WIP: iterable PVs
Main goal: allow PV MC checking with matching by tracks.
links:
- new data layout LHCb!2257 (closed)
- updates to brunel Brunel!923 (closed) (for reference as discussed in the discussion there)
- updates to moore Moore!446 (closed)
This MR updates the TrackBeamLineVertexFinderSOA (TBL) to store to output PVs references from which PVs they were built, for forward and backward tracks. The internal data structure of the TBL already holds a signed integer for reference keeping. The so-far unused range of strictly negative numbers is (ab)used for that and references are restored to normal indices when preparing the output data.
This MR furthermore introduces a converter from these new PVs to v1 RecVertices, assuming the tracks from which the PVs have been reconstructed have already been converted to Track v1s. This means that the PV MC checking algorithm can continue using the v1 rec vertex design. (presumably not a hot code path).
In my local tests this MR does introduce a small slowdown. (TODO: copy and paste the numbers together)
NB: seems some clang-format managed to sneak some whitespace changes in … be sure to view with whitespace changes disabled and/or unfold the changes to TBL.