Norm and SemiNorm are not clear and probably wrong

[campospinto]

No docstring is provided and it is not clear whether these objects represent the square of a norm, or a norm itself (eg when assembled). Also, it should be possible to add two (semi)norms.

Also, their evaluation after discretization is sometimes wrong: see Psydac issue #campospinto/psydac_dev#12

[e-moral-sanchez]

At the moment the L2 norm of Sympde does $\sqrt{\int_D u_h(x)\cdot u_h(x) dx}$, instead of $\sqrt{\int_D u_h(x)\cdot \overline{u_h(x)} dx}$

I think that the problem is in the Inner class, line https://github.com/pyccel/sympde/blob/fdc19138e86948cb72c8ab8bee7f58020a57dc7a/sympde/calculus/core.py#L510C22-L510C25 where it does c = Mul(*c1)*Mul(*c2)