From: Jody G. <jod...@gm...> - 2009-06-16 12:54:21
|
Hi Andrea: On Tue, Jun 16, 2009 at 4:32 PM, Andrea Aime<aa...@op...> wrote: > I'm missing how this can be any manageable in the long term. Indeed; can we have a set of useful user map keys stored anywhere? Perhaps in the FeatureTypes utility class? > So next time we add a new interesting key we'll have to modify all > the points in which we need a strict equality? So how about ... if( target.equalsExact( origional, FeatureType.LIBRARY_KEYS) ){ >> ... >> } > What if a datastore adds a new one but its not being developed by > a core developer? How is that going to be handled? Yeah perhaps this idea of using the user map is not a good one; if we end up using it for the library in a way that prevents users from using it we kind of lose the point of having a map for user data. > I've asked this question a couple of times already but you seem > to dodge it every time :-) Not intensionally; I am still trying to sort out if we need an API change. However that can honestly come later - we are trying to use the user map to sort out of the approach is a good one. At worst this equals check will cause some extra work (if the user has placed some more stuff in the user map). >> Unless you would like to confirm that our retyping code is going to >> copy the entire user map along as it goes? > > Humm.... copy? The retyper creates new features and attaches them > to the target feature type, which has the desired user map. > There is no user map copying. When that new feature type was created; it is often created by the subtype method; we should consider copying the user map during the sub type method. > Cheers > Andrea > > -- > Andrea Aime > OpenGeo - http://opengeo.org > Expert service straight from the developers. > |