[postgis-devel] MultiPoint weirdness

[strk at refractions.net]

On Tue, Jan 18, 2005 at 02:07:18PM +0100, Markus Schaber wrote:
> Hi, strk,
> 
> strk at refractions.net schrieb:
> 
> > mm... 'MULTIPOINT(0 0)'::geometry also doens't work.
> > I've found out this is due to the GEOMETRYCOLLECTION(EMPTY)
> > support I've added recently (so should not be in RC1).
> 
> So this answers the question how this bug could have been undiscovered -
> it existed only a small time.
> 
> BTW, some other geometries seem to suffer from the same problem.
> 
> > What do you think about dropping all EMPTY geoms except
> > GEOMETRYCOLLECTION ? It would simplify the code and the usage
> > of it. Canonical empty output is:
> >
> > 	GEOMETRYCOLLECTION EMPTY
> >
> > Accepted inputs would be:
> >
> > 	GEOMETRYCOLLECTION EMPTY
> > 	GEOMETRYCOLLECTION(EMPTY)
> >
> 
> This is okay, I think. Simplifying the code is always a good idea.
> 
> So I'll drop all other EMPTY tests from my new jdbc test suite.
> 
> I'm not really shure what the WKT spec says about empty geometries, it
> seems to be like swiss cheese (lots of holes).

Mmm.. I've checked it. Actually no holes there, previous behaviour
was strict-standard.

Valid constructs:

	POINT EMPTY
	LINESTRING EMPTY
	POLYGON EMPTY
	MULTIPOINT EMPTY
	MULTILINESTRING EMPTY
	MULTIPOLYGON EMPTY
	GEOMETRYCOLLECTION EMPTY

Our old GEOMETRYCOLLECTION(EMPTY) was NON-standard... maybe we should
*only* allow GEOMETRYCOLLECITON(EMPTY) and all the other standard ones ?

REF: 99-049 -- 3.2.5.2 Language Constructs

--strk;


> 
> Thanks,
> Markus



-- 

For standing up against patentability of software,

  Thank You, Poland!

Read the intervention:    http://kwiki.ffii.org/ConsPolon041221En
Send your thanks:         thankyoupoland.info
Read/do more:		  http://www.noepatents.org/