Abstract
We withdraw this note because our calculation of the A(3,3) example, which initially contradicted one of the results of a 2005 paper by Fomin-Fulton-Li-Poon, was incorrect. In the second version of the prepublication arXiv:2303.11653, we explain how the description of the cone A(p,q) obtained by Fomin-Fulton-Li-Poon refines that obtained using the O'Shea-Sjamaar theorem.