[geos-devel] [GEOS] #959: Reversed result in GEOSClipByRect

[GEOS]

#959: Reversed result in GEOSClipByRect
-------------------------+---------------------------
 Reporter:  Algunenano   |       Owner:  geos-devel@…
     Type:  defect       |      Status:  new
 Priority:  major        |   Milestone:  3.5.2
Component:  Default      |     Version:  3.5.0
 Severity:  Significant  |  Resolution:
 Keywords:               |
-------------------------+---------------------------

Comment (by Algunenano):

 > I think by definition invalid inputs result in invalid outputs

 Fair enough, but then ST_AsMvtGeom shouldn't be using it. To be honest,
 I'm already on the verge of removing the geos path and leave wagyu as the
 only option for clipping and validation for MVT polygons.

 > Given that this geometry is only slightly invalid (a single collapsed
 line segment), and given the relatively simplicity of clipping to a
 rectangle, it seems like it might be technically feasible to make clipping
 more tolerant of some invalid inputs. This might require an entirely new
 algorithm, though.

 What surprised me the most is that it gives this invalid input although
 the invalid spike is outside the clipping rectangle (which is the only
 ring). I haven't checked the algorithm that Geos uses yet.

 As a reference, here is the current quick clip mechanism in wagyu:
 https://github.com/postgis/postgis/blob/svn-
 trunk/deps/wagyu/include/mapbox/geometry/wagyu/quick_clip.hpp#L67

-- 
Ticket URL: <https://trac.osgeo.org/geos/ticket/959#comment:4>
GEOS <http://trac.osgeo.org/geos>
GEOS (Geometry Engine - Open Source) is a C++ port of the Java Topology Suite (JTS).