If the work has all been done, and the first author will continue to take responsibility for the paper throughout the submission process (which may include multiple rounds of revisions, some with new analyses or even new experiments), there is no reason for the authorship order to be altered.
If the original first author is not willing to participate in all these steps, there will eventually develop some gray area where even if that person did most of the original work, they can't really be responsible for the sum total of the paper anymore. In that case, it may be more sensible to have someone else take the first author position. It would not be fair, though, for that author to be somehow prevented from participating. Also, if I were advising the supervisor rather than the student in this case, I'd strongly encourage them to let the student keep first authorship. It's just not good business sense to be a jerk.
Ultimately, if there is an authorship dispute the primary recourse is to oppose publication of the paper. Assuming everything in the publishing process works how it's supposed to work, papers can only be published when every author agrees with publication, including authorship order. At that point, it's basically between the parties to decide whether their place in the authorship order is worth not having a paper published at all.
(as a side note: present affiliation doesn't really matter at all; I'd expect your affiliation to be listed as the place you did the work originally, possibly with an added affiliation for your current location if still in academia, otherwise probably nothing else Affiliation on a paper written mostly in previous position Changing affiliation on publication )