Remove "inhibitory" and "excitatory" spiking input port specifiers

[clinssen]

Depends on #938.

"Excitatory" and "inhibitory" input port specifiers are confusing and should be eliminated.

Current sitation: if specifiers are present, spikes get "routed" to the right buffer depending on weight of the incoming event.

exc_spikes <- excitatory spike
inh_spikes <- inhibitory spike

(this example: actually only one in-port (one rport)

Behaviour is such that:

w < 0:
	gets routed to inh_spikes
w > 0:
	gets routed to exc_spikes

Proposal: change to multisynapse model (behavioural/API change), or preserve current situation by putting the check for the sign of the weight inside the onEvent() handler function.

in_spikes <- spikes

onEvent(in_spikes):
	if in_spikes < 0:
		g_syn_inh += 1
	else:
		g_syn_exc += 1