[Python-checkins] r58288 - in python/trunk: Misc/NEWS Objects/object.c Objects/tupleobject.c

[Brett Cannon]

I know this broke some tests.  I could have sworn I ran regrtest when
I first made the changes.

I am looking into possible fixes.

-Brett

On 9/30/07, brett.cannon <python-checkins at python.org> wrote:
> Author: brett.cannon
> Date: Sun Sep 30 21:45:10 2007
> New Revision: 58288
>
> Modified:
>    python/trunk/Misc/NEWS
>    python/trunk/Objects/object.c
>    python/trunk/Objects/tupleobject.c
> Log:
> tuple.__repr__ did not consider a reference loop as it is not possible from
> Python code; but it is possible from C.  object.__str__ had the issue of not
> expecting a type to doing something within it's tp_str implementation that
> could trigger an infinite recursion, but it could in C code..  Both found
> thanks to BaseException and how it handles its repr.
>
> Closes issue #1686386.  Thanks to Thomas Herve for taking an initial stab at
> coming up with a solution.
>
>
> Modified: python/trunk/Misc/NEWS
> ==============================================================================
> --- python/trunk/Misc/NEWS      (original)
> +++ python/trunk/Misc/NEWS      Sun Sep 30 21:45:10 2007
> @@ -12,6 +12,12 @@
>  Core and builtins
>  -----------------
>
> +- Issue #1686386: Tuple's tp_repr did not take into account the possibility of
> +  having a self-referential tuple, which is possible from C code.  Nor did
> +  object's tp_str consider that a type's tp_str could do something that could
> +  lead to an inifinite recursion.  Py_ReprEnter() and Py_EnterRecursiveCall(),
> +  respectively, fixed the issues.
> +
>  - Issue #1164: It was possible to trigger deadlock when using the 'print'
>    statement to write to a file since the GIL was not released as needed.  Now
>    PyObject_Print() does the right thing along with various tp_print
>
> Modified: python/trunk/Objects/object.c
> ==============================================================================
> --- python/trunk/Objects/object.c       (original)
> +++ python/trunk/Objects/object.c       Sun Sep 30 21:45:10 2007
> @@ -408,7 +408,12 @@
>         if (Py_Type(v)->tp_str == NULL)
>                 return PyObject_Repr(v);
>
> +       /* It is possible for a type to have a tp_str representation that loops
> +          infinitely. */
> +       if (Py_EnterRecursiveCall(" while getting the str of an object"))
> +               return NULL;
>         res = (*Py_Type(v)->tp_str)(v);
> +       Py_LeaveRecursiveCall();
>         if (res == NULL)
>                 return NULL;
>         type_ok = PyString_Check(res);
>
> Modified: python/trunk/Objects/tupleobject.c
> ==============================================================================
> --- python/trunk/Objects/tupleobject.c  (original)
> +++ python/trunk/Objects/tupleobject.c  Sun Sep 30 21:45:10 2007
> @@ -216,6 +216,15 @@
>         PyObject *s, *temp;
>         PyObject *pieces, *result = NULL;
>
> +       /* While not mutable, it is still possible to end up with a cycle in a
> +          tuple through an object that stores itself within a tuple (and thus
> +          infinitely asks for the repr of itself). This should only be
> +          possible within a type. */
> +       i = Py_ReprEnter((PyObject *)v);
> +       if (i != 0) {
> +               return i > 0 ? PyString_FromString("(...)") : NULL;
> +       }
> +
>         n = Py_Size(v);
>         if (n == 0)
>                 return PyString_FromString("()");
> @@ -226,7 +235,10 @@
>
>         /* Do repr() on each element. */
>         for (i = 0; i < n; ++i) {
> +               if (Py_EnterRecursiveCall(" while getting the repr of a tuple"))
> +                       goto Done;
>                 s = PyObject_Repr(v->ob_item[i]);
> +               Py_LeaveRecursiveCall();
>                 if (s == NULL)
>                         goto Done;
>                 PyTuple_SET_ITEM(pieces, i, s);
> @@ -261,6 +273,7 @@
>
>  Done:
>         Py_DECREF(pieces);
> +       Py_ReprLeave((PyObject *)v);
>         return result;
>  }
>
> _______________________________________________
> Python-checkins mailing list
> Python-checkins at python.org
> http://mail.python.org/mailman/listinfo/python-checkins
>