[postgis-users] regress_ogc. failed

[strk at refractions.net]

I just committed a patch for this (equals).

--strk;

On Wed, Sep 13, 2006 at 10:03:37AM -0600, Michael Fuhr wrote:
> On Wed, Sep 13, 2006 at 08:22:15AM +0200, Heinz-Josef Lcking wrote:
> > When does postgis / the user need regress_ogc ?
> 
> regress_ogc is a regression test; it submits a query and compares
> the result to the expected result to check for bugs.  In this case
> the query is
> 
>   SELECT 'unite_garray',
> 	 astext(unite_garray(geom_accum('{POINT(0 0)}', 'POINT(2 3)')));
> 
> The regression test expects
> 
>   unite_garray|MULTIPOINT(2 3,0 0)
> 
> With GEOS 2.2.3 I get the expected result; with GEOS 3.0.0rc1 I get
> 
>   unite_garray|MULTIPOINT(0 0,2 3)
> 
> These geometries are spatially equal; only their point order differs.
> According to the OpenGIS Simple Features Specification, Section
> 2.1.4, the elements of a MultiPoint "are not connected or ordered,"
> which is probably why strk wondered if the regression test should
> be equals('MULTIPOINT(2 3,0 0)', ...) instead of checking for a
> specific point order.
> 
> -- 
> Michael Fuhr