A MR 45780 failed an automatic 21.3 sweep from master due to a difference in file content between master and 21.3 . I therefore port manually the same fix in this MR.
Below original MR description.
This is a bug-fix for TRT material layer building in FATRAS. The code should build a number of material layers (m_modelBarrelLayers) from a larger number of straw layers (nTotalBarrelLayers). It however uses a too large step (nMaterialLayerStep). Therefore an incorrect number of material layers is built.
The fix provides the correct step size, for each of TRT barrel and TRT endcaps, and issues a WARNING in case of the incorrect number of material layers.
Fix and implications were discussed at the Fast Chain meeting. The p5 of the linked slides shows a numerical example of why the original step was too large, p6 a numerical example of the fix. The bug only affects the TRT tracking geometry building for FATRAS. The TRT tracking geometry of reconstruction geometry is not affected, as this does not use the complex geometry fixed here.