eventplot raises an exception for empty sequences

[jgosmann]

plt.eventplot([[]], colors=[(0.0, 0.0, 0.0, 0.0)])
will fail with the following exception:

/home/jgosmann/virt_env/default/local/lib/python2.7/site-packages/matplotlib-1.3.1-py2.7-linux-x86_64.egg/matplotlib/pyplot.pyc in eventplot(positions, orientation, lineoffsets, linelengths, linewidths, colors, linestyles, hold, **kwargs)
   2718                            lineoffsets=lineoffsets, linelengths=linelengths,
   2719                            linewidths=linewidths, colors=colors,
-> 2720                            linestyles=linestyles, **kwargs)
   2721         draw_if_interactive()
   2722     finally:

/home/jgosmann/virt_env/default/local/lib/python2.7/site-packages/matplotlib-1.3.1-py2.7-linux-x86_64.egg/matplotlib/axes.pyc in eventplot(self, positions, orientation, lineoffsets, linelengths, linewidths, colors, linestyles, **kwargs)
   3987 
   3988         if len(positions) > 0:
-> 3989             minpos = min(position.min() for position in positions)
   3990             maxpos = max(position.max() for position in positions)
   3991 

/home/jgosmann/virt_env/default/local/lib/python2.7/site-packages/matplotlib-1.3.1-py2.7-linux-x86_64.egg/matplotlib/axes.pyc in <genexpr>((position,))
   3987 
   3988         if len(positions) > 0:
-> 3989             minpos = min(position.min() for position in positions)
   3990             maxpos = max(position.max() for position in positions)
   3991 

/home/jgosmann/virt_env/default/local/lib/python2.7/site-packages/numpy/core/_methods.pyc in _amin(a, axis, out, keepdims)
     19 def _amin(a, axis=None, out=None, keepdims=False):
     20     return um.minimum.reduce(a, axis=axis,
---> 21                             out=out, keepdims=keepdims)
     22 
     23 def _sum(a, axis=None, dtype=None, out=None, keepdims=False):

ValueError: zero-size array to reduction operation minimum which has no identity

I think, this case should be handled more gracefully. Having a timeslice without any events can be a quite standard thing depending on the analysis one performs. Having to handle this special case always outside the function before plotting can be quite annoying (and one has also take care of labels matching after removing empty sequences out of the list).