I'll try to answer as much as I can, as accurately as I can (but I'm starting to stretch my knowledge going into physical chemistry territory -- someone may very well come in and knowledge slap me). Your chemistry book is, in general, correct that the p, d, l orbitals fill with arbitrary spins until all the spins are filled, and then the other "anti-parallel" spins come in. The only reason I say in general is because there may be a couple exceptions, I can't remember either way, to this rule. You see, it all has to do with what configuration gives the lowest ground state energy. This order of filling is this way so that the ground state energy is the lowest. The bit about fermions comes into play because once all the p,l,d,etc. shells are filled with spin states, it is still more favorable to continue filling the ml levels than to add another n level (sorry for mixing quantum numbers, but didn't want s to get confused with spin). However, as a fact of nature, fermions (which electrons fall into the category of) do not share the same energy states. As such, the additional electrons filling the atomic "orbitals" have to pair in an opposite spin.
Something to get at your second question is that the spins of electrons aren't physical spins as you may be thinking of them. Electrons are point charges, so you have to be careful you think of giving them a direction. Really, the electrons are just filling intrinsic states, and the multiplicity of spin for an electron is 1/2, which means it can either be +1/2 or -1/2 state.
For your 3rd question, what's to say the magnesium ion takes the electron you shot into the trap? Furthermore, are you continuously monitoring the spin of the electron you shoot in? If not, the electron can go back to following its eigenvalues, its probability.
Maybe someone more expert than I can fill in the gaps.