@forest
Your are right to specify that a sinewave is needed.
An inductance is an inductance whatever the way it is made. But if it is non linear and depends on parameters, as an inductance with an iron core is when near saturation, the method will work while the signal is in a range where the inductance can be considered constant, for example the primary coil of a mains transformer tested with a 1v pp signal. The measured inductance will give the inductance for this range only. If a cc current was superposed and saturated the core, another correct value would be measured for this new condition, always providing that the voltage range from the signal generator is low enough to not modify the inductance.
The method is based on equation E =  L*(dI/dt). If L is not constant E = d(LI)/dt =  L*(dI/dt)  I*dL/dt: the method fails if I*dL/dt is not negligible in comparison with L*di/dt, and so "I" must be low and consequently the voltage from the generator also, so that I*dL/dt~0.
