Need to remove arbitrary N=1000 cutoff on number of cross-sections / weights
The cross-section code currently applies a hard limit of 1000 entries: https://gitlab.cern.ch/hepmc/HepMC3/-/blob/d939db8c73ef0b248bede26cdfe0cc1f7b3cabb2/src/GenCrossSection.cc#L68 for reasons not quite remembered: !347 (comment 7928083)
This isn't compatible with making the number of xsecs equal to the number of weights: the weights will be greater than 1000 with a single run on an NNPDF 1000-replica set, let alone other ever-increasing weight combinations. We'd better figure out what the issue was, and fix it rather than than apply an arbitrary limit.