>>> class I(int):
...     def __str__(self):
...         return 'spam'
... 
>>> '%i' % (I(42),)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
SystemError: Objects/unicodeobject.c:13595: bad argument to internal function

This issue is related to issue18738 and will be fixed when use PyNumber_ToBase() for any int subclass (not only for bool).

formatlong() should copy the string if Py_REFCNT(str) != 1.

I think we should just use PyNumber_ToBase() for any int and int subclass. The calling __str__() for special case looks weird.

I'll get my patch separated and over here when I get back home (on a business trip at the moment).

Okay, simple fix, patch and tests attached.

lgtm

New changeset 33727fbb4668 by Ethan Furman in branch 'default':
Close #18780: %-formatting now prints value for int subclasses with %d, %i, and %u codes.
http://hg.python.org/cpython/rev/33727fbb4668

New changeset 833246d42825 by Ethan Furman in branch 'default':
Issue #18780: code cleanup.
http://hg.python.org/cpython/rev/833246d42825

Should this patch also go into the 3.3 branch?  It only went into 3.4.

If yes, how should I go about doing that?

The 3.3 branch is now only open for security fixes so this issue doesn't appear to warrant backporting there.

Cool, leaving it closed.

On Tue, Jun 24, 2014 at 2:29 PM, Ned Deily <report@bugs.python.org> wrote:

>
> Ned Deily added the comment:
>
> The 3.3 branch is now only open for security fixes so this issue doesn't
> appear to warrant backporting there.
>
> ----------
>

These questions keep popping up in various places, being asked even by core
devs like here.

Is there a place where this is described formally? I.e. 3.5 is trunk, 3.4
is bug fixes, 3.3 is security fixes, doc fixes good everywhere, and some
info about 2.7? I imagine since this only changes once per 18 months or so,
it should not be difficult to maintain and can just be part of the release
manager's job.

It is described in the developer's guide.  The current status is summarized here:

https://docs.python.org/devguide/devcycle.html#summary

On Tue, Jun 24, 2014 at 3:18 PM, Ned Deily <report@bugs.python.org> wrote:

>
> Ned Deily added the comment:
>
> It is described in the developer's guide.  The current status is
> summarized here:
>
> https://docs.python.org/devguide/devcycle.html#summary
>
>
Excellent. Thanks Ned.