Astrocytes example fails in brian2cuda

[denisalevi]

When running the brian2 astrocyte example, I get an device-side assert:

code_objects/synapses_pre_codeobject.cu:255: void kernel_synapses_pre_codeobject(int, int, int, int, int, signed int *, int, double *, int, signed int *, int, signed int *): block: [2,0,0], thread: [31,0,0] Assertion `spiking_neuron_idx == eventspace[neurongroup_size]` fai
led.

Could be that the synapses to synapses implementation is not implemented correctly for SynpaticPathway with Synapses source? See brian-team/brian2#1287.

Synapses to synapses support was added in #186, closing issue #133.

[mstimberg]

I think this can be closed, example_2_gchi_astrocyte.py works on my machine, and all the other astrocyte examples work as well except for example_4_synrel.py which fails with an error related to the number of neurons recorded. The error message refers to #50, but this issue has been about several problems with StateMonitor and is now closed.

Oh, and note that the plotting code for the examples fails for recent matplotlib versions, but that's not Brian2CUDA's fault 😉

[denisalevi]

Great, I can confirm its working now. Thanks!

[denisalevi]

Oh and the StateMonitor is failing because of #201. I will fix that one next.