Why do you force push to their branch, instead of committing your fixes as separate commits later on?
Sometimes I will actually create the fixup commits and push them separately to communicate with the submitter, but then squash them into the appropriate places in history before merging.
The only technical issue is when someone has signed their commits, as I obviously cannot re-sign them :-). In that case, I call it out to them and let them re-sign if they want before I merge.