I've just pushed new code to the "rotation" branch in my fork [1] that makes panning and zooming supported at the MapToPixel level.
The MapToPixel class is used by the renderer to transform inputs to output coordinate space, right after performing geographical projection transformations. Rendering wise, the rotation doesn't seem to give nice results. Arrow decorations and labels are "flattened", and "parallel" labels are parallel to the original (non-rotated) lines. [1] https://github.com/strk/qgis/tree/rotation I suspect there's some redundancy I'm missing out, whereas local classes have local view on what the canvas extent or scale is. For example I saw QgsMapRenderer has mExtent and mMapUnitsPerPixel members but also a mRendererContext which can provide a MapToPixel object having another copy of that information (extent and scale). What are these duplications for ? Could we get rid of them to have a unified configuration for coordinate transforms ? --strk; () ASCII ribbon campaign -- Keep it simple ! /\ http://strk.keybit.net/rants/ascii_mails.txt _______________________________________________ Qgis-developer mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/qgis-developer
