Including the search for missed primary hits during HGTD reco, adding info to track decorations

Change to the HGTD reco output due to an oversight in the initial implementation. During the reco the modules close to the point of extrapolation are now searched for truth matched primary hits that might have been missed in the reconstruction, leading to a special class of misreconstruction. Improved reconstruction algorithm are expected to (at least partly) recover these, so we want to know how frequent this happens.

These changes should probably go to master as well.

Tagging @wyswys @cohm @dshope @vraskina @tawang @tadej @tstreble