[menes97]

Fixes #58

NwcMeshClipper.Clip tested the mesh's local vertices against section planes given in world space, and kept the original Transform on the result. A translated or rotated mesh was therefore cut in the wrong place. This affects the IFC export with SectionPlanes/SectionBox and ClipGeometryWithSectionPlanes.

I took the first option from the issue: when the mesh has a Transform, each active plane is mapped into the mesh's local space before the cut. For a world position x = M·l + t, using the same matrix as GetWorldVertices, the plane n·x = d becomes (Mᵀn)·l = d − n·t, normalized again. The result keeps its local vertices and its Transform, so the IFC export with WorldCoordinates = false stays consistent. If a transform yields a degenerate or non-finite plane, the plane is used unchanged, the same fallback GetWorldVertices uses for a transform it can't apply. Meshes without a Transform are clipped exactly as before.

Tests (SectioningTests):

• NwcMeshClipper_Clip_TranslatedMesh_CutsInWorldSpace: the case from the issue. A triangle moved to world Z = 100 is kept whole by CreateBottom(50) and removed by CreateTop(50). The kept triangle keeps its Transform, and its world vertices are still at Z = 100.
• NwcMeshClipper_Clip_RotatedMesh_CutsInWorldSpace: a quarter turn about Z plus a translation puts a triangle at world Y 2..4. CreateFront(3) keeps exactly its Y ≥ 3 part, with world vertices at Y = 3 and Y = 4.

Both fail without the change. dotnet build NwcIO.slnx succeeds, and NwcIO.Tests passes locally: 241 passed, and 2 skipped because they need private Navisworks files (.NET 10.0.400 on Windows, with global.json pinned down to it temporarily; not committed).

🤖 Generated with Claude Code