Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Fix SpikeMonitor for subgroups (weave, cython, cpp_standalone) #454
For subgroups not starting at 0, the code for searching through the list of all spikes for spikes that should be recorded was not correct. If some neuron in the full group spiked, but none of them >= the start index of the recorded subgroup, then all (instead of none) spikes were recorded. The good news is that this should not have led to undetected incorrect recordings because 1) it would have recorded spikes with negative indices, which would be incorrect in an obvious way and 2) it most of the time should have led to a segmentation fault due to accessing the
I'm going to merge this now. One of the travis test suites stalled for some reason, but since this is a single commit on top of current master, the two tests suites ran on the identical code and are therefore redundant. AppVeyor failed because of the standalone filename issue that we discussed, it will be fixed in a separate pull request (it also timed out, though...).