I wish to generalise the equations governing electromagnetic induction. Consider a loop whose size is changing, and is subject to a changing magnetic field. Then,
Definitions: $$\phi \equiv\int_{S(t)}\mathbf{B(t)}\cdot d\mathbf{a} \tag{1.1}$$ $$\mathrm{emf} \equiv\oint \mathbf{f}\cdot d\mathbf{l} = \oint \mathbf{(E+v\times B)}\cdot d\mathbf{l} \tag{1.2}$$ Laws: $$e=-\dfrac{d\phi}{dt}\tag{2.1}$$ $$\nabla \times \mathbf{E}=-\dfrac{\partial \mathbf{B}}{\partial t}\tag{2.2}$$
If the loop is fixed ($v=0$), and the magnetic field changes with time, then it's the $\oint \mathbf{E}\cdot d\mathbf{l}$ that produces the emf (transformer emf).
If the loop boundary changes (for e.g. a metal rod sliding on metallic rails, with speed $v$) and the magnetic field doesn't, then $\nabla \times \mathbf{E}=0$, so $\oint \mathbf{E}\cdot d\mathbf{l}=0$, and only $\oint (\mathbf{v \times B})\cdot d\mathbf{l}$ is responsible for the emf (motional emf).
For these two cases individually, it seems so far so good.
Question (part a): Is this generalisation indeed correct?
What's particularly bothering me is the equation $\nabla \times \mathbf{E}=-\dfrac{\partial \mathbf{B}}{\partial t}$. If we were to put it in the integral form, (using Stokes' theorem), there seems to be no reference to $v$ or $B$, which seems to be incorrect as per definition $(1.2)$.
A possible explanation for this is that Stokes' theorem does not hold when the surfaces/boundaries in question also change.
The lack of literature on the most general case ($B$ changing and the loop changing), makes it difficult to verify this generalisation completely, and brings me to:
Question (part b): Using the "definitions", can one of the "laws" be derived? By assuming the other one as an axiom? (or shockingly, derive both the "laws" from simply the definitions: I am almost certain this is blasphemous)
The basis of this question is more or less mathematical in nature.