When fired, the primer is pushed out of the case and conversely the case is pushed forward to the limit allowed by the head space. The hammer does not restrain it. There is a simple experiment you can perform here. Simply load a primer in a case and fire the empty case with primer in a revolver. The primer will back out, quite possibly tying up the revolver. For this reason, cases for blanks and primer fired wax or plastic bullets have enlarged flash holes. Back at the dawn of time there were pistol designs which used the force of the primer blowing back to actuate the pistol mechanism. In a normal load the primer is reseated when the powder ignites and the case is pushed to the rear. Note that the case DID NOT stick to the chamber walls.
In a blow back action, the case and breach block are blown to the rear. This action would not work if the case stuck to the chamber walls. Over the course of time, various attempts have been made to corrugate the chamber walls, etc. in an attempt to get the case to stick and retard the blow back. They have met with VERY limited success. Its quite obvious that if the case were to stick to the chamber walls a blow forward action would not work. The force tending to push the barrel or cylinder forward is due to the bullet drag on the walls. In the blow forward action the case is free enough to allow the blow forward force to overcome not only the inertia of the barrel but the force of the return spring.
A gas in a container can exert a force on the walls only in a direction normal to the wall at any given point. Hence the pressure tends to cause the chamber to expand, but it cannot produce any significant force tending to push the chamber either to the front or to the rear with a straight case such as the typical revolver case.
As for action and reaction, we are dealing with three components here, not two, the bullet is pushed forward, the case and recoil shield are pushed back, and the third component, the cylinder, is dragged forward with the bullet. Any physical system can be modeled in any of several equally valid ways. Some ways make a given point more obvious than others. If instead of action/reaction, you examine the pressure induced force balance, you will find a force in all directions at right angles to the axis on the chamber walls. The forces in this class sum to zero, producing no net movement. The force on the base of the case is balanced by the force on the base of the bullet so they tend to move in opposite directions. The drag of the bullet on the chamber/throat walls is the only unbalanced force acting on the cylinder so it tends to move forward.
Finally, all this should be made obvious by the fact that in S&Ws we are forever stretching yokes or adding shims to compensate for wear at the yoke tail and cylinder well bottom which limits forward travel, but almost never replacing the ratchet which limits rearward travel. We rarely have problems with headspace. (I am excepting ratchets replaced because of pawl wear which occurs during rotation of the cylinder and has nothing to do with the discussion here.)
The revolver does not normally tie up because the case is springy and after firing the case contracts and is free to move within the chamber.
Its true the 22 Jet would work well enough if kept religiously clean. Other bottle necked cartridges in revolvers have done better as the 22 Jet case has a shape designed to provoke set back problems. But the point is not howwell it did or didnt work, but that the problem was one of shoving the cylinder forward and not allowing it to fall back thus tying up the revolver.
In the case of a recoil operated arm, the entire breech block, chamber, and barrel assembly recoil as a locked unit until the unlocking operation occurs.
All this said and done, we still dont know what the correct end play clearance is for a DW revolver.