Can anyone suggest how to make the below model render without showing a
"WARNING: Object may not be a valid 2-manifold and may need repair! "
https://drive.google.com/uc?export=download&id=0Bxg0L-sj9zVtQUlNWVhoS2oxRzQ
p1 = [10,10,0];
p2 = [10,-10,0];
p3 = [-10,-10,0];
p4 = [-10,10,0];
p5 = [ 0,  0,10];

polyhedron(
points = [ p1, p2, p3, p4, p5],
faces = [
[0,1,4], [1,2,4], [2,3,4], [3,0,4],
//[0,1,2], [0,2,3] ,
//[0,1,2,3]
]);
translate([0,0,-2.5])
cube([20,20,5], true);

I've tried increasing or decreasing the size of cube().
Without the bottom cube, no warning is shown, but once I add the below cube,
I get the warning.

--
View this message in context: http://forum.openscad.org/Polyhedron-and-cube-Not-a-Valid-2-Manifold-tp14283.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

Try adding [0,3,2,1] to faces.

$  Runsun Pan, PhD

$ libs: doctest , faces ( git ), offline doc ( git ),runscad.py( 1 , 2 , git );

$ tips: hash( 1 , 2 ), sweep , var , lerp , animGif

--
View this message in context: http://forum.openscad.org/Polyhedron-and-cube-Not-a-Valid-2-Manifold-tp14283p14285.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

The term 'solid' means what it says. A solid object.
There are several ways to restate this that might be useful when trying to
determine "Is my object a solid?"

• No zero thickness walls. All triangles or quads etc have no thickness.
• All faces must fully connect with each other to form a continuous
  surface. All vertices must meet.
• The object is watertight:
• I.e. if filled with water there are no gaps where water might leak out.
• E.g. if you have a triangle and another two triangles meet in the
  middle of one of it's edges. It appears to be solid but that middle edge is
  not connected at the vertices to the other two. So when filled with water it
  will leak out the gap.
• Think of the faces as sheets of metal and the vertices as welds.

Check out many links online for help on what a "solid" model is.

--
View this message in context: http://forum.openscad.org/Polyhedron-and-cube-Not-a-Valid-2-Manifold-tp14283p14287.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

In another  thread
http://forum.openscad.org/3d-hull-with-2d-primitives-needed-tp5845p14276.html
I have placed a piece of code that demonstrates that nophead and Neon22's
comments are at variance with how polyhedron() actually behaves.

In that code, change any of the StartVal_ or EndVal_  parameters and see how
an ellipsoid changes into a vase. If you count the vertices or faces, there
are exactly as many as there should be, and thus the shape is watertight.
But it does not have a volume, irrespective of what OpenSCAD reports, as all
faces are of zero thickness.

But if you create an stl and export it into MeshLab, the number of vertices
triples, and the ellipsoid / vase would no longer be watertight. Whose
messages are then to be trusted?

To me, it means that someone has buggered the C++ code, either in creating
the stl mesh, or in reading it. Does anybody have an idea how to find the
bug, preferably without me first learning C++?

Wolf

--
View this message in context: http://forum.openscad.org/Polyhedron-and-cube-Not-a-Valid-2-Manifold-tp14283p14293.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

Looking at @wolf's code the vertex location calc is done 4(I think) times for
each vertex.
I.e. 4 faces join at a vertex but that vertex is calculated 4 times instead
of once.

Perhaps a way to do it where the vertices are shared, is to calculate all
the points once, in a list, and then build the faces list using those
vertices.
Then there will be no floating point error as the vertices will really be
shared. And therefore it will be a solid.

Sorry don't have time to do it right now :(

--
View this message in context: http://forum.openscad.org/Polyhedron-and-cube-Not-a-Valid-2-Manifold-tp14283p14296.html
Sent from the OpenSCAD mailing list archive at Nabble.com.