Boost logo

Geometry :

Subject: [ggl] incorrect result for an intersection
From: Barend Gehrels (barend)
Date: 2011-04-17 05:29:43


Hi Arash,

On 17-4-2011 7:00, Arash Partow wrote:
> Barend Gehrels wrote:
>>
>> Arash, your input is welcome, especially if it is a once-off effort.
>>
>
> Fair enough! :D

OK, seeing this combination this is not what I meant, so let me
rephrase: Your input is always welcome! And a once-off effort for
verifying polygon validity or intersections is also welcome.

>
>
>> A GPC-backend: might be possible but not supported now.
>
> That would be really nice. because it could mean, that one could
> plug-in ggl bool-ops or gpc bool-ops or cgal bool-ops etc.
>
>
>> Besides that: it might be more convenient the other way round: GGL
>> supporting GPC polygons. That might be interesting e.g. to read WKT
>> into GPC polygons.
>
> Barend, there is no gpc polygon format. It is up to the end-user to
> figure out how they persist the in-memory contour ds. That said wkt
> doesn't look too hard, just one question are islands described in
> reverse order of their containing hole in wkt?

No, a polygon with a hole and inside that hole an island must be
presented as a multi-polygon (containing two outer rings).

OK, I was using gpc_vertex and gpc_polygon for my test, 2 years ago, but
apparently I've to look again into the implementation then.

Regards, Barend


Geometry list run by mateusz at loskot.net