Fix PeaksInRegion edge case

[Owen Arnold]

There might be an edge case that needs fixing for the PeaksInRegion algorithm #7233. When checking that the peaks radius intersects a face, we are actually checking whether it intersects an infinite plane. This could lead to false positives.

If this does turn out to be a problem, and I'll confirm it first using a unit test, then it can be fixed by determining whether a point at the end of the line defined by the distance (already calculated), the peak centre, and the plane normal is within the bounds of that face.

[Owen Arnold]

refs #7261. Unit test confirms false positive bug.

Changeset: 758b31ee33552dbd26a1883eb1b407fb71af4e81

[Owen Arnold]

refs #7261. Fix edge case.

Changeset: adfc19931b69488c80e3c1458699b75e728e3c75

[Owen Arnold]

Tester: Follow the general test instructions given on this ticket #7233.

• Peak a peak from the file and note it's index as well as it's coordinates in QLab.
• Open the PeaksInRegion algorithm, set the CoordinateFrame to QLab and construct the extents box such that the peak is not included. Run the algorithm with CheckExtents = false, and confirm that the Intersecting column in the output workspace for that peak index gives False.
• Now experiment with setting CheckExtents to on, and providing the PeakRadius parameter. It should be logical that the Intersecting column for that peak in the output can be made true if and only if the peak passes through the box defined by the extents.

[Owen Arnold]

Branch is bugfix/7261_edge_case

[Peter Peterson]

This works as advertised, but there are some compiler warnings associated with line 157 in PeaksInRegion.cpp. Two unused variables "peakCenter" and "distance." Fix this and I'll pass the ticket.

[Owen Arnold]

refs #7261. Remove unused vars.

Changeset: c0d5d0fac0343a24e62ee39a7a1c7179f6806975

[Stuart Campbell]

This ticket has been transferred to github issue ​8107