When you make a Ctrl-c break you are presented with the following choices:

how do you want to proceed? [(S)kip current test/(e)nd detection phase/(n)ext parameter/(c)hange 
verbosity/(q)uit] c

After choosing (c) and entering a value, what basically happens to the test phase is unfortunately the same as if (S)kip current test had been requested.

Due to the linearity of the relevant code and the way the exception handling is currently scoped, it looks like a non-trivial amount of refactoring might be required to get fix this. But I might have missed something, anyone please prove me wrong ;)