Empathy List Archives

NH

nop head

Sun, Dec 3, 2017 9:49 AM

I am getting this error from a fairly simple polyhedron when CGAL operates
on it. It looks fine in the thrown together view and I think the code used
to work in older versions of OpenSCAD. I don't remember seeing this error
before but I could be wrong about that.

ERROR: CGAL error in CGAL_Nef_polyhedron3(): CGAL ERROR: assertion
violation! Expr: ss_circle.has_on(sp) File:
/opt/mxe/usr/x86_64-w64-mingw32.static.posix/include/CGAL/Nef_3/polygon_mesh_to_nef_3.h
Line: 263

If I export it as an STL before involving CGAL and import it I can use it
with CGAL. So whatever the problem is STL export cleans it up.

polyhedron(points =

[[0, 10, 0], [10, 0, 0], [0, -10, 0], [-10, 0, 0], [0, 10, 100],

[9.80581, 0, 98.0388], [0, -10, 100], [-9.80581, 0, 101.961], [30, 10,
150], [38.5749, 0, 144.855], [30, -10, 150], [21.4251, 0, 155.145]],

faces = [[3, 2, 1, 0], [0, 1, 5, 4], [4, 5, 9, 8], [1, 2, 6, 5], [5, 6,
10, 9], [2, 3, 7, 6], [6, 7, 11, 10], [3, 0, 4, 7], [7, 4, 8, 11], [8, 9,
10, 11]]);

Perhaps it doesn't like quads because they will be triangles when read from
STL.

I am getting this error from a fairly simple polyhedron when CGAL operates on it. It looks fine in the thrown together view and I think the code used to work in older versions of OpenSCAD. I don't remember seeing this error before but I could be wrong about that. ERROR: CGAL error in CGAL_Nef_polyhedron3(): CGAL ERROR: assertion violation! Expr: ss_circle.has_on(sp) File: /opt/mxe/usr/x86_64-w64-mingw32.static.posix/include/CGAL/Nef_3/polygon_mesh_to_nef_3.h Line: 263 If I export it as an STL before involving CGAL and import it I can use it with CGAL. So whatever the problem is STL export cleans it up. polyhedron(points = [[0, 10, 0], [10, 0, 0], [0, -10, 0], [-10, 0, 0], [0, 10, 100], [9.80581, 0, 98.0388], [0, -10, 100], [-9.80581, 0, 101.961], [30, 10, 150], [38.5749, 0, 144.855], [30, -10, 150], [21.4251, 0, 155.145]], faces = [[3, 2, 1, 0], [0, 1, 5, 4], [4, 5, 9, 8], [1, 2, 6, 5], [5, 6, 10, 9], [2, 3, 7, 6], [6, 7, 11, 10], [3, 0, 4, 7], [7, 4, 8, 11], [8, 9, 10, 11]]); Perhaps it doesn't like quads because they will be triangles when read from STL.

P

Parkinbot

Sun, Dec 3, 2017 10:03 AM

On 2017.1.20 it renders without any probs

--
Sent from: http://forum.openscad.org/

On 2017.1.20 it renders without any probs -- Sent from: http://forum.openscad.org/

NH

nop head

Sun, Dec 3, 2017 10:30 AM

Will it union with a unit cube though?

On 3 December 2017 at 10:03, Parkinbot rudolf@parkinbot.com wrote:

On 2017.1.20 it renders without any probs

--
Sent from: http://forum.openscad.org/

OpenSCAD mailing list
Discuss@lists.openscad.org
http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

Will it union with a unit cube though? On 3 December 2017 at 10:03, Parkinbot <rudolf@parkinbot.com> wrote: > On 2017.1.20 it renders without any probs > > > > > -- > Sent from: http://forum.openscad.org/ > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >

P

Parkinbot

Sun, Dec 3, 2017 11:05 AM

sorry forgot to tell this. Yes!

--
Sent from: http://forum.openscad.org/

sorry forgot to tell this. Yes! -- Sent from: http://forum.openscad.org/

NH

nop head

Sun, Dec 3, 2017 11:23 AM

Thanks. So looks like a regression then as I am on Windows snapshot version
2017.11.12 (git 7cb1e93).

On 3 December 2017 at 11:05, Parkinbot rudolf@parkinbot.com wrote:

sorry forgot to tell this. Yes!

--
Sent from: http://forum.openscad.org/

OpenSCAD mailing list
Discuss@lists.openscad.org
http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

Thanks. So looks like a regression then as I am on Windows snapshot version 2017.11.12 (git 7cb1e93). On 3 December 2017 at 11:05, Parkinbot <rudolf@parkinbot.com> wrote: > sorry forgot to tell this. Yes! > > > > -- > Sent from: http://forum.openscad.org/ > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >

NH

nop head

Sun, Dec 3, 2017 12:09 PM

Seems to be this as my library info shows CGAL 4.10:
https://github.com/openscad/openscad/issues/2139

On 3 December 2017 at 11:23, nop head nop.head@gmail.com wrote:

Thanks. So looks like a regression then as I am on Windows snapshot
version 2017.11.12 (git 7cb1e93).

On 3 December 2017 at 11:05, Parkinbot rudolf@parkinbot.com wrote:

sorry forgot to tell this. Yes!

--
Sent from: http://forum.openscad.org/

OpenSCAD mailing list
Discuss@lists.openscad.org
http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

Seems to be this as my library info shows CGAL 4.10: https://github.com/openscad/openscad/issues/2139 On 3 December 2017 at 11:23, nop head <nop.head@gmail.com> wrote: > Thanks. So looks like a regression then as I am on Windows snapshot > version 2017.11.12 (git 7cb1e93). > > > > On 3 December 2017 at 11:05, Parkinbot <rudolf@parkinbot.com> wrote: > >> sorry forgot to tell this. Yes! >> >> >> >> -- >> Sent from: http://forum.openscad.org/ >> >> _______________________________________________ >> OpenSCAD mailing list >> Discuss@lists.openscad.org >> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >> > >

MK

Marius Kintel

Sun, Dec 3, 2017 5:50 PM

On Dec 3, 2017, at 7:09 AM, nop head nop.head@gmail.com wrote:

Seems to be this as my library info shows CGAL 4.10: https://github.com/openscad/openscad/issues/2139

That sounds likely; CGAL used to have a precondition check for planar faces, but they removed this in CGAL-4.10, causing the polyhedron builder to fail later in the process if given slightly non-planar polygons (note: I guess most polygons are technically non-planar when specified using floating point numbers).

Since we don’t know exactly what CGAL accepts, we catch the exceptions thrown from CGAL and rebuild the polyhedron using triangles if it fails.
..but we don’t yet know exactly what exceptions are caused by non-planar input..

-Marius

> On Dec 3, 2017, at 7:09 AM, nop head <nop.head@gmail.com> wrote: > > Seems to be this as my library info shows CGAL 4.10: https://github.com/openscad/openscad/issues/2139 > That sounds likely; CGAL used to have a precondition check for planar faces, but they removed this in CGAL-4.10, causing the polyhedron builder to fail later in the process if given slightly non-planar polygons (note: I guess most polygons are technically non-planar when specified using floating point numbers). Since we don’t know exactly what CGAL accepts, we catch the exceptions thrown from CGAL and rebuild the polyhedron using triangles if it fails. ..but we don’t yet know exactly what exceptions are caused by non-planar input.. -Marius

NH

nop head

Sun, Dec 3, 2017 6:31 PM

I converted my quads to triangles and forced my end caps to be exactly
planer polygons and it works. So the latest Windows snapshot was released
with this regression.The test suite appears to catch it so it appears not
to have been tested.

On 3 December 2017 at 17:50, Marius Kintel marius@kintel.net wrote:

On Dec 3, 2017, at 7:09 AM, nop head nop.head@gmail.com wrote:

Seems to be this as my library info shows CGAL 4.10:

https://github.com/openscad/openscad/issues/2139

That sounds likely; CGAL used to have a precondition check for planar
faces, but they removed this in CGAL-4.10, causing the polyhedron builder
to fail later in the process if given slightly non-planar polygons (note: I
guess most polygons are technically non-planar when specified using
floating point numbers).

Since we don’t know exactly what CGAL accepts, we catch the exceptions
thrown from CGAL and rebuild the polyhedron using triangles if it fails.
..but we don’t yet know exactly what exceptions are caused by non-planar
input..

-Marius

OpenSCAD mailing list
Discuss@lists.openscad.org
http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

I converted my quads to triangles and forced my end caps to be exactly planer polygons and it works. So the latest Windows snapshot was released with this regression.The test suite appears to catch it so it appears not to have been tested. On 3 December 2017 at 17:50, Marius Kintel <marius@kintel.net> wrote: > > On Dec 3, 2017, at 7:09 AM, nop head <nop.head@gmail.com> wrote: > > > > Seems to be this as my library info shows CGAL 4.10: > https://github.com/openscad/openscad/issues/2139 > > > That sounds likely; CGAL used to have a precondition check for planar > faces, but they removed this in CGAL-4.10, causing the polyhedron builder > to fail later in the process if given slightly non-planar polygons (note: I > guess most polygons are technically non-planar when specified using > floating point numbers). > > Since we don’t know exactly what CGAL accepts, we catch the exceptions > thrown from CGAL and rebuild the polyhedron using triangles if it fails. > ..but we don’t yet know exactly what exceptions are caused by non-planar > input.. > > -Marius > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >

MK

Marius Kintel

Sun, Dec 3, 2017 6:35 PM

On Dec 3, 2017, at 1:31 PM, nop head nop.head@gmail.com wrote:

I converted my quads to triangles and forced my end caps to be exactly planer polygons and it works. So the latest Windows snapshot was released with this regression.The test suite appears to catch it so it appears not to have been tested.

We don’t test productions builds, and we don’t currently run tests on Windows, so anything triggered by 3rd party libraries rely on Travis CI being run against those libraries.

Can you reproduce this using the test suite locally?

-Marius

> On Dec 3, 2017, at 1:31 PM, nop head <nop.head@gmail.com> wrote: > > I converted my quads to triangles and forced my end caps to be exactly planer polygons and it works. So the latest Windows snapshot was released with this regression.The test suite appears to catch it so it appears not to have been tested. > We don’t test productions builds, and we don’t currently run tests on Windows, so anything triggered by 3rd party libraries rely on Travis CI being run against those libraries. Can you reproduce this using the test suite locally? -Marius

NH

nop head

Sun, Dec 3, 2017 7:09 PM

I can try when I get home. I am in Tenerife at the moment. I don't have the
dev setup on this laptop. It uses a hell of lot of disk space.

On 3 December 2017 at 18:35, Marius Kintel marius@kintel.net wrote:

On Dec 3, 2017, at 1:31 PM, nop head nop.head@gmail.com wrote:

I converted my quads to triangles and forced my end caps to be exactly

planer polygons and it works. So the latest Windows snapshot was released
with this regression.The test suite appears to catch it so it appears not
to have been tested.

We don’t test productions builds, and we don’t currently run tests on
Windows, so anything triggered by 3rd party libraries rely on Travis CI
being run against those libraries.

Can you reproduce this using the test suite locally?

-Marius

OpenSCAD mailing list
Discuss@lists.openscad.org
http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

I can try when I get home. I am in Tenerife at the moment. I don't have the dev setup on this laptop. It uses a hell of lot of disk space. On 3 December 2017 at 18:35, Marius Kintel <marius@kintel.net> wrote: > > On Dec 3, 2017, at 1:31 PM, nop head <nop.head@gmail.com> wrote: > > > > I converted my quads to triangles and forced my end caps to be exactly > planer polygons and it works. So the latest Windows snapshot was released > with this regression.The test suite appears to catch it so it appears not > to have been tested. > > > We don’t test productions builds, and we don’t currently run tests on > Windows, so anything triggered by 3rd party libraries rely on Travis CI > being run against those libraries. > > Can you reproduce this using the test suite locally? > > -Marius > > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >

discuss@lists.openscad.org

Anybody know what ss_circle.has_on(sp) means?