Well you can compute the gate size and use less than 0 to max size in each direction. and feed the numbers, assuming the gate fits the view that is (and it dont get too small)
what if I just want to list the objects that the camera sees?
Select save selection reset back to what you had before. The api call is just doing the same thing as drawing a marque box in the screen. If you dont want to rasterize yourself then that's what you get the same raw function.
But obviously if time is no issue you can use a software render to sort this out (i used to do this back in the day when maya couldn't cull visibility properly*, the cool thing about thsi is that it can be affected by transparency factors in way that opengl can not such as only select objects that are behind a second object. Plus its easy to say its the actual resolution of what the render is, a but slow tough.)
*and still for some ambient occlusion testing based selection, joint tweaking etc