Semicircular Mesh

[Anonymous]

Hello,
I have run into something strange with the mesh tool.
The attached image shows a semicircle drawn with arcs and lines.
Above that is a mesh made by tracing the figure made with arcs and lines.
When I go to the 3D window to view the mesh I get a warning
saying "Invalid polygon, self-intersection or hole intersects boundary".
Nothing appears in the 3D window.
If I trace the same figure using the slab tool or the roof tool and
go to the 3D window I get no warning and the slab or roof
show in 3D.
If I make an object out of the mesh using save selection as GDL Object
and open the objects 3D script and check script I get an OK but
if I go to the 3D view I get an error message saying
"Invalid curve definition at line......" and the 3D image fails to process.
I have tried this in both AC 8.1 and AC 9.
Would someone verify this behavior and perhaps
suggest a way to make such a mesh shape work.
I know that if I make the included angle one degree less than 180°
or make the inner radius 1/64" less than 10'-0" the error does not occur.
This work around is not satisfactory.
Thank you,
Peter Devlin

[Anonymous]

Hello Z.Bauer,
I am gratified to hear that you have verified the problem.
I have drawn two meshes that are identical to the one
that is flawed except that the inner arc is drawn with the
three point arc method in such a way that at the top midpoint
of the arc is 1/64" less and 1/64" more than at the
start and end points of the inner arc which are exactly 10'-0".
I believe this makes arcs that are not exact semicircles.
These two meshes do not generate an error.
Thanks,
Peter Devlin

[Anonymous]

TEST RESULTS:

I have created 4 point mesh with dims there are represented on the image. As already have sad error runs in 3d. I have saved mesh as Mass.gsm GDL Object to check the coords.
Have found nothing that can bring any conflict.
I think this bug in MASS function for specific dims, because than I have added Mass coords in to POLY_ , EXTRUDE , SLAB function to check and generate script in 3d. And they work fine!

Tested objects are attached!

Anyone can prove that that's not a "bug again from GS"

[Anonymous]

Yes Peter,
I have added 3rd point in the middle of inner arc too and it's work's fine without any problem. And actually the problem must not be. I remember that something else was when I worked on AC 6.5(7). Most of curved meshes and slabs in my projects I had done with min 3 points on a curve due problem like this one.

[Anonymous]

Hello Z.Bauer,
I am a bit confused by your last post.
You say:
"And actually the problem must not be"

Are you saying you have concluded
there is no problem with the mesh tool in this context ?
Thank you,
Peter Devlin

[Anonymous]

Hello Z.Bauer,
I opened your library part with the MASS command and viewed in 3D
and get the same error as I got when I did the same thing.
I also could not see anything in the script that would cause the problem.
Thank you,
Peter Devlin

[Anonymous]

Doing a little experimenting, it seems that if an exact semicircle
is drawn on an edge of a mesh, either concave or convex,
the mesh fails to generate in 3D and an invalid polygon error is reported.
It seems clear to me that the unique properties of an exact semicircle
is causing the problem, but only for the mesh tool and only if the semicircle
is drawn on the edge, semicircular holes don't cause a problem.
Other tools are not affected.
So, hopefully, GS will have something to say about this issue.
Thank you,
Peter Devlin

[Anonymous]

This has truly been one of the most interesting issues / bugs I've ever seen in ArchiCAD.. congrats Peter

cheers,
dan

[Anonymous]

Peter wrote:
Hello Z.Bauer,
I am a bit confused by your last post.
You say:
"And actually the problem must not be"

Are you saying you have concluded
there is no problem with the mesh tool in this context ?

Hello Peter again,

no, as I have sad and shown in attached files the problem exist, but just looking generated coords I sad "actually I see no problem, this is "crafty bug""!

But I wasn't attentive:
AC has generated not correct coords for Mass. I have done some additional tests to find the problem. And here is that:

!METHOD TO FIND THE PROBLEM:
!1. The autogenerated for MASS object coords was added in to SLAB function.
!2. 2D and 3D HOTSPOTS with the same coords was added.
!3. After generating 3D for SLAB and zooming left corner/point on the inner curve,
! One additional problematic point was found (see on attached image & in file)

I would say is bug and GS must solve problem.

[Anonymous]

Problematic Point:

[Anonymous]

Hello Z.Bauer,
"crafty bug"" indeed !
I found the same thing with a slab with a concave quarter arc cut into it.
In my case the extra point formed a protrusion at one corner.
Peter Devlin