3DPrinting/PrusaSlicer-NonPlainar
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Another attempt to fix unit test for 2D convex hull of sinking object on RasperryPi

Browse Source
This commit is contained in:
enricoturri1966 2021-05-21 15:37:25 +02:00
parent 49f29e8353
commit dfa85c2c37
1 changed files with 10 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
tests/libslic3r/test_3mf.cpp
View File

@ -110,13 +110,17 @@ SCENARIO("2D convex hull of sinking object", "[3mf]") {
{ -4242640, -8537523 }
};
bool res = hull_2d.points == result;
std::cout << "hull_2d vertices count: " << hull_2d.points.size() << "\n";
std::cout << "hull_2d vertices:\n";
for (size_t i = 0; i < hull_2d.points.size(); ++i) {
std::cout << hull_2d.points[i].x() << ", " << hull_2d.points[i].y() << "\n";
bool res = hull_2d.points.size() == result.size();
if (res) {
for (size_t i = 0; i < result.size(); ++i) {
Vec3d hull_p(unscale<double>(hull_2d.points[i].x()), unscale<double>(hull_2d.points[i].y()), 0.0);
Vec3d res_p(unscale<double>(result[i].x()), unscale<double>(result[i].y()), 0.0);
res &= res_p.isApprox(hull_p);
}
}
// this does not work on RaspberryPi for float precision problem: the x of 1st vertex ending up being -4242640 instead of -4242641
// bool res = hull_2d.points == result;
THEN("2D convex hull should match with reference") {
REQUIRE(res);