Bug in CalculateCoverageDGS

[AndreiSavici]

There is a problem when using the planning tool for direct geometry inelastic. In particular, the trajectories do not correspond to results from ConvertToMD if one uses projections
Here is the results from ConvertToMD

The script to generate it is:

Load(Filename='CNCS_7860_event.nxs', OutputWorkspace='CNCS_7860_event')
ConvertUnits(InputWorkspace='CNCS_7860_event', OutputWorkspace='CNCS_7860_event', Target='DeltaE', EMode='Direct', EFixed=3)
SetUB(Workspace='CNCS_7860_event', a=10, b=10, c=10)
SetGoniometer(Workspace='CNCS_7860_event', Axis0='30,0,1,0,1')
ConvertToMD(InputWorkspace='CNCS_7860_event', QDimensions='Q3D', Q3DFrames='HKL', QConversionScales='HKL', Uproj='1,1,0', Vproj='1,-1,0', Wproj='0,0,1', OutputWorkspace='md', MinValues='-4,-4,-4,-4', MaxValues='4,4,4,4')
BinMD(InputWorkspace='md', AxisAligned=False, BasisVector0='[H,H,0],in 0.889 A^-1,1,0,0,0', BasisVector1='[H,-H,0],in 0.889 A^-1,0,1,0,0', BasisVector2='[0,0,L],in 0.628 A^-1,0,0,1,0', BasisVector3='DeltaE,DeltaE,0,0,0,1', OutputExtents='-2.4,2.4,-0.8,4,-0.1,0.1,-0.1,0.1', OutputBins='100,100,1,1', Parallel=True, OutputWorkspace='md_rebinned')

When using the planning tool, one can use the projection u to be (1,1,0), projection v (1,-1,0)

One can also reverse the order of projections, but still plot (HH0) along x, and (H,-H,0) along y. I would expect the same result. But what I get is

I tracked down the issue, and is the way we store matrices. In CalculateCoverageDGS.cpp I set the rows of W matrix as the projection vectors, instead of the columns (lines 295-297)