Don't vertex correct PFOs when vertex is outside tracker
This adds a requirement to the PFO vertex correction that the primary vertex being corrected to must not be outside the Inner Detector volume (otherwise no correction is applied). This was prompted by an issue reported in ATLJETMET-1477 in 5 TeV data, where some pathological vertices were reconstructed outside of the tracker (or even outside the calorimeter), leading to ill-defined 4-momenta for the "corrected" clusters, e.g. when the cluster is closer to the coordinate system origin than the vertex.
Users who may be interested in real (very) displaced vertices should not be affected by this; they wouldn't use the standard vertex-corrected PFOs for anything as far as I'm aware.