difference() problem with extruded polygon in CGAL (F6) mode

[Harvie]

This renders properly with F5, but not with F6. IMHO there are no conflicting verticles, etc...

module pagoda(scale=1, height=1, ratio=2, center=true) {
    scale([scale,scale])
    linear_extrude(height=height, center=center)
    polygon(points=[[-1,1],[1,1],[1*ratio,-1],[-1*ratio,-1]], paths=[[0,1,2,3]]);
}

module male() {
    difference() {
        pagoda(ratio=2, scale=0.9);     
        translate([0,-0.21]) pagoda(ratio=2, scale=0.7, height=1.1);
    }
}

male();

Parsing design (AST generation)...
Compiling design (CSG Tree generation)...
Compilation finished.
Rendering Polygon Mesh using CGAL...
Number of vertices currently in CGAL cache: 926
Number of objects currently in CGAL cache: 99
Top level object is a 3D object:
Simple: yes
Valid: yes
Vertices: 8
Halfedges: 24
Edges: 12
Halffacets: 12
Facets: 6
Volumes: 2
Total rendering time: 0 hours, 0 minutes, 0 seconds
Rendering finished.

[Harvie]

update: i've tried latest version and git version. actually i've got two modules (male() and female()), male() suddenly started to work, but female() doesn't and in latest git (saying: OpenSCAD 2011.05.27) it's just vice-versa (male() doesn't work again, but female() does).

module pagoda(scale=1, height=1, ratio=2, center=true) {
    scale([scale,scale])
    linear_extrude(height=height, center=center)
    polygon(points=[[-1,1],[1,1],[1*ratio,-1],[-1*ratio,-1]], paths=[[0,1,2,3]]);
}

module female() {
    difference() {
        cube([4,3,2], center=true);
        translate([0,0.6,0]) pagoda(ratio=1.5);
    }
}

module male() {
    difference() {
        pagoda(ratio=2, scale=0.9);     
        //translate([0,-0.7]) pagoda(ratio=2);
        translate([0,-0.21]) pagoda(ratio=2, scale=0.7, height=1.1);
    }
}

//female();
male();

[Harvie]

i've restarted openscad again and both modules are working. must be some race condition or something. i don't understand...

[Harvie]

Happend again!
Design -> Flush Caches helped me.
But i still can't intentionaly reproduce after flusing (to let you know what triggers the bug)

[tinkerology]

I'm having the same or a similar problem. Here's a simple code example that shows it. Does anyone have any ideas on how to work around the problem? Flush caches does not work for me at all.

Thanks, Scott

FRAGMENTS=40;

module drawCylinder(depth, radius)
{
// cylinder(depth, radius, radius, $fn=FRAGMENTS);
linear_extrude(height = depth, center = true, convexity = 10, twist = 0)
circle(radius, $fn=FRAGMENTS);
}

module cubeWithHoles(x,y,z)
{
difference()
{
cube([x,y,z]);

    translate([x/2,y/2,z/2])
    % drawCylinder( y * 2, x * 0.4 );

    rotate([0,90,0]) 
    translate([-1 * x/2,1 * y/2,z/2])
    % drawCylinder( y * 2, x * 0.4 );
}

}

cubeWithHoles(5,5,5);

[kintel]

I believe this was fixed in ad0f9f5
Could you try again with the latest version (i.e. any build of master newer than Nov 28)?

[tinkerology]

With the 12/7/2011 Windows install, I am still seeing failures with
difference with the following test:

module fails1()

{

difference() {

cube([50,50,50]);

rotate(90, [0,0,0])

% cylinder(25, 67.5, center = true);

}

}

module fails2()

{

difference() {

cube([50,50,50]);

rotate(90, [0,0,0])

%cylinder(25, 65.5, center = true);

}

}

fails1();

translate([60,0,0])

fails2();

Any ideas?

Thanks, Scott

On Mon, Dec 12, 2011 at 2:39 PM, Marius Kintel <
reply@reply.github.com

wrote:

I believe this was fixed in ad0f9f5
Could you try again with the latest version (i.e. any build of master
newer than Nov 28)?

---

Reply to this email directly or view it on GitHub:
#10 (comment)

[kintel]

It looks good here.
Could you describe in what way this fails for you, perhaps with an illustrating screenshot?

Note the difference between the % operator and the # operator: The former makes the object not take part in a CSG operation, the latter does.

[kintel]

Any updates on this issue?
I suspect the last report to be a misunderstanding of how the % and # operators work, but it would be nice with a confirmation so we can fix or close this issue.

[tinkerology]

Thanks! I got back to testing this weekend and everything works as
expected now. I appreciate all of your hard work on OpenSCAD!

Do you have a way to donate to the project?

Thanks again, Scott

On Mon, Dec 12, 2011 at 2:39 PM, Marius Kintel <
reply@reply.github.com

wrote:

I believe this was fixed in ad0f9f5
Could you try again with the latest version (i.e. any build of master
newer than Nov 28)?

---

Reply to this email directly or view it on GitHub:
#10 (comment)

[kintel]

Right now all we've got is a Flattr account: https://flattr.com/thing/121566/OpenSCAD-The-Programmers-Solid-3D-CAD-Modeller