> fully document this change.
I'm waiting for a bug report because I bet it is a bug and not the deliberate choice to map IndexError et al to quit. Consider that code that catches these errors already was broken with `-d:danger` too and for years `-d:release` was what is now `-d:danger`.
