Get active set from solver

[gianluca-garofalo]

No description provided.

[darnstrom]

Thanks for this contribution!
Since it changes the result struct, I am a bit hesitant in merging (I am trying to keep the structs as slim as possible to keep the API minimal.)

The (non-weakly) active set is encoded in the dual variables that are already in the result struct. Simply checking which componentd lamda are non-zero should give you the (non-weakly) active constraints. Also note that the signs of lambda encode weather the upper or lower bound is active. A "get_active_set" function that takes in the result struct and extract the active set from the dual variables could therefore be an alternative.

But if there is a good argument for including the active set directly in the struct, I am open for merging.

[darnstrom]

This is a great addition to the Eigen interface! Thanks :)