It really depends on your motives. If you will be needing to port to new platforms in the future, it's better to have a hard-and-fast crash now, so you can learn and avoid the undefined behavior now, rather than face a large bug backlog years down the line.
But you're correct that sometimes it can be expedient to exploit such technically undefined behavior. (I've committed this sin myself, most commonly in serializers/deserializers)
