I am currently working on the angular momentum part (chapter 3) of Sakurai's QM 2nd. From eq. 3.159 and eq. 3.154, the book says $[J^2, J_+] = 0$ and $[J^2, J_z] = 0$, but I don't understand why $| l, m \rangle $ is not an eigenstate of $J_+$ (eq. 3.164)? Doesn't the commutation relation tell us $J^2$ and $J_+$ share the same eigenstates, so the eigenstates of $J^2$ are also the eigenstates of $J_+$?
Mathematically, if we apply two states $ | l, m \rangle $ and $ | l', m' \rangle $, $$ \langle l', m'|[J^2, J_+]| l, m \rangle = \hbar ^2(l'(l'+1) - l(l+1))\langle l', m'|J_+ | l, m \rangle = 0$$ Doesn't this mean $ | l, m \rangle $ is an eigenstate of $J_{+}$?