Geometry Boost List: Re: [geometry] Possible bug in boost::geometry::equals ?

Subject: Re: [geometry] Possible bug in boost::geometry::equals ?
From: Adam Wulkiewicz (adam.wulkiewicz_at_[hidden])
Date: 2014-01-25 19:40:15

Hi,

2014-01-21 <yohann.benedic_at_[hidden]>

> Hi there,
>
> To summarize, if two polygons are built spatially equal, then
> boost::geometry::equals sometimes fails to recognize it.
> I hope this post is not a duplicate. I googled around and did not find
> anything.
>
> <snip>

>
> From my investigations, the problem comes from the building of the
> "point-direction" vectors, which end up having different sizes. At some
> point in the algorithm, the direction of the last point is compared to the
> one of the first point and the latter is removed if equal. Here it happens
> between the last point of the interior ring and the first of the exterior
> ring : on p1 they are different but on p2, hence the difference of size
> that causes equals to fail.
>
>
Thanks for the info. It was indeed caused by an error in collect_vectors
internal utility used by equals(). It should be fixed in the develop branch.

Regards,
Adam