Logical attacks on smart cards have been used for many years, but their attack potential is hindered by the processes used by issuers to verify the validity of code, in particular bytecode verification. More recently, the idea has emerged to combine logical attacks with a physical attack, in order to evade bytecode verification. We present practical work done recently on this topic, as well as some countermeasures that can be put in place against such attacks, and how they can be evaluated by security laboratories.