[postgis-devel] st_isvalid linestring and multilinestring

[Martin Davis]

Thanks for catching this, strk.  I'll add the unit test you provided, 
and have a look at cleaning up this variable.

strk wrote:
> I've found the bug in GEOS, but also unused code in JTS
> which probably confused the porting process.
> In IsValidOp.java you're using an 'isChecked' private
> boolean which you then never use.
> I guess it used to be used before like it is in GEOS,
> only it prevents (in GEOS) checking validity of any 
> component past the first of any MULTI* !!!
>
> Haven't checked in previous versions but the fact
> neither JTS nor GEOS contained *any* testcase for
> invalid geoms being after first one is pretty scary.
>
> --strk;
>
>   Share. It's fair.
>   http://iwouldntsteal.net
>
>   

-- 
Martin Davis
Senior Technical Architect
Refractions Research, Inc.
(250) 383-3022