@osschar showed that mkFit often gets odd number of hits from T4/T5 (no-IT/pixel LST inputs), which looks like the result of skipping the first S-side hits when filling hitsForSeed.
IIUC the TrackCandidate variant of the T4/T5 output gets the full list of hits.
I think that reinserting the dropped hit could just work (need to check with the mkFit side that somewhat incompatible hits can be dropped during a bw fit in mkFit)
@osschar showed that mkFit often gets odd number of hits from T4/T5 (no-IT/pixel LST inputs), which looks like the result of skipping the first S-side hits when filling
hitsForSeed.IIUC the
TrackCandidatevariant of the T4/T5 output gets the full list of hits.I think that reinserting the dropped hit could just work (need to check with the mkFit side that somewhat incompatible hits can be dropped during a bw fit in mkFit)