[postgis-devel] KML from PostGIS patch

[Paul Ramsey]

Which points up something: it doesn't look like transform() has a short 
circuit right at the top to simply return the input geometry if the 
requested SRID is the same as the current geometry SRID.

Markus Schaber wrote:

> This way, those people who don't need to transform first (because they
> keep the data in a compatible coordinate system) can skip the
> transform() overhead.
> 
> HTH,
> Markus