Hello all, i made a little "thing" (attached below). If i press F5 i get it rendered with no problem. By pressing F6 i get : ERROR: CGAL error in CGAL_Nef_polyhedron3(): CGAL ERROR: assertion violation! Expr: ss_plane.has_on(sv_prev->point()) File: /usr/include/CGAL/Nef_3/polygon_mesh_to_nef_3.h Line: 294 i am using OpenSCAD version 2017.12.21 sometimes it will be rendered, but without the polyhedron. can you help me, what i did wrong ? /* * Thing starts here */ translate([0,35,0]) bodenprofil(); bodenprofil(); translate([3+4,0,5+3])cube([6,40,3]); translate([40-7,0,5+3])cube([6,40,3]); translate([3+5,18,5+3])cube([4,4,60]); translate([3+4.25,16,5+3+60]) difference() { cube([5.5,8,7]); translate([2,-1,2]) cube([1.5,10,6]); } module bodenprofil() { translate([0,5,0]) rotate([90,0,0]) linear_extrude(height = 5) polygon([[0,0],[3,0],[3,5],[7,5],[7,8],[18,8],[18,5],[28,5],[28,8],[39,8],[39,5],[43,5],[43,0],[46,0],[46,11],[0,11]]); } polyhedron([ [39-5,0,11], [43-5,0,11], [43-5,5,11], [39-5,5,11], [12,18,58-5], [12,18,58], [12,22,58], [12,22,58-5] ], [ [0,1,2,3], // bottom [4,5,1,0], // front [7,6,5,4], // top [5,6,2,1], // right [6,7,3,2], // back [7,4,0,3]] ); polyhedron([ [39-5,0+35,11], [43-5,0+35,11], [43-5,5+35,11], [39-5,5+35,11], [12,18,58-5], [12,18,58], [12,22,58], [12,22,58-5] ], [ [0,1,2,3], // bottom [4,5,1,0], // front [7,6,5,4], // top [5,6,2,1], // right [6,7,3,2], // back [7,4,0,3]] ); /* * Thing stops here */ best regards corpsman

I had no problem to render (F6) your model with snapshot version 2018.01.06. The "front" facets of your polyhedra are not planar and that could be the reason of CGAL complaints for version 2017.12.21. Either install the new snapshot version or triangulate those facets.

I think a new snapshot should fix this. If not, please open a new issue. -Marius

my fault, i do not run Windows, (i tested it on Windows, there it works with the version from 6.01.) is there a way to get the Linux version from 6.01 ? to double check? i am using Ubuntu Mate64 and tested it with the nightly build from tonight, see attachment. - Corpsman On 01/15/2018 09:05 PM, Uwe Schächterle wrote: > Hello all, > > i made a little "thing" (attached below). > > If i press F5 i get it rendered with no problem. By pressing F6 i get : > > ERROR: CGAL error in CGAL_Nef_polyhedron3(): CGAL ERROR: assertion > violation! Expr: ss_plane.has_on(sv_prev->point()) File: > /usr/include/CGAL/Nef_3/polygon_mesh_to_nef_3.h Line: 294 > > i am using OpenSCAD version 2017.12.21 > > sometimes it will be rendered, but without the polyhedron. > > can you help me, what i did wrong ? > > > /* > * Thing starts here > */ > translate([0,35,0]) bodenprofil(); > bodenprofil(); > > translate([3+4,0,5+3])cube([6,40,3]); > translate([40-7,0,5+3])cube([6,40,3]); > > translate([3+5,18,5+3])cube([4,4,60]); > > translate([3+4.25,16,5+3+60]) > difference() > { > cube([5.5,8,7]); > translate([2,-1,2]) cube([1.5,10,6]); > } > > module bodenprofil() > { > translate([0,5,0]) rotate([90,0,0]) linear_extrude(height = 5) > polygon([[0,0],[3,0],[3,5],[7,5],[7,8],[18,8],[18,5],[28,5],[28,8],[39,8],[39,5],[43,5],[43,0],[46,0],[46,11],[0,11]]); > } > > > polyhedron([ > [39-5,0,11], > [43-5,0,11], > [43-5,5,11], > [39-5,5,11], > > [12,18,58-5], > [12,18,58], > [12,22,58], > [12,22,58-5] > ], > [ > [0,1,2,3], // bottom > [4,5,1,0], // front > [7,6,5,4], // top > [5,6,2,1], // right > [6,7,3,2], // back > [7,4,0,3]] > ); > > polyhedron([ > [39-5,0+35,11], > [43-5,0+35,11], > [43-5,5+35,11], > [39-5,5+35,11], > > [12,18,58-5], > [12,18,58], > [12,22,58], > [12,22,58-5] > ], > [ > [0,1,2,3], // bottom > [4,5,1,0], // front > [7,6,5,4], // top > [5,6,2,1], // right > [6,7,3,2], // back > [7,4,0,3]] > ); > /* > * Thing stops here > */ > > best regards corpsman > > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > -- Auf meiner Homepage www.Corpsman.de ist immer was los, ständig wird sie aktualisiert und erweitert. Da ist für jeden was dabei. -- Schütze deine Privatsphäre, nutze E-mail Verschlüsselung. Wie das geht steht z.B. hier : https://support.mozilla.org/en-US/kb/digitally-signing-and-encrypting-messages#w_installing-gpg-and-enigmail

To inform the ones that are interested in. The Problem under Linux still exists, but i found a workaround. if you write hull() { } around the polyhedron then F6 is possible under Linux too. On 01/16/2018 12:32 PM, Uwe Schächterle wrote: > my fault, i do not run Windows, (i tested it on Windows, there it works > with the version from 6.01.) is there a way to get the Linux version > from 6.01 ? to double check? > > i am using Ubuntu Mate64 and tested it with the nightly build from > tonight, see attachment. > - Corpsman > > > > On 01/15/2018 09:05 PM, Uwe Schächterle wrote: >> Hello all, >> >> i made a little "thing" (attached below). >> >> If i press F5 i get it rendered with no problem. By pressing F6 i get : >> >> ERROR: CGAL error in CGAL_Nef_polyhedron3(): CGAL ERROR: assertion >> violation! Expr: ss_plane.has_on(sv_prev->point()) File: >> /usr/include/CGAL/Nef_3/polygon_mesh_to_nef_3.h Line: 294 >> >> i am using OpenSCAD version 2017.12.21 >> >> sometimes it will be rendered, but without the polyhedron. >> >> can you help me, what i did wrong ? >> >> >> /* >> * Thing starts here >> */ >> translate([0,35,0]) bodenprofil(); >> bodenprofil(); >> >> translate([3+4,0,5+3])cube([6,40,3]); >> translate([40-7,0,5+3])cube([6,40,3]); >> >> translate([3+5,18,5+3])cube([4,4,60]); >> >> translate([3+4.25,16,5+3+60]) >> difference() >> { >> cube([5.5,8,7]); >> translate([2,-1,2]) cube([1.5,10,6]); >> } >> >> module bodenprofil() >> { >> translate([0,5,0]) rotate([90,0,0]) linear_extrude(height = 5) >> polygon([[0,0],[3,0],[3,5],[7,5],[7,8],[18,8],[18,5],[28,5],[28,8],[39,8],[39,5],[43,5],[43,0],[46,0],[46,11],[0,11]]); >> } >> >> >> polyhedron([ >> [39-5,0,11], >> [43-5,0,11], >> [43-5,5,11], >> [39-5,5,11], >> >> [12,18,58-5], >> [12,18,58], >> [12,22,58], >> [12,22,58-5] >> ], >> [ >> [0,1,2,3], // bottom >> [4,5,1,0], // front >> [7,6,5,4], // top >> [5,6,2,1], // right >> [6,7,3,2], // back >> [7,4,0,3]] >> ); >> >> polyhedron([ >> [39-5,0+35,11], >> [43-5,0+35,11], >> [43-5,5+35,11], >> [39-5,5+35,11], >> >> [12,18,58-5], >> [12,18,58], >> [12,22,58], >> [12,22,58-5] >> ], >> [ >> [0,1,2,3], // bottom >> [4,5,1,0], // front >> [7,6,5,4], // top >> [5,6,2,1], // right >> [6,7,3,2], // back >> [7,4,0,3]] >> ); >> /* >> * Thing stops here >> */ >> >> best regards corpsman >> >> >> >> _______________________________________________ >> OpenSCAD mailing list >> Discuss@lists.openscad.org >> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >> > > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > -- Auf meiner Homepage www.Corpsman.de ist immer was los, ständig wird sie aktualisiert und erweitert. Da ist für jeden was dabei. -- Schütze deine Privatsphäre, nutze E-mail Verschlüsselung. Wie das geht steht z.B. hier : https://support.mozilla.org/en-US/kb/digitally-signing-and-encrypting-messages#w_installing-gpg-and-enigmail

​Interesting workaround! It is "reasonable" since the hull operation depends only on the vertex positions. So, missing facets in a polyhedron are irrelevant. For instance, this code renders fine: intersection(){ hull(){ polyhedron( points=[[0,0,0],[0,10,0],[10,10,0], [0,0,10],[0,10,10],[10,10,10]], faces =[[0,1,2],[3,4,5]]); } sphere(8); } ​If you drop the hull, CGAL complaints. 2018-01-17 14:10 GMT-02:00 Uwe Schächterle <corpsman@corpsman.de>: > To inform the ones that are interested in. > > The Problem under Linux still exists, but i found a workaround. > > if you write > > hull() > { > } > > around the polyhedron then F6 is possible under Linux too. >

Here is another interesting case. The hull is a tetrahedron but the inside polyhedron was defined with just one face ( a simple list of all vertices). It is surprising that the preview of the polyhedron alone is correct but the preview of the intersection with the sphere is void. intersection(){ hull(){ polyhedron( points=[[0,0,0],[0,10,0],[10,10,0],[0,0,10]], faces =[[0,1,3,2]]); } sphere(8); // the preview is correct when this line is commented } The same render comes to light for any order of vertices in the list faces. 2018-01-17 16:06 GMT-02:00 Ronaldo Persiano <rcmpersiano@gmail.com>: > > ​Interesting workaround! It is "reasonable" since the hull operation > depends only on the vertex positions. So, missing facets in a polyhedron > are irrelevant. > For instance, this code renders fine: > > intersection(){ > hull(){ > polyhedron( points=[[0,0,0],[0,10,0],[10,10,0], > [0,0,10],[0,10,10],[10,10,10]], > faces =[[0,1,2],[3,4,5]]); > } > sphere(8); > } > > > ​If you drop the hull, CGAL complaints. > > 2018-01-17 14:10 GMT-02:00 Uwe Schächterle <corpsman@corpsman.de>: > >> To inform the ones that are interested in. >> >> The Problem under Linux still exists, but i found a workaround. >> >> if you write >> >> hull() >> { >> } >> >> around the polyhedron then F6 is possible under Linux too. >> >

> On Jan 17, 2018, at 5:34 PM, Ronaldo Persiano <rcmpersiano@gmail.com> wrote: > > It is surprising that the preview of the polyhedron alone is correct but the preview of the intersection with the sphere is void. > Both previews should work, and do work here. This sounds a bit like a rendering/driver issue. -Marius

On 2018-01-17 23:34, Ronaldo Persiano wrote: > Here is another interesting case. The hull is a tetrahedron but the > inside polyhedron was defined with just one face ( a simple list of > all vertices). It is surprising that the preview of the polyhedron > alone is correct but the preview of the intersection with the sphere > is void. > >> intersection(){ >> hull(){ >> polyhedron( points=[[0,0,0],[0,10,0],[10,10,0],[0,0,10]], >> faces =[[0,1,3,2]]); >> } >> sphere(8); // the preview is correct when this line is commented >> } Since faces are of no interest in the context of hull() this works. But the single face is totally warped, i.e. points not in the same plane. According to https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/The_OpenSCAD_Language#polyhedron this should be an error: "Points which describe a single face must all be on the same plane." Perhaps one could also say that a polyhedron with fewer than 4 faces is always an error, perhaps with the exception of zero faces in the case of using it within hull()? Carsten Arnholm

cacb wrote > Perhaps one could also say that a polyhedron with fewer than 4 faces is > always an error, perhaps with the exception of zero faces in the case of > using it within hull()? I am heavily using this feature - to preview 2D-polygons placed in 3D space, - to see, what sweep() will do with my code - to check for selfintersections Indeed I even don't see much sense in restricting STL output to "proper" manifolds. Is it part of the format specification? -- Sent from: http://forum.openscad.org/