blueprint: negative volumes in sub-divided meshes from polytopal inputs

[cyrush]

Describe the bug

See details in: https://github.com/Alpine-DAV/ascent/issues/1261

We need to explore what cases we can avoid negative volumes in VisIt's expression calc.

Helpful additional information

• Did VisIt crash: nope
• Did you get wrong results: yes

To Reproduce

Using blueprint_v0.8.2_polytess_test_data/polyess_3d_example_hdf5.root, plot mesh_quality/volume

Expected behavior

Positive volumes.

Desktop

• OS and version: [e.g. Ubuntu 14.04]
• VisIt Version: 3.4.1
• Server info (if applicable):

[biagas]

I think there is possible value in allowing negative volumes, they can point out problems with connectivity information. There is already the ability to tell avtVMetricVolume to only return positive volumes (it just converts the sign of the result). avtWeightedVariableSummationQuery sets this to true. Perhaps this should be exposed as a user-settable option?

[markcmiller86]

In Silo db plugin, I think there is a tad of logic to test the first encountering each type of new zoo type element in a UCD mesh and check whether the volume is positive or negative. Negative indicates an issue with node ordering and when we encounter negative volume in first zone of a given type, we decide to reverse ordering of all zones of that type...hmmm...maybe we do that only for tets.

[JustinPrivitera]

@cyrush does this mean polytess is wrong? Node ordering is incorrect for the cubes, perhaps?

[cyrush]

related: should we look into our verdict options in this day in age?