On Sat, Mar 23, 2024 at 10:07:31AM -0400, Marius Kintel wrote:

Correct! I thought nothing significant would've changed in the past
month or so as I didn't see any big announcements.

Verfied: Version mid-february still shows the problem. Version march23
does not.

Roger. 

--
** R.E.Wolff@BitWizard.nl ** https://www.BitWizard.nl/ ** +31-15-2049110 **
**    Delftechpark 11 2628 XJ  Delft, The Netherlands.  KVK: 27239233    **
f equals m times a. When your f is steady, and your m is going down
your a is going up.  -- Chris Hadfield about flying up the space shuttle.

The big announcement will come once Manifold moves out of experimental :)
We doing a lot of small fixes these days, some of which will enable us to process previously tricky corner cases.

-Marius

Even the latest version 28-03-2024 has the same issue with manifold.

Luckily I have a 14th Jan 2024 version which works perfectly without any
errors.

On Mon, 25 Mar 2024 at 19:56, Marius Kintel via Discuss <
discuss@lists.openscad.org> wrote:

Hi Marius,
This same example of Hook has this issue:
you can download it from here:
https://github.com/sprabhakar2006/openSCAD/blob/main/hook.scad

In case any issue in downloading, will send the file through mail
Regards
Sanjeev

On Sat, 30 Mar 2024 at 01:18, Marius Kintel marius@kintel.net wrote:

Sanjeev,

Oh, you’re referring to that it stops on a warning if you have “stop on first warning” enabled?

We’re considering whether or not we should issue warnings on user-supplied non-manifold topologies, but the easy workaround is to disable that setting in preferences.

-Marius

Hi Marius,
Ok, Thanks for the clarification.
Yes with the "stop on first warning" disabled, it is OK.
Regards
Sanjeev

On Sat, 30 Mar 2024 at 09:17, Marius Kintel marius@kintel.net wrote:

On Fri, Mar 29, 2024 at 11:47:39PM -0400, Marius Kintel via Discuss wrote:

That's a "works as designed". I remember posting something to that
effect before.

OK! I vote: YES, complain on creation of the object that it is not
manifold.

I remember in the old days, say:

difference () {
        cube ([40,40,20]);
    translate ([2,2,2]) cube ([36,36,18]);
    }
    translate ([20,20,2]) cylinder (d=10,h=5);

would throw errors when rendering or result in unusable STLs. That
would be annoying because you'd have no clue where in your design you
did something wrong and it would be hard-to-find when the design is
real-world and not this simple example.

Issue a warning when importing or creating an object that is not
manifold.

Maybe it shouldn't be an error: If you cut off the part that has the
anomaly, then it might still be OK. That might be exactly the reason
the object is being manipulated in openscad.

Roger. 

--
** R.E.Wolff@BitWizard.nl ** https://www.BitWizard.nl/ ** +31-15-2049110 **
**    Delftechpark 11 2628 XJ  Delft, The Netherlands.  KVK: 27239233    **
f equals m times a. When your f is steady, and your m is going down
your a is going up.  -- Chris Hadfield about flying up the space shuttle.

In this case we issued a warning on a nonmanifold topology in a polyhedron(),  which we managed to repair and import correctly. We warned before attempting the repair.

On Sat, Mar 30, 2024 at 08:48:43AM -0400, Marius Kintel wrote:

IMHO, "repairing" is always a bit fishy. The obvious repair may seem
obvious, but it is not guaranteed to have been the intent. In theory
it can be "wrong".

Suppose we have a cube with the top face missing. The obviously right
thing to do is to add two triangles closing the cube.

Now two of the vertices are raised a bit above the other two. Now you
can choose which diagonal to use and get different results is the two
cases.

Roger. 

s=20;

difference () {
cube ([s,s,s*(1+sqrt(.5))]);
translate ([0,0,s*(1+sqrt(.0))]) rotate ([0,-45,-45])cube ([s,2*s,s]);
}

translate ([1.5*s,0,0])

difference () {
cube ([s,s,s*(1+sqrt(.5))]);
translate ([0,0,s]) rotate ([0,-45,45])translate ([0,-s,0]) cube ([s,2s,s]);
translate ([s,s,s]) rotate ([0,-45,45])translate ([0,-s,0]) cube ([s,2s,s]);
}

--
** R.E.Wolff@BitWizard.nl ** https://www.BitWizard.nl/ ** +31-15-2049110 **
**    Delftechpark 11 2628 XJ  Delft, The Netherlands.  KVK: 27239233    **
f equals m times a. When your f is steady, and your m is going down
your a is going up.  -- Chris Hadfield about flying up the space shuttle.