Plot values sometimes jump up/down at end about nengo-1.4

[tbekolay]

[Comment migrated from SourceForge | Submitted by 'bryan_tripp']

This was happening in plots of unfiltered sums of spikes, during very short timesteps at the end of a simulation. Sometimes after the expected number of time steps, accumulated rounding errors made the time slightly less than the specified end time, so a very small step was taken to make up the difference. Typically no spikes or very few spikes occurred during this time, so the value of unfiltered sum of spikes was zero. This is basically a cosmetic problem. It has been resolved by stretching the last time step rather than taking an additional time step, in LocalSimulator.