Proposed fix to segmentation fault

[kirichoi]

This is proposed fix to segmentation fault caused by fix to issue #411.
Added extra step that checks 'time' is part of the selection.

[0u812]

This fixes the segfault, but I don't think the behavior is correct for models with events. I don't think time-shifting the output row will work in general.

Antimony:

model mymodel()
  var species X = 10
  at time>=5: X = 5
end

Python:

mymodel.reset()
mymodel.simulate(-5,5,100)
mymodel.plot()

[kirichoi]

I see. This requires a bit more complex solution.