[postgis-tickets] [PostGIS] #3899: BTree opclass does not take object length into account

[PostGIS]

#3899: BTree opclass does not take object length into account
----------------------+---------------------------
  Reporter:  komzpa   |      Owner:  pramsey
      Type:  patch    |     Status:  reopened
  Priority:  medium   |  Milestone:  PostGIS 2.4.1
 Component:  postgis  |    Version:  2.4.x
Resolution:           |   Keywords:
----------------------+---------------------------

Comment (by komzpa):

 With this change, some people may need to rebuild their btrees (say on
 UNIQUE columns) once again.
 Without it, they will have corrupted uniqueness for some edge cases.
 Cases are quite obscure (a polygon that is ST_Reverse of itself can be
 sorted in wrong order if one copy of it has a box and another does not).

 It just occured to me that we're still not looking enough into header -
 for two geometries with the geometry and different SRID we don't define a
 stable order. Looks like this thing is a candidate for oss-fuzz?

--
Ticket URL: <https://trac.osgeo.org/postgis/ticket/3899#comment:3>
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.