Oh, one more reason I do not believe "nested attack" should be a thing we interpret - such treatment would virtually neglect the bonus attack limitation per turn.
During the aftermath of primary attack, before said attack would be said to have been performed, one would perform a bonus attack. But this bonus attack would then have it's aftermath in which, theoretically, another bonus attack could be performed.
We know for a fact one can only perform one bonus attack per turn, but if performing bonus attack would've been a part of performing an attack, then analogically performing additional bonus attack would have to be a part of performing the previous bonus attack.
(Attack (Bonus Attack (Bonus Attack (Bonus Attack (...))))) could, theoretically, go on forever because for each additional bonus attack added there would've been no previous bonus attack which has already been performed.
There are multiple problems and ambiguities when nested / recursive logic is incorporated. Problems definitely too complex for a board game rule set to handle.
I advocate for reasoning that aftermath is an after math indeed, takes place after the attack and itself is a part of the procedure of performing the attacker but not a part of the attack itself. A procedure I, intuitively, see as choosing a weapon, choosing target, making the attack and finally resolving the aftermath of the attack.
Edited by Ryfterek