[postgis-devel] Possible server crash detected

[strk at refractions.net]

Another note: MULTIPOINT(EMPTY) is interpreted as a 
MULTIPOINT with a single, EMPTY element (which is converted
internally to a COLLECTION, thus making it an invalid 
representation).

I think you meant to really provide an EMPTY geometry, which
is obtained by 'MULTIPOINT EMPTY'.

I think we should first fix the interpretation.

MULTIPOINT(EMPTY) should probably be the EMPTY geometry,
whereas MULTIPOINT( ( EMPTY ) ) would be a multipoint
with an EMPTY element [ possibly, *hopefully* :) unsupported ].

For a recent discussion about MULTIPOINT WKT format see:
See: http://lists.jump-project.org/pipermail/jts-devel/2006-April/001572.html

--strk;

On Tue, Apr 18, 2006 at 12:15:25PM +0200, strk at refractions.net wrote:
> I'm taking a look at this bug. Question really is: should we
> allow EMPTY components in non-collection geometries ?
> 
> A simpler exploit for the bug is:
> 
> select summary('MULTIPOINT(EMPTY, 0 0)');
> 
> --strk;
> 
> On Mon, Apr 03, 2006 at 04:38:38PM +0200, Markus Schaber wrote:
> > Hi, Strk,
> > 
> > strk at refractions.net wrote:
> > > Confirmed on 1.1.2 as well.
> > > Could you file a bug report please ?
> > 
> > http://postgis.refractions.net/bugs/bug.php?op=show&bugid=82&pos=20
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel

-- 

 /"\    ASCII Ribbon Campaign
 \ /    Respect for low technology.
  X     Keep e-mail messages readable by any computer system.
 / \    Keep it ASCII.