MCTupleToolDalitz crashes with photons

If a decay descriptor is used that optionally matches photons (=>) MCTupleToolDalitz considers the photon a stable particle, when it finds one, and tries to create columns in the ntuple with pairs involving the photon. If this happens in the first event, fine, but if it doesn't and then happens at a later event, I think this causes the unexpected column to be attempted and a nasty seg-fault to occur.

The "problem" call seems to be line 18 https://gitlab.cern.ch/lhcb/Analysis/-/blob/run2-patches/Phys/DecayTreeTupleDalitz/src/TupleToolDalitz.icpp#L18

Can we simply exclude photons from TupleToolDalitz, or add a flag to allow their treatment only in well-controlled cases (e.g. when ->, not =>, is used)?

Assigning to @pkoppenb just because I found his name in MCTupleToolDalitz.cpp, but feel free to decline responsibility ;)

Also tagging @dathomps and @cburr whose analysis productions randomly failed because of this (if I've understood the situation correctly)