0:36:22stassatsso everything is unwound except it ends up back where the signal hit
0:45:43stassatsif that's some fundamental problem with rosetta, unable to unwind from syscalls, then i can live with that
0:46:59stassatsbut just exactly does it end up back there
1:00:26stassatsand can't reproduce with a simpler test case
1:37:08stassatslooks like if i have two nested interrupts and unwind from the second one to the first one, the first one returns to the places where the second one would've returned
2:03:08stassatsi don't even know if our assumption that returning from a nested signal handler should return at that nesting level is right
2:26:04stassatsi supposes a solution would be not to run any lisp code from signal handlers, but use arrange_return_to_lisp_function
5:01:26stassatsi hooked up unwind-protect with signal handlers, making it not unwind across the signals, but continue unwinding when returning from the handler