t_start is not set correctly when running simulation in multiple steps

[jakobj]

when trying to run a simulation in multiple steps via a for loop:

for i in xrange(1000):
    sim.run(100)

pyNN does not set t_start properly, which causes an error when trying to retrieve spikes:

Traceback (most recent call last):
  File "test_loop.py", line 9, in <module>
    print n.get_data('spikes',clear=True).segments[0].spiketrains
  File "/home/jordan/opt/pyNN.0.8.install/lib/python2.7/site-packages/pyNN/common/populations.py", line 474, in get_data
    return self.recorder.get(variables, gather, self._record_filter, clear)
  File "/home/jordan/opt/pyNN.0.8.install/lib/python2.7/site-packages/pyNN/recording/__init__.py", line 270, in get
    data.segments.append(self._get_current_segment(filter_ids=filter_ids, variables=variables, clear=clear))
  File "/home/jordan/opt/pyNN.0.8.install/lib/python2.7/site-packages/pyNN/recording/__init__.py", line 234, in _get_current_segment
    for id in sorted(self.filter_recorded('spikes', filter_ids))]
  File "/home/jordan/opt/neo/lib/python2.7/site-packages/neo/core/spiketrain.py", line 219, in __new__
    _check_time_in_range(obj, obj.t_start, obj.t_stop, view=True)
  File "/home/jordan/opt/neo/lib/python2.7/site-packages/neo/core/spiketrain.py", line 46, in _check_time_in_range
    (value, t_start))
ValueError: The first spike ([ 700.02357704  711.3158928   722.60820856  733.90052432  745.19284007
  756.48515583  767.77747159  779.06978735  790.36210311]) is before t_start (700.1)

[apdavison]

fixed in eeec7e7