[postgis-devel] [PostGIS] #1345: ST_StartPoint, ST_EndPoint regress failure

[PostGIS]

#1345: ST_StartPoint, ST_EndPoint regress failure
-----------------------+----------------------------------------------------
  Reporter:  robe      |       Owner:  pramsey      
      Type:  defect    |      Status:  closed       
  Priority:  critical  |   Milestone:  PostGIS 2.0.0
 Component:  postgis   |     Version:  trunk        
Resolution:  fixed     |    Keywords:               
-----------------------+----------------------------------------------------

Comment(by pramsey):

 Judging by the differences in code from 1.5 to 2.0 the change was a side
 effect of the serialization change, since the original code worked against
 "inspected geometries", that is, direct access to the serialization. The
 original code was also pretty clearly deliberately unwrapping the first
 instance of the targeted type within any multi* geometries it was handed.
 So if it was handed a geometry collection or multilinestring, it would
 unwrap and then use the *first* instance it found to answer the call.
 Whether the benefit of auto-unwrapping of things like single-entry
 linestrings outweighs the surprise of the end point of a double entry
 multlinestring being the end of the first entry, I leave as an exercise to
 the user.

 I noticed the problem working on my tutorial, which includes some calls to
 NumInteriorRings that previously transparently unwrapped a multipolygon to
 the first entry.

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/1345#comment:9>
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.