BUG: sampling: face sets: incomplete sampling when intersecting processor patches

Summary

The face sampling across a given line does not produce data points beyond the intersection point between the sampling line and a processor boundary.

Comment: There is a problem with the face-based sampling crossing odd number of boundary (processor) faces but this is a general issue.

Steps to reproduce

• Execute Allrun-parallel in the minimal test case: NACA4412.zip
• Compare the data files: output.serial/postProcessing/sample.lines/1/xbyc0.90_columnAverage_UMean__columnAverage_UPrime2Mean_.xy and output.parallel/postProcessing/sample.lines/1/xbyc0.90_columnAverage_UMean__columnAverage_UPrime2Mean_.xy
• Compare the figures: serial and parallel runs, respectively:

Environment information

• api = 2412
• HEAD = 8e98779d (!717 (merged))
• compiler = Clang (system) = clang version 15.0.7
• mpi = SYSTEMOPENMPI = mpirun (Open MPI) 1.10.7.0.5e373bf1fd
• OS = Description: openSUSE Leap 15.5

added bug label

marked this issue as related to #2067 (closed)

marked this issue as related to #1456

mentioned in merge request !717 (merged)

It has something to do with the restarting of the sampling once it hits a boundary. The sampling line enters a processor, tracks correctly to the next processor patch but then does not re-enter.