Certain Irish lines are good candidates, but not the one indicated in the above link. These early Irish lines have to be examined on a case-by-case basis, and the Ui Chennselaig line of Leinster (of which the famous "Eve of Leinster" was a member, as well as the above Crimthann) can only be accepted back to about ca. 850 before the chronology becomes suspicious. Much better is the line of Eve's mother Mor from the Ui Dunlaing dynasty of Leinster, which includes descents from several well-documented marriages of the late seventh and early eighth centuries, leading to several lines which go back to the late sixth century at least (and probably a bit earlier).
Probably the best example if you also want to factor in the quality of the documentation.
> The kings of Dál Riata could also be a candidate.
Maybe, but the quality of the documentation for this line during the eighth century is not that good (but also, in my opinion, not as bad as some of its critics have claimed).
> If you are OK with gaps in the line; then it is almost assured that the Armenian rulers were descended from the Arsacids in Parthia which could get you back into the B.C. Maybe someone will dig up better evidence here.
Only if you allow the possibility that the "gaps" are REALLY large. On this basis, I could get us an extra millennium by claiming a descent "with gaps" from Ramses II of Egypt on the following basis: He had more than 70 documented children, and therefore presumably had thousands of descendants after a few hundred years and millions after a millennium or so. By now, his list of descendants would undoubtedly include everybody in the world (with the possible exception of small groups that remained isolated until relatively recent times). If you make the reasonable demand that the "gaps" should be relatively small, and that the descent should be at least approximately known within those gaps, then even the supposed Armenian example becomes troublesome, because the conjectured Armenian lines involve certain key intermarriages which cannot be convincingly demonstrated to be part of the picture in many of the cases where this example is put forward. The most plausible such gap-filled line goes something like
Parthian Arsacids > Armenian Arsacids > Gregorids > Mamikonids > Armenian Bagratids > Georgian Bagratids
the last of which go up to the early nineteenth century. However, I have never seen a plausible attempt to get into medieval Western Europe without a huge amount of conjecture.
With regard to Cerdic of Wessex, with whom this thread started, his candidacy should not be taken seriously. Even if he existed at all (possible, but not certain), the evidence is too late and contradictory to regard the line of descent as plausible.
Also worth mentioning is the line going back through the kings of Dyfed in Wales, which goes back to a king mentioned as a contemporary by Gildas (writing probably in the first half of the sixth century), with the earliest surviving version written probably in the eighth century (surviving in independent Irish and Welsh versions which agree on the important parts). Not as well-documented as some of the Irish dynasties, but not too shabby either.
I am much less well informed on Asian dynasties, but based on what I have read, the case for ancestry of the Japanese emperors would appear to be somewhat comparable to these European dynasties, like the Irish genealogies blending from history into myth and legend in a way that the boundary between them is difficult to determine.
Stewart Baldwin