Improve performance network.copy

[eendebakpt]

The network.copy makes a deep copy of the frequency object, and this copy happens again in the Network constructor.
This PR removes the explicit copy of the frequency object and adds a test.

Note: are the copies at https://github.com/scikit-rf/scikit-rf/blob/master/skrf/calibration/calibration.py#L245 required? If not, we could remove them as well

[jhillairet]

I guess that, originally, the copy() was made to avoid modifying the frequency objects of the copies if the original one was modified... Can you check it is not the case after the change?

[eendebakpt]

I guess that, originally, the copy() was made to avoid modifying the frequency objects of the copies if the original one was modified... Can you check it is not the case after the change?

I added test cased to cover this

[FranzForstmayr]

Just thinking loud....
Do we still need copy(), as a frequency object should be immutable?
I know that we intended to change the frequency object to be immutable, but I'm not sure if we checked every case.

[jhillairet]

Just thinking loud.... Do we still need copy(), as a frequency object should be immutable? I know that we intended to change the frequency object to be immutable, but I'm not sure if we checked every case.

You're right, was supposed to be done in #700. But it seems that changing an element of the array is still possible, as done in the test of this PR.

[eendebakpt]

Just thinking loud.... Do we still need copy(), as a frequency object should be immutable? I know that we intended to change the frequency object to be immutable, but I'm not sure if we checked every case.

The Frequency object is not immutable right now. One can change the unit and also the frequency (via the Frequency.f numpy array). There are three setter methods, of which 2 are already deprecated. I created #878 to see whether we can actually make it immutable (except for the unit).

[eendebakpt]

@FranzForstmayr @jhillairet I think this PR is ready. The Frequency is not immutable right now, and will not be during some deprecation time.

[FranzForstmayr]

Thank you, I'm fine with this change 👍🏻

[jhillairet]

Thanks!