I came across a situation where a client just connected to a large shared dataset and the syncing (downloading) was taking a while.
He started modifying (renaming) items in the shared folder but the changes weren’t shared until the downloading process was completed. It seemed to me that the client has just a single to do list which gets only updated once it completed one item of it.
How about having several sync lists:
one for downloading which gets filled continuously every new file created on server side.
a second one for changing files (renaming, moving, deleting)
and a second for uploading.