PyNN API

refactoring in progress on parameter_space branch - revamp of the projection/connection API

FYI: available in NEST since 2.0.0 both for original (broken) and correct numerics.

fixed: The problem turned out to be due to caching. It was not anticipated that data would be retrieved in the middle of a run (although it should have been) and so the results of getting data from file were cached. The cache is now invalidated when ... [More] run() is called (see r1245). I don't see the bug in trunk because the spike data seem to be being recorded in memory, rather than to file. [Less]

Fixed ticket:236 in 0.7 branch

status, version, description changed It is not that the neurons stop firing - recording the membrane potential demonstrates that, but that the recorder stops working. If you remove the first get_spike_counts() then the final call shows {1: 249, 2: 249}

fixed: well, getWeights() is still broken in pyNN.brian in trunk, but that didn't stop me applying and testing the patch, in the end. (see r1244)

Applied Alexandre Gravier's patch for ticket:235 in trunk

component, description changed Fixed in the 0.7 branch in r1243 There are other bugs in pyNN.brian in trunk at present which prevent me fixing it there for now.

Applied Alexandre Gravier's patch for ticket:235 in 0.7 branch

sorry, in previous comment, should be: (2) replace record_v(p1, True) by p1.record_v() (and similarly for record())