“ could pure attitude control for pointing at different targets continue to be accomplished by a set of gyros and reaction wheels (electrically powered by solar”
That’s what JW is doing right now . We thought of that . Thruster propellant is NOT used for pointing- a total waste of a constrained resource, mass. The consumable is only consumed where a shortcoming of the reaction wheels (a non-consumable) raises its head.
As a megaproject, such an observatory has extra reaction wheels, for redundancy. Should the mission fall to two wheels- insufficient authority for three axes- then an emergency mode might- might - be resorting to the “trim flap”. The momentum flap counterbalances the craft against solar pressure, because the shield area is asymmetric. Trimming the flap, like an aircraft trim tab, might allow some attitude moment. But this would require that some wheels fail, not others- can’t tell prima facie . Even then, the flap force would be low; the observing program would have to plan targets of favorable attitudes only.
There is no orbital maintenance once propellant runs out. Gyros are for attitude, and cannot do translation. Thrusters are overwhelmingly for translations (thrusters have short moment arms) and are thus conserved for translations .