Why B(tB) - A(tB) = rAB?

Here's why -- the whole length B(tB) - A(tA) the photon travels along, is equal to the length of the rod rAB in the stationary system plus the length A(tB) - A(tA) which the beginning of the rod A covers during the time of travel of the light until it reaches B:

B(tB) - A(tA) = rAB + (A(tB) - A(tA))

or

B(tB) - A(tB) = rAB


or even simpler

B(tB) - A(tB) = [A(tA) + rAB + v(tB - tA)] - [A(tA) + v(tB - tA)] = rAB