[geos-devel] Exposing Precision model through C-API

Sandro Santilli strk at keybit.net
Fri Mar 7 00:16:32 PST 2014


On Fri, Mar 07, 2014 at 12:54:19AM +0530, Varun Saraf wrote:

> I have made a temporary proposal. Below is the link to the wiki page.
> 
> http://trac.osgeo.org/geos/wiki/GSoC/CAPI_PrecisionModel

You propose to expose a new "PrecisionModel" object, but not
the exposure of a "GeometryFactory" ? Is that correct ?

Did you understand the proposal of Sean Gillies ?
He proposed to expose no new objects, but simply pretend the
"PrecisionModel" parameters are really parameters of a "context".

What problems do you see with that proposal ?
Have you read the problems I individuated about that ?
Can you think about solutions to them ?

--strk;

> On Thu, Mar 6, 2014 at 5:44 PM, Sandro Santilli <strk at keybit.net> wrote:
> 
> > On Thu, Mar 06, 2014 at 05:23:43PM +0530, Varun Saraf wrote:
> > > Hi,
> > >
> > > We either create the PrecisionModel and the GeometryFactory in the
> > > GEOSContext_setPrecisionModel or pass a pointer to the
> > > new PrecisionModel to this method. But the signature of
> > initGEOS/initGEOS_r
> > > cannot be changed so how to pass it to the initGEOS. Please clarify.
> >
> > Please make proposals.
> > My proposal was in the mail you quoted, did you understand it ?
> >
> > --strk;
> >
> > > On Thu, Feb 20, 2014 at 12:00 PM, Sandro Santilli <strk at keybit.net>
> > wrote:
> > > >
> > > > I agree about making the precision model part of the context.
> > > > But we can't change the initGEOS signature, so we'll need
> > > > some GEOSContext_setPrecisionModel kind of method.
> > > > At that point, we have to decide what happens with a previously
> > > > set PrecisionModel, if it has to be allowed at all.
> > > >
> > > > Deleting the PrecisionModel and the GeometryFactory on finishGEOS
> > > > is fine, but note that this would make any created geometry also
> > > > broken (they hold a reference to the GeometryFactory that created
> > > > them). Not a big deal, as long as it's clearely documented. Dunno
> > > > if it would make sense for the opaque GEOSGeometry type to also
> > > > hold a pointer to the context, just in case...
> > _______________________________________________
> > geos-devel mailing list
> > geos-devel at lists.osgeo.org
> > http://lists.osgeo.org/mailman/listinfo/geos-devel
> >

> _______________________________________________
> geos-devel mailing list
> geos-devel at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/geos-devel


-- 

 ()  ASCII ribbon campaign  --  Keep it simple !
 /\  http://strk.keybit.net/rants/ascii_mails.txt  


More information about the geos-devel mailing list