Meanwhile I found that it is more likely to get a snap roll assuming my theoretical scenario with high differerential lift at speed well above 1G stall.
Right, you can prevent it to develop, but I think the issue is that if you get a wingdrop and the wing tucks under, the nose will drop, and in some airplanes it might happen really fast, right - it's not the case of a C172, and as soon as you got the nose down it would be impossibe to recover at low altitude i.e. on final, or after takeoff. So it's still an issue at low altitude.
Also during flare, after using the crab method to compensate for a crosswind, when you de-crab you are cross-controlling close to stall speed, why there's no wing drop or even an incipient spin behaviour? I figure out it may be a slip here, but you are wings levek, not banked as in a slip and being so close to stall, why opposite aileron doesn't increase the angle of attack of the opposite wing? I guess it is not yet fully stalled tohave the tendency to drop a wing, right?
To answer your last question first, the reason you don't get a wing drop is because the inputs in a crosswind landing, even if using full rudder and aileron, are not increasing the the angle of attack enough on the retreating wing to stall the wing. What you say about landing level is confusing to me, I land in a slip during the x-wind, slipping into the wind at the same rate as the wind is trying to blow me off the runway. I surely don't land level - my upwind wing is down, the upwind wheel hits first and I am not level until all wheels are on the deck, and I have stopped flying.
Go do some falling leafs, you will understand more what I am talking about. Try it in a Pitts - go up to 5000' AGL in the Pitts, - which drops a wing very quickly when stalled uncoordinated, btw. Pull the stick full back and hold it there. The aircraft will stall and the nose and one wing will probably drop. The nose will bob up and down, and you can use the rudder to bring up the low wing. After some practice you will be able do this for a very long time, dropping several thousand feet, holding the stick back all the time, bringing up one wing and then the other, dancing on the rudders. You will be using full, fast, hard rudder inputs. You can do it power on and power off, it's more challenging with power on. It's fun, and once you get good at it, you won't ever again worry about an accidental spin. You will have a feel for the plane, and won't need airspeed or AOA's either. You will feel uncoordinated flight, and when you are nearing the critical angle of attack. It's nice to know you can still fly safely if the gadgets in your plane stop working. My instructor made me learn to do landings with the airspeed indicator covered, using attitude and feel to control airspeed. That''s the way I was trained, anyway, and it still makes sense to me.
As far as the first question, if you are so low to recover from a stall where you get a wing to drop, then you are probably too low to recover from the stall anyway, much less a spin. You are not really in a spin by definition until after two full turns, and if you can bring a wing up using the rudder, you are preventing the spin. You are likely to add rudder input as soon as you feel the wing start to drop, once you have some training under your belt. If you still hit the ground, you stalled to low. If you are trained and practice to use your rudder instead of aileron when a wing drops, then your reaction will be instinctive, whether it's a Pitts, RV or 172. But too low is too low, and if you stall to close to the ground to regain enough lift to prevent the aircraft from contacting the ground at a dangerous rate, then you are too low, period, and it doesn't make any difference whether a wing drops or not.