[postgis-tickets] [PostGIS] #2307: ST_MakeValid outputs invalid geometries

[PostGIS]

#2307: ST_MakeValid outputs invalid geometries
--------------------------+-------------------------------------------------
 Reporter:  dmiranda      |       Owner:  pramsey      
     Type:  defect        |      Status:  new          
 Priority:  high          |   Milestone:  PostGIS 2.0.4
Component:  postgis       |     Version:  2.0.x        
 Keywords:  st_makevalid  |  
--------------------------+-------------------------------------------------
Changes (by pramsey):

 * cc: strk (added)


Comment:

 OK, so some interesting results... the result of ST_MakeValid on the
 geometry cited in the comment is this:
 {{{
 select st_astext(st_makevalid(geom)) from test;

  MULTIPOLYGON(((1.38241720848787e+306
 1.38241720848787e+306,1.38241720848787e+306
 1.38241720848787e+306,1.38241720848787e+306
 1.38241720848787e+306,1.38241720848787e+306
 1.38241720848787e+306,1.38241720848787e+306
 1.38241720848787e+306,1.38241720848787e+306 1.38241720848787e+306),(0
 1.38241720848787e+306,1.38241720848787e+306
 1.38241720848787e+306,1.38241720848787e+306
 2.13091231225152e-314,6.95322297557074e-310
 2.12970193392818e-314,6.95322297557153e-310 1.38241720848787e+306)))
 }}}
 Fun, hey? But it gets better. This is a single-part multi-polygon... what
 happens if we strip out the multi and clean the polygon as a singleton.
 {{{
 select st_astext(st_makevalid(st_geometryn(geom,1))) from test;
 POLYGON((-41.1931560828767 -7.32572934869855,-41.161610776897
 -7.32573074083104,-41.1569353108121 -7.32573081181504,-41.1569356001464
 -7.34833623254846,-41.193158194274 -7.34833476367411,-41.1931560828767
 -7.32572934869855),(-41.161610776897 -7.32573074083104,-41.1879093408465
 -7.32573084378415,-41.1879095682369 -7.34254610131274,-41.1616110087225
 -7.34254610942963,-41.161610776897 -7.32573074083104))
 }}}
 Hey presto, it works, and this output is actually valid (though looking at
 it in the JTS testbuilder, it still looks invalid until you zoom up quite
 close to see that it doesn't have a hole/hull edge coincidence, just a
 Very Very close hole and hull.

 So, first thing is to examine what the cleaning functions are doing on
 multi-geometries and see if we can just plug that easy problem to start
 with.

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/2307#comment:1>
PostGIS <http://trac.osgeo.org/postgis/>
The PostGIS Trac is used for bug, enhancement & task tracking, a user and developer wiki, and a view into the subversion code repository of PostGIS project.