F.CHILD in F.MAP_INPUT leads to functor compilation error (using FunTuple, probably also Moore)

When using the F.CHILD functor in a F.MAP_INPUT I seem to run into a compilation error of the functor. The interesting bit is, I think, the following:

stack/LHCb/InstallArea/x86_64_v2-centos7-gcc11-opt/include/Event/Particle.h:500:52: note:   no known conversion for argument 1 from 'LHCb::Particle* const' to 'const LHCb::Particle&'
  500 |   inline auto decayProducts( const LHCb::Particle& p ) {

meaning there is something that might need a dereferencing-if-pointer for the involved functors.

The way I got here was by using this simple functor dict:

    func_coll_vals["DTF_Jpsi_MyMassA"] = F.MAP_INPUT((F.CHILD(1, F.MASS)), DTF_jpsi.OutputRelations);
    func_coll_vals["NoDTF_Jpsi_MyMass"] = (F.CHILD(1, F.MASS))
    func_coll_vals["DTF_Jpsi_MyMassB"] = F.MAP_INPUT(F.MASS, DTF_jpsi.OutputRelations);

the first line is the one which invokes the segfault at running DaVinci, the other two are fine. I have put the full log at ~/public/segfault_davinci_17jun2022. I suspect this problem is also present for Moore.

@graven @pkoppenb