Hi Cory, I'm about to dive in with a review of your changes, but one question does come to mind about the general sequence of events. When the WM_DESTROY message is received has the window and graphics context been destroyed by the OS yet, or is it still just about to happen? Also what thread generates this event?
Robert. On Wed, Jul 1, 2009 at 6:40 PM, Cory Riddell<[email protected]> wrote: > Added setGraphicsThread(0) to GraphicsWindowWin32::destroyWindow(). This > is needed for cases where OSG is not rendering in a frame window. > > When it is in a frame window, OSG will get a WM_CLOSE message and the > event handler will handle this is a CLOSE_WINDOW and shutdown proceeds > correctly. > > If OSG is in a non-frame child window, it never gets the WM_CLOSE > message. It only receives a WM_DESTROY message during the actual window > destruction phase. In this situation, the graphics thread is never > terminated and an access violation results. > > Cory > > _______________________________________________ > osg-submissions mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org > > _______________________________________________ osg-submissions mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
