This release is also available from our maven repository. This release is made in conjunction with GeoServer 2.5.3.
A few highlights from the GeoTools 11.3-Release Notes:
- Rendering fixes related to cut geometries/labels at map tile borders
- Several improvements/fixes to the NetCDF readers
- Table hints for SQL Server can be specified at the store level, and it's now possible to force SQL Server to use spatial indexes
- A good set of JDBC related fixes, for joins, multi-geometry tables, spurious error reports against invalid sql views
- Make sure SortedSimpleFeatureCollection makes full use of the merge-sort sorter and respects the system wide in memory limits (was going straight and fully to disk before)
Thanks to Andrea for this release (GeoSolutions).
About GeoTools 11
Summary of the new features for the GeoTools 11 series:
- The DataStore API has a new removeSchema method to drop feature types. This new optional feature is currently implemented by the JDBCDataStore family (all spatial database backed stores), other stores will likely throw an UnsupportedOperationException
- JDBCDataStore now exposes facilities to list, create and destroy indexes on database columns.
- Ability to create and drop databases from the PostgisNGFactory
- PostGis data store will now call ST_Simplify when the GEOMETRY_SIMPLIFICATION hint is provided, significantly speeding up loading of complex geometries (the renderer can perform scale based simplification already, but doing it before sending the data speeds up data retrieval significantly)
- ImageMosaic can now manage vector footprints for its granules, allowing to filter out no-data or corrupted sections of the imagery
- All properties in a SLD style can now have a local unit of measure, as opposed to specifying the unit of measure per symbolizer. For example, if you need to have a line width to be 10 meters, its value can now be "10m"
- Improved handling of data with 3D coordinates in JDBC data stores
- A number of small improvements to the rendering engine, such as improved raster icon placement resulting in cleaner, less blurry output, improved label grouping, better handling of icons at the border of the map and in general much improved estimation of the buffer area needed to include all symbols in a map (for features that sit outside the map, but whose symbols are big enough to enter it).
