jmarrec
commented
11 months ago @macumber thanks for the fix. I wonder if we shouldn't just erase these edges in the polyhedron instead of retaining them and putting them being shared twice to the same surface. Have you tried it by any chance?
Edit: I toyed with the idea of removing them directly in Surface3d 's ctor, but that doesn't work because the Surface3d isn't convex in that case
Tried implementation, at end of current Ctor:
// #5002 - special case to find and remove edges that are used to "cut" in to a surface to remove an interior hole
const bool keep_edge = true;
std::vector<bool> mask = std::vector<bool>(edges.size(), keep_edge);
for (size_t i = 0; i < edges.size() - 1; ++i) {
for (size_t j = i + 1; j < edges.size(); ++j) {
if (edges[i].reverseEqual(edges[j])) {
mask.at(i) = false;
mask.at(j) = false;
}
}
}
edges.erase(std::remove_if(edges.begin(), edges.end(), [this, &mask](const auto& edge) { return mask.at(&edge - edges.data()); }), edges.end());
ci-commercialbuildings
Pull request overview
Supports polygons with interior holes cut out with reverse winding, this fixes a regression from 3.6.0 that was discovered in 3.7.0-rc1.
Pull Request Author
No API or IDD changes needed
Labels:
IDDChangeAPIChangePull Request - Ready for CIso that CI builds your PRReview Checklist
This will not be exhaustively relevant to every PR.