If you look at the revision history metadata you'll see that the order of commits is actually tracked by following prov:wasRevisionOf so the information is there to reconstruct the exact order but I suspect that date is being used as a proxy because of the difficulty in constructing these types of queries that Evren wrote about in a recent post.
I would think that the result order from the sparql query could not be used for an absolute ordering in the case of identical commit times and you would have to follow the wasRevisionOf links to get that.