Readout Geometry local position revisions

As discussed in ATLSWUPGR-103, some inconsistencies were found in the methods for converting between local positions and SiCellIds. An identically named method exists for both the DetectorElement and the DetectorDesign associated with that Element, which do subtly different things; the DetectorElement poentially applies a correction for the lorentz drift and a coordinate flip, while the DetectorDesign does not.

As a result, the SCT_RDOAnalysis suffered some unexpected effects where conversions between positions and cells became inconsistent, resulting in RDOs being plotted in incorrect global positions.

To address this, the following changes were made:

• Changing SolidStateDetectorElementBase::localPositionOfCell to replicated the functionality of SolidStateDetectorElementBase::rawLocalPositionOfCell so that it is consistent with XXXDesign::localPositionOfCell (which is what is called by rawLocalPositionOfCell in any case)
• adding SolidStateDetectorElementBase::correctedLocalPositionOfCell to reproduce the previous behaviour of SolidStateDetectorElementBase::localPositionOfCell
• Migrating the previous uses of SolidStateDetectorElementBase::rawLocalPositionOfCell to SolidStateDetectorElementBase::localPositionOfCell
• Migrating the (limited) uses of SolidStateDetectorElementBase::localPositionOfCell to use SolidStateDetectorElementBase::correctedLocalPositionOfCell (with the exception of the previously mentioned SCT_RDOAnalysis)

Some of the small number of cases migrated to use SolidStateDetectorElementBase::correctedLocalPositionOfCell may in fact more correctly use SolidStateDetectorElementBase::localPositionOfCell (i.e. without lorentz correction), and these have been left consistent but commented.

NB some FTK classes migrated have warnings from some bitwise operations - I have not tried to fix these since they are a) complicated and b) not needed in future.

FYI @asalzbur @cohm @dshope @ncalace @tstreble @hhayward @tlari