Time for China to master the art of "strategic face slapping". If the other party is not conducting diplomacy in good faith some face slapping on purpose is good for deterring opportunistic politicians. In this case China knows not allowing construction will eventually force the ship to disband, so they allowed some face saving for both sides. However if Philipine does not take the L quietly, there is also the option to make them take the L more obviously, which is to block food supply. Eventually the occupant either have mutiny, or need to be rescued off shore.
Such face slapping exercise may result in some whining, but it serves well to deter other nations looking to start up trouble with China. China does not want trouble yes, but this works both ways. Most other country don't like trouble either. China can avoid trouble by displaying ability to create bigger trouble than other party. That is the alternative to the "save face both sides" approach.
One example of this success is the infamous "create the conditions first" interaction between Justin Trudeau and Xi jinping. Despite Canada being the western poster boy the western media all made fun of Trudeau instead of the 'big bad Xi'. The truth is western society values looking tough than mutual face saving. They rather praise Xi looking tough even though he is the 'bad guy'. Gestures like having others save face will instead be interpreted as weakness. The gesture to moderately bully others is seen as desirable and respectable. This may seems tribal and primitive to people of Chinese culture, but this is how western civilization functions.
Emphasize on 'moderate bullying'. The gesture must be strong enough to look tough, but not enough to elicit terror. In this case I recommended blocking food instead of evict them by force. These is some art in balancing the act to fit the description.
