[OpenSCAD] example that works in 2009.11, but not it 2010.02

Britton Kerin britton.kerin at gmail.com
Mon Apr 5 17:38:23 CEST 2010


On Mon, Apr 5, 2010 at 3:26 AM, Marius Kintel <marius at kintel.net> wrote:

> On Apr 5, 2010, at 03:36 AM, Britton Kerin wrote:
>
> > The attached set of files render as desired in 2009.11, but not in
> 2010.02.
> > I haven't tried to distill the problem down to its essence, but it looks
> like
> > maybe something isn't working right with radial extrusion in 2010.02.
> >
> AFAIK, rotate_extrude() was introduced in 2010.01.
>

Looks like you're right, I have a build in that source dir but not 2009.  A
tiny
little useful feature to add is the GNU-style --version flag.  Though now I
think about it its right there in About menu item... but as you can see
OpenSCAD
is attractive to people who don't think in GUI too well :)


> Anyway: Your problem is because you specified your polygon in the wrong
> winding order.
> You can verify this by flipping it around (e.g. add a scale([-1,1]) in
> front of it)
>

This fixes it.  For whatever reason, it isn't necessary for rendering with
openscad 2010-01 though.  However,
the quickview never worked right even in 2010-01 with high convexity
arguments everywhere.  So its a lovely
relief to know of this issue :)

I added a small note about this to the bottom of
http://en.wikibooks.org/wiki/OpenSCAD_User_Manual/2D_Primitives#polygon

There are still a couple of things I don't understand though:

1.  Why doesn't reversing the path order (i.e. 20, 19, ..., 1, 0 in the
original example) fix the problem?

2.  The 2D subsystem draws in X and y as the above polygon page shows.  But
the rotate_extrude command appears
     rotate about the Z axis (by default at least).  Is there an implicit
step in which 2D figure are reoriented before extrusion?

Thanks again,
Britton

Polygon winding order is a bit confusing in OpenSCAD atm., especially for
> 2d->3d operations as you end up with an invisible object. I don't have a
> good idea of how to improve this without introducing incompatibe behavior
> though.
> It might be doable to do some analysis and print out some fat warnings if
> smth. looks fishy..
>
> Any takers? ;)
>
> ~/= Marius
>
> --
> We are Elektropeople for a better living.
>
>
>
>
> _______________________________________________
> OpenSCAD mailing list
> OpenSCAD at rocklinux.org
> http://rocklinux.net/mailman/listinfo/openscad
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rocklinux.net/pipermail/openscad/attachments/20100405/fc30188e/attachment.html


More information about the Openscad mailing list