# MeshSplit Bug Example

**URL:** <https://discourse.mcneel.com/t/meshsplit-bug-example/15636>\
**Category:** Rhino for Windows\
**Created:** [January 13, 2015, 8:10pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636 "2015-01-13T20:10:50Z")\
**Posts on this page:** 11\
**Page:** 1

<div class="post-metadata">

**Author:** ![wattzie](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/wattzie/32/743497_2.png) [@wattzie](https://discourse.mcneel.com/u/wattzie)\
**Post date:** [January 13, 2015, 8:10pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/1 "2015-01-13T20:10:50Z")

</div>

Hi,

I am having some trouble accurately splitting/trimming the mesh in this file.

[MeshSplit Error.3dm](https://sea3.discourse-cdn.com/mcneel/uploads/default/15409/061c13e7afd27903.3dm) (140.3 KB)

The result does not follow the polyline at the corner, almost creating a fillet at this point. I need a corner here!

Additionally, projecting the polyline to the mesh follows this contour created by MeshSplit however placing a point at this corner and projecting it to the mesh gives an accurate result.

Does anyone know of a workaround for this?

Thanks

---

<div class="post-metadata">

**Author:** ![pascal](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/pascal/32/479395_2.png) [@pascal](https://discourse.mcneel.com/u/pascal)\
**Post date:** [January 13, 2015, 11:15pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/2 "2015-01-13T23:15:16Z")

</div>

Hi Waltzie- I see that, thanks - it is the Projection part that is messing up. Workaround: Extrude the curve through the mesh, Mesh the result, MeshIntersect to get a new curve in place and and use that curve to split the mesh with SplitMeshWithCurve (not MeshSplit) …

thanks,

-Pascal

---

<div class="post-metadata">

**Author:** ![wattzie](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/wattzie/32/743497_2.png) [@wattzie](https://discourse.mcneel.com/u/wattzie)\
**Post date:** [January 14, 2015, 2:06pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/3 "2015-01-14T14:06:54Z")

</div>

Hi Pascal,

Thanks for that. I see that it also works if I use MeshSplit with the extruded curve (surface). Is this a fluke? Is your method the most reliable way of achieving an accurate result or can I just use the surface without meshing it and intersecting the meshes? I will be placing this into a script and I want it to work every time.

Thanks for the help.

---

<div class="post-metadata">

**Author:** ![pascal](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/pascal/32/479395_2.png) [@pascal](https://discourse.mcneel.com/u/pascal)\
**Post date:** [January 14, 2015, 4:10pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/4 "2015-01-14T16:10:11Z")

</div>

Hi wattzie- I got the same discrepancy when I used the extruded curve directly in MeshSplit and had to resort to SplitMeshWithCurve - I’ll double check how that worked but I’d say this is a bug/weakness… Rhino’s mesh intersection/split/trim stuff is not all that reliable in many cases.

-Pascal

---

<div class="post-metadata">

**Author:** ![wattzie](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/wattzie/32/743497_2.png) [@wattzie](https://discourse.mcneel.com/u/wattzie)\
**Post date:** [January 14, 2015, 5:24pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/5 "2015-01-14T17:24:54Z")

</div>

> [@pascal](#):
>
> Rhino’s mesh intersection/split/trim stuff is not all that reliable in many cases.

Hi Pascal,

Is this being looked at for V6?

Thanks

---

<div class="post-metadata">

**Author:** ![pascal](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/pascal/32/479395_2.png) [@pascal](https://discourse.mcneel.com/u/pascal)\
**Post date:** [January 14, 2015, 5:28pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/6 "2015-01-14T17:28:17Z")

</div>

I am lobbying for this… yes.

-Pascal

---

<div class="post-metadata">

**Author:** ![wattzie](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/wattzie/32/743497_2.png) [@wattzie](https://discourse.mcneel.com/u/wattzie)\
**Post date:** [January 14, 2015, 5:56pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/7 "2015-01-14T17:56:41Z")

</div>

Thanks. I think that rhino is a great program however it often falls down with the mesh tools. I am using stl meshes obtained from 3d scans quite extensively in rhino and I keep encountering little bugs like this in the mesh tools in Rhino and RhinoPython.

Thanks for the help

---

<div class="post-metadata">

**Author:** ![wattzie](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/wattzie/32/743497_2.png) [@wattzie](https://discourse.mcneel.com/u/wattzie)\
**Post date:** [January 16, 2015, 3:15pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/8 "2015-01-16T15:15:04Z")

</div>

> [@pascal](#):
>
> Workaround: Extrude the curve through the mesh, Mesh the result, MeshIntersect to get a new curve in place and and use that curve to split the mesh with SplitMeshWithCurve (not MeshSplit) …

Hi Pascal,

I have been trying to implement this in python and it seems like most of these commands ( \_Mesh, \_MeshIntersect and \_SplitMeshWithCurve ) are not available in there. Do you have any idea how i could do this reliably in python?

Thanks

---

<div class="post-metadata">

**Author:** ![pascal](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/pascal/32/479395_2.png) [@pascal](https://discourse.mcneel.com/u/pascal)\
**Post date:** [January 16, 2015, 3:47pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/9 "2015-01-16T15:47:47Z")

</div>

Well, however you invoke it, the underlying code is, well, weak, so it will not be any better or worse, in terms of success, than the Rhino commands I would say.

-Pascal

---

<div class="post-metadata">

**Author:** ![bmartin](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/bmartin/32/514301_2.png) [@bmartin](https://discourse.mcneel.com/u/bmartin)\
**Post date:** [September 11, 2015, 5:39pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/10 "2015-09-11T17:39:23Z")

</div>

Hello,

Same for me. I find myself trying to implement a robust semi-automatic design workflow based on meshes from surface laser scans and the curve/mesh, mesh/mesh intersections are really the weakest point of Rhino.

Please, use [CGAL](http://www.cgal.org/) or another robust geometric library to re implement the core algorithms of your mesh objects.

---

<div class="post-metadata">

**Author:** ![bmartin](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/bmartin/32/514301_2.png) [@bmartin](https://discourse.mcneel.com/u/bmartin)\
**Post date:** [February 15, 2016, 7:23pm UTC](https://discourse.mcneel.com/t/meshsplit-bug-example/15636/11 "2016-02-15T19:23:20Z")

</div>

See

[http://discourse.mcneel.com/t/projectcurvetomesh-bug-on-triangle-edges/24760](http://discourse.mcneel.com/t/projectcurvetomesh-bug-on-triangle-edges/24760)

for the possibility of using CGAL to bypass these problems.