Vertex coordinate order changed after cull

[changliao1025]

After the cull operation, some mesh cell vertex coordinates are switched in order, which creates invalid polygons.
Examples: using both the base and culled mesh, we can plot the mesh cell using the same cell center location:

and

The test mesh files are stored at:
/compyfs/liao313/mpas/out

[changliao1025]

@xylar I shared a minimal reproducible code at: https://gist.github.com/changliao1025/939376071ead9dfc89ccc0ce27b04ed9

[xylar]

@changliao1025, thanks for reporting this problem. Sorry that it's happening!

Just to be sure I understand, is the base mesh before or after it has gone through the mesh converter? (After, I assume, so we can eliminate that that is responsible?)

[changliao1025]

@changliao1025, thanks for reporting this problem. Sorry that it's happening!

Just to be sure I understand, is the base mesh before or after it has gone through the mesh converter? (After, I assume, so we can eliminate that that is responsible?)

The base mesh was produced using the mesh converter, so it is not the JIGSAW triangular.

[changliao1025]

There are a total of 189 incorrect cells in the mesh, and they are (possibly) all located on the coastline:

For example: