Ok, anyway I will try to look at this case and let you know if I find something interesting.
I'm not sure I understand. As far as I'm concerned, the problem is solved. The disassembly I posted shows that this code would be affected by the optimization bug, and the ultimate result would likely match the symptoms seen (the PROPCASE_ASPECT_RATIO value would be out of range, causing the viewport offset functions to return random values, causing the img pointer to point to some incorrect location). I didn't have quite enough information to readily verify the specific value that appears in your logs, but I don't see any need to do that.