I have finished a first implementation of multi screen capabilities for dtEntity. It is now possible to open additional windows from JavaScript or C++. The input handler handles events on a per-window basis, which makes it easy to write separate motion models for each screen. Windows are addressed by their context id. The main screen has context id 0, the openWindow function returns the context id of the newly created window. InputHandler.getKey(“k”, 0) would return true if the main window was focused and the k key pressed.
I also did some work on the integration of osgManipulator. There are some issues with the draggers not being picked when they are not the nearest intersection, also it is not yet possible to manipulate more than one object at once. But still this seems to have promise.
