GMSH quad mesh with some cells with negative weight

[tjhei]

I have a valid 2d mesh from GMSH that consists of two parts and GMSH decided to export half of the cells with a negative and half with a positive measure.

Reading it currently fails, but it works fine if we remove the AssertThrow:

dealii/source/grid/grid_tools.cc

Line 912 in 9df2fdc

AssertThrow(n_negative_cells == 0 || n_negative_cells == cells.size(),

@singima is working on a patch with a test case.

[bangerth]

That is probably relatively easy to fix for a 2d mesh in 2d space. I don't know what to do about it for a 2d mesh in 3d, or in general a 3d mesh in 3d, though.

[tjhei]

That is probably relatively easy to fix for a 2d mesh in 2d space

It is. Patch is coming soon. I am coaching @singima through that.

3d mesh in 3d

For that case we also flip all cells with negative measure already. Why do you think that is an issue?

I don't know what to do about it for a 2d mesh in 3d,

I agree. The code for checking/changing measures currently only runs for dim=spacedim. For other cases we probably fail if we encounter negative cell measures during a computation, see #13346

[bangerth]

What do we do in 3d? Just swap the upper and lower four vertices? If we already do that, much the better of course!

[tjhei]

Just swap the upper and lower four vertices?

dealii/source/grid/grid_tools.cc

Lines 889 to 896 in 9df2fdc

	else if (dim == 3)
	{
	// swap the front and back faces in 3D
	std::swap(cell.vertices[0], cell.vertices[2]);
	std::swap(cell.vertices[1], cell.vertices[3]);
	std::swap(cell.vertices[4], cell.vertices[6]);
	std::swap(cell.vertices[5], cell.vertices[7]);
	}

[agrayver]

Maybe related #3201

[tjhei]

Indeed, thank you! @singima here you have a simple test case!

[tjhei]

fixed by #13444