Subject: Re: [geometry] closed wrt closedness?
From: Barend Gehrels (barend_at_[hidden])
Date: 2012-02-16 13:00:21
On 16-2-2012 16:45, Volker Schöch wrote:
> Pardon the pun... Looking at the example from my previous post, I noticed that the result of the asymmetric difference algorithm is a closed polygon although it should be the same type as the input (not closed).
> A = MULTIPOLYGON(((1740 1056,2640 1326,3540 1776,3540 32767,1740 32767)))
> B = MULTIPOLYGON(((1740 1794,2640 2280,3540 2352,3540 32767,1740 32767)))
> A^B = MULTIPOLYGON(((1740 1794,1740 1056,2640 1326,3540 1776,3540 2352,2640 2280,1740 1794)))
> I am using boost 1.48.0. My polygon type is oriented counter-clockwise and not closed.
> Is this intentional/tolerable, or do you consider this a bug?
Yes, this is indeed a shortcoming. So you can call it a bug. Though it
is normally not really harmful (at least, if you can handle degenerecies).
Will be fixed. open/counterclockwise is well-tested as input for
intersections and unions but (as you will have noticed) most focus and
care has been for closed/clockwise... Sorry for that. There is still
work to do there.
Geometry list run by mateusz at loskot.net