Matplotlib animation with Qt backend in jupyter

It seems the Qt backend cannot be used in Jupyter to animate a plot. Is that true in general?

The code shown below runs fine using the Qt4Agg or TkAgg backend in a script. It also runs fine using the notebook backend (%matplotlib notebook) or the tk backend (%matplotlib tk) in a Jupyter notebook.

However, when using %matplotlib qt (or %matplotlib qt4) in Jupyter, the window freezes and the Kernel dies.

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.animation 

y = np.cumsum(np.random.normal(size=30))

fig, ax = plt.subplots()
line, = ax.plot(np.arange(len(y)),y)
ax.set_xlim(0,30)
ax.set_ylim(y.min(),y.max())

def update(i):
    x = np.arange(i)
    line.set_data(x,y[:i])

ani = matplotlib.animation.FuncAnimation(fig, update, frames=30, repeat=False)
plt.show()

Commenting out the line ani = matplotlib.animation.FuncAnimation(...), the window appears and stays responsive also for %matplotlib qt backend in use. So it seems %matplotlib qt will not work with animations in Jupyter.

I'm using python 2.7, matplotlib 2.0, notebook server 4.4.1.

My question is:

1. Is this behaviour expected?
2. Can someone reproduce this?
3. Does the Qt backend really not work in Jupyter with matplotlib animations?
4. And if so, does anyone know the reason for this?