[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: what does "Error while clipping PBO_SUB: 3" mean?



On Wed, 2010-02-10 at 02:11 -0600, Mark Rages wrote:
> I filed a bug and uploaded a file which exhibits this:
> 
> https://sourceforge.net/tracker/?func=detail&aid=2948916&group_id=73743&atid=538811
> 
> On Wed, Feb 10, 2010 at 1:29 AM, Mark Rages <markrages@xxxxxxxxx> wrote:
> > I'm trying to work with a rather compex polygon ground plane.  I get
> > the following spew on the terminal, then crash:

I see a ground plane and some tracks - on different layers. Is that
expected. I don't get any polygon clipping errors on the console with my
experimental branches - but I _do_ see a rendering error, which looks
like the sort of error we encounter if PCB produces a self-intersecting
polygon.

Did you computer generate the polygon? (Is it possible there is an
erroneous coordinate in it somewhere?)

The fact I don't see complaining - either means that I'm not taking the
required steps to reproduce (ie.. what is subtracting from the polygon
if your lines are on a different layer?)

Adding vias, or things cutting the polygons - I can't seem to make it
complain on the console. This suggests that the bug isn't manifesting in
the same way on my branch - which gives a (small) clue.

It could be (I've seen it before), that the no-holes polygon dicer gets
stuck into a recursive loop due to a polygon which keeps dicing in two
to give something with holes in it. That can run the program out of
stack space and crash it, although your console warnings don't "quite"
look like that. Running under GDB might tell more - but I've not got the
time to do that just now.

Unfortunately, if it is the ever lurking "self-intersection" problem - I
can't fix it easily. We need to introduce a topology preserving,  snap-
rounding based polygon intersection routine, and that is _hard_. I can
only suggest attempting to work around the issue by tweaking your
polygon geometry.

If you generated it progmatically, be sure to check your coordinates
don't cause a self-intersection.

> 




_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user