950] Dynamic classes contain non-breakable reference cycles

[Kristján Valur Jónsson]

Kristján Valur Jónsson added the comment:

Thanks Guido.
The current patch provides the property you ask for.  I will see if I can make the "fiddling" of the internal tuple less magical.

I have one other question for you:  The standard "mro" puts the class in the 0th position.
But apparently, there is a mechanism for a "custom" mro, by calling an mro() function on the type (as far as I understand).
However, the order of objects in the returned tuple is not verified, only the types of the objects therein (this is in mro_internal())
Yet there is code that manually skips the 0th element, e.g. this code 

/* Initialize tp_dict properly */
    bases = type->tp_mro;
    assert(bases != NULL);
    assert(PyTuple_Check(bases));
    n = PyTuple_GET_SIZE(bases);
    for (i = 1; i < n; i++) {
        PyObject *b = PyTuple_GET_ITEM(bases, i);
        if (PyType_Check(b))
            inherit_slots(type, (PyTypeObject *)b);
    }
(from PyType_Ready())

I'm not familiar with the mro() function.  What kind of black magic is it supposed to provide?  And how can we make sure that its free-form return value is reconciled with the 1-based loop above?

----------

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue17950>
_______________________________________