Added an additional test for structured data

[vtnerd]

The routing_trait module was made pub due to an issue that appears to be a bug. Unfortunately this means there are two names referring to the same object, even outside of this crate. I don't have a solution for both right now.

[vtnerd]

This also only uses the name as the get_name, but doesn't include the owner. The name should be randomly generated anyway, so I'm not sure if adding the owner would help anyway. Hashing to an owned spot can always be an error that generates a new random value.

[dirvine]

Ah I see the issue @vtnerd , it was down to creating maidsafe_types when routing was not around. The RoutingTrait in types was a temporary solution to allow types to be written with the correct trait impl. The answer here is to remove the RoutingTrait from maidsafe_types.
We should get the trait deceleration from routing and not in our own library here.Removing the RoutingTrait here will solve the double object issue (means this lib depends on routing now, which means we need to get it compiling this morning :-) ).

[vtnerd]

So you want maidsafe_types to depend on routing? Do I move just the trait to routing, or everything? And when/if routing eventually needs to call this trait (assuming these types also move to routing, as previously discussed), it will have the same compilation error.