Without resorting to calculations, the trajectory seems to put cash in the neighborhood of $200b within three or four quarters. Plus or minus a bit doesn't much matter. Soon or sooner.
I don't know the breakdown of overseas money vs. funds held in the U.S. offhand, but it seems to me more political gamesmanship than anything else when corporations argue for repatriation holidays. They will shift the money to the U.S. when they need or want it, either way. Waiting for Congress to give you a special deal may be good politics, but it is bad business. Apple should and I think will focus capital investments where they have the best growth opportunities. That's probably not in the the U.S. right now anyway. And keep in mind, they have more than enough capital to expand wherever they want, so I don't think where they have the money stashed is a huge issue.
We don't know if the buyback program the board initiated last year was ever spent out. At this point though they should probably be buying back/supporting the stock price when the multiples drop way low, as they have recently.
The reparation issue is actually a pretty significant business issue. Congress (especially the Republican Party) has suggested they would be open to another tax holiday. Since that seems possible, all US companies who can afford to are holding cash overseas in case this happens again. Now, or course you don't hold off on important investments for this. But as you point out Apple has plenty of money to do all the US investments it wants. Apple will basically never bring that cash back if it means paying 35% on it. At least not when changes to the tax code are so frequently discussed. And yes the overseas cash is most of the cash pile because Apple actually uses cash here in the US for development and to pay its dividends.
The buyback program is ongoing. But it is modest and mainly based on buying back shares to cancel the shares that are issued to executives as part of their competition. It isn't focused on any particular stock price or earning multiple.