Issue2442

I have a tuple subclass that overrides those methods.  When trying to iterate over it, it enters an infinite loop between `__iter__` and `__len__`:

```
  File "spf.py", line 187, in __iter__
    yield getitem(node, 0)
  File "spf.py", line 203, in __len__
    return sum(1 for _ in self)
  File "spf.py", line 203, in <genexpr>
    return sum(1 for _ in self)
```

Here `getitem` is a call to `super(<tuple_subclass>, type(self)).__getitem__`, which in this case evaluates to `tuple.__getitem__`.

My working hypothesis for why this happens is that `tuple.__getitem__` calls the __len__ method of my subclass.

```
>>> class C(tuple):
...   def __len__(self):
...      print('Length')
...      return super(C, type(self)).__len__(self)
>>> c = C([1, 2])
>>> c[1]
Length
2
>>> tuple.__getitem__(c, 1)
Length
2
```

In general, my understanding is that calls to the methods of builtin types shouldn't call overridden methods.  This code works on CPython and PyPy.

This seems to be the fault of this code in org.python.core.__builtin__:

    public static int len(PyObject obj) {
        return obj.__len__();
    }

Rather related discussion here:
http://sourceforge.net/p/jython/mailman/message/34677839/

Also, I believe __getitem__ and __iter__ are implemented properly when used implicitly with [] and a foreach loop respectively. So it's just len/__len__ we have to fix.

Target RC

The problem is in __iter__, per Ceridwen's original message in msg10519

Fixed as of https://hg.python.org/jython/rev/85d99f6152b1 - I refactored iteration for list and tuple classes to not use other special methods.