Store InDetTrackParticles in ESD/AOD new JO test job

This is realy a test if adding to ESD/AOD ItemList from reco fragments would work.

I find it a bit mouthful (the if + merge ...) and was consiering adding helper functions:

addToESD(ca, flags, collections)
addtoAOD(ca, flags, collections)

But this breaks the symmetry a bit so in the end decided not to do it. Tagging @wlampl @emoyse @goetz @tadej @jchapman ... for an opinion.

The place where it is done is maybe not optimal. I.e. I was thinking about putting it in TrackParticleCnvAlgCfg. One thing that caused a lot of wasted time was that the TrackParticleCnvAlg converter silently ignored missing inputs (and above all) consequently failed to produce promised output. I have modified it to produce errors now. FI @goetz @mvozak

Side note, the handles are perfectly capable of declaring the fact that they are "empty". Like this:

m_readHandleKey.initialize(SG::AllowEmpty)

instead:

m_readHandleKey.initialize(m_shoulReadThis).

We do not need an extra boolean as it requires additional effort to keep it consistent with the keys.