Hi,
You already found the one ironclad rule: Don't study Microsoft code, disassembled or leaked C code.
Disassembling the game code is generally OK because we do not reimplement the game.
Wrt your suggestion of comparing the control flow in the game with builtin and native d3dxof: This is getting awfully close to disassembling of the native DLL. A cleaner approach would be to replicate the d3dxof calls the game makes in your own program and see where your program starts to receive different results. The good side effect of this is that you can use this code as a test case to submit along with your fix :-). See dlls/d3dxof/tests for existing tests.
Cheers, Stefan
Thank you, I'll try my best.
Regards, Fabian Maurer