So far we've seen lots of references to the Americanized version of Roubanis (which misses some of the details I am familiar with, at least listening to the Mandolirium version), but little to indicate provenance beyond that. Normally that would provide corroboration for the claims that Roubanis wrote it. But I still doubt that. And nobody has explained the existence of an early Sephardic version, according to Q's link (which version, by the way, DOES have the details I remember, which rather causes me to think that Roubanis SIMPLIFIED the original for western tastes as part of his adaptation).
The Greeks would never use Misirlu for a circle dance, by the way - we would consider it inappropriate, too "Arabic" for that. As somebody mentioned, the rhythm is the beguin, and the Greeks would (do!) use it for facing-pair freestyle dance, which we have copied from the Arabic countries via Turkey.
The name "Misirlu" gives conflicting clues: "Misir"-unhappy-Latin root Supports the Sephardic theory. The ending "lu" however is the Turkish "li" (without the dot over the "i"), that commonly turns a noun to an adjective ("misery" to "miserable").
Listening to the tune, I would classify it as "Mediterranean folk"; could be Arabic (Lebanese, Maroccan, Egyptian) or Hasidic/ Sepharding. Almost certainly not Greek (I would bet money on that actually). But it appeals to the peoples of all the Mediterranean countries (even Southern French and Italian, though they would deem it more "exotic").
Me, I just like it...and remember times in my misspent youth humming it with friends on deserted beaches during the magic hour between evening and night.