compilation fails when using batch sizes larger than 1 in combination with is_prev_post_spike_time_required=True

[FabianSchubert]

If the model contains a weight update class that uses
is_prev_post_spike_time_required=True and model.batch_size is set to any number larger than 1, building the model fails with the error class "MergedNeuronPrevSpikeTimeUpdateGroup0" has no member "numNeurons"
A code example (any value of NBATCH larger 1 produces the error):

import numpy as np
from pygenn.genn_model import (GeNNModel,
                               create_custom_weight_update_class,
                               init_var)

NBATCH = 10
R_TARG = 1.0

T = 100.0
DT = 0.1
NT = int(T/DT)

cs_wu = create_custom_weight_update_class(
        "cs_wu_class",
        param_names=["mu"],
        var_name_types=[("g", "scalar")],
        sim_code="$(addToInSyn, $(g));",
        learn_post_code=f"""
            $(g) += $(mu)*(
                {R_TARG}*($(t)-$(prev_sT_post))
                - 1.0
            );
        """,
        is_prev_post_spike_time_required=True
        )

model = GeNNModel("float", "test")
model.batch_size = NBATCH

poiss_pop = model.add_neuron_population(
        "poiss_pop",
        10, "Poisson",
        {"trefract": 0.0,
         "tspike":0.0,
         "Vspike":1.0,
         "Vrest":0.0},
        {"V":0.0, "spikeTime": 0.0}
        )

lif_pop = model.add_neuron_population(
        "lif_pop",
        10, "LIF",
        {"C": 1.0,
        "Ioffset": 0.0,
        "TauM": 1.0,
        "TauRefrac": 0.0,
        "Vreset": 0.0,
        "Vrest": 0.0,
        "Vthresh": 1.0},
        {"RefracTime": 0.0, "V": 0.0}
        )

syn_pop = model.add_synapse_population(
        "syn_pop", "DENSE_INDIVIDUALG", 0,
        "poiss_pop", "lif_pop",
        cs_wu, {"mu": 0.01},
        {"g": init_var("Uniform",{"min":0.0,"max":0.0})},
        {}, {},
        "DeltaCurr", {}, {})

model.build()

poiss_pop.set_extra_global_param("firingProb", np.random.rand(10))
poiss_pop.set_extra_global_param("offset", 0)

model.load()

for t in range(NT):
    model.step_time()

print("done")