Michael Bedward wrote: > Hi folks, > > Sorry to still be banging on about BBox filters but I've spent a while > following the creation of the filter in my debugger (using the code > posted previously) and it just looks broken to me. > > When you pass a ReferencedEnvelope to the bbox factory method the > first thing that happens is that it's converted to a JTS Envelope and > subsequently to a Polygon. The BBoxExpressionImpl constructor checks > to see if it has been passed a ReferencedEnvelope in order to get the > SRS but, of course, it hasn't been. The SRS is sort of kludged in at > the end of the bbox(Expression, double, double, double, double, String > srs) method but then not used at all during evaluation of the filter. > > Is this not broken ? :-)
Sorry I did not have time to look into it and I might not have any until next week. However I know that all datastores and filters are projection unaware, they expect everything to be in the same SRS as the data. What we do in GeoServer is to apply filter visitors to reproject the filters to the store native SRS before they reach the data at all Trying for a library wide fix will be a very big effort. Fixing just the filter itself to work when it's actually evaluated in memory will be easier provided we know about the SRS of the geometries we're receiving (which I don't think we can know from a general point of view) Cheers Andrea ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Geotools-devel mailing list Geotools-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geotools-devel
