smcameron

This commit appears to break lighting on the weapons screen (maybe elsewhere too, but most visibly on the weapons screen): 4bb5e1868a55684283d5ba681cc6b409f59302e5 Correct lighting:  Incorrect lighting: 

f298a80fa5de379b9bea1069edffcb4448405573 revert avoid nans in process vertex normals Reverted 4bb5e1868a55684283d5ba681cc6b409f59302e5

@jv4779 A long shot, but... any idea why is this multiplying a vector by an angle? (only thing I can think of is that for small angles, sin(angle) ~= angle,...

Ok, I think I got rid of the NaNs from process_vertex_normals() with this: 82e0c9bdf154d63b0d70793269e73474ef12be62

* 773d9c14574b0254fc1c12aca9f63dfe509e67b9 Fix divide by zero in damcon_robot_move()

This patch will avoid the divide by zero in mat33_inverse_transpose_ff(), but I'm not sure it does the right thing. I'm not quite sure why the matrix we're trying to inverse...

* 801331dbb20c73bd7137a7c45fe97a39790fbaef Avoid divide by zero in render_entities() when culling small objects

* c5fa0d3a5a65aeb846325ce2a054fb1da7150775 Do not use model normal matrix in graph_dev_draw_nebula() I think the following two fix the root cause of the NaNs in mat33_inverse_transpose_ff() -- some components of entity->scale could...

At this point, I'm not aware of any more sources of NaNs. It will run for quite a awhile with "--trap-nans" option active (Haven't tried overnight though.)

Holy crap! Awesome! Haven't looked at the mission script yet. Congratulations on being the first person to create a mission script other than me!