Not sure if this is stalled because it's not convincing enough, but if it helps, the debugger check is not merely conjecture. Disassembly around the crashing site is:
``` 4d97fd: 68 e8 13 19 01 pushl $0x11913e8 4d9802: e8 c9 98 51 00 call 9f30d0 4d9807: 83 c4 04 add $0x4, %esp 4d980a: a0 90 01 00 00 mov 0x190, %al ```
where 0x11913e8 points to a string "Causing exception to test for debugger..", and 0x9f30d0 is presumably some sort of logging function.