Hi Eric-
I'm an optimistic guy, but in reality, I know that the 37mm effective baselength of the R-D1's rangefinder means that focusing a 90mm at f2 (or even 2.8) is what Tom Abrahamsson might call "hope over experience", meaning that getting accurate focus at close range and wide open (without the 58mm precision of the .85 Leica finders or the 49mm accuracy of the .72 M finder) is really a matter of luck over the ability to repeat the task.
You should be able to do this with these two Leica finders, but with the 90 f2, 75 f1.4 or 50 Noctilux, it really will be next to impossible to do reliably. The tendency to miss focus is usually the user, not the rangefinder. When I am off, it's almost always focused slightly *past* the intended point of focus. Since DOF works as 1/3 front, 2/3 back of focus point (portrait shooters figure this one out the hard way), it's easy to blow focus even with precise finders at close range.
I use my 75mm Voigt with the Bessa R3A (same finder as the R-D1 at 1:1) and it will focus reasonably accurately at f2.5 or f2.8, but the focusing accuracy and consistency will increase SIGNIFICANTLY with the M finders, especially the .85x. Adding the 1.25x magnifier to that will give you a 1.06x finder, something that will not only give you the best accuracy possible, but also make the 90mm brightlines fill the frame nicely instead of the postage stamp in the center.
In other words, nothing is wrong with the R-D1 in terms of defect, it is just a byproduct of the shorter focusing accuracy, reliably out of the range of the requirements for the 90mm. I don't think it's an accident that there aren't brightlines for anything longer than a 50mm (75 FOV) in the R-D1 finder. Returning the camera or lens will not remedy the problem as they all have the same focusing accuracy. Fine with wide to normal, not enough for telephoto, especially at f2. You can see the DOF charts on Leica's site, also, which confirm that there is essentially zero latitude at f2.
If you look at the front of the camera (compared to the M6/M7/MP or Zeiss Ikon), you'll see that the two rangefinder windows are extremely close together. This is the physical baselength for the camera's rangefinder. The accuracy of this finder is improved slightly over the R2a in that the finder is a 1.0x magnification (or 1:1). Look at where the finder windows are for a Leica. They are significantly further apart. The magnfication on a Leica is lower, but the focusing accuracy is greater because the physical baselength that the focusing system uses for triangulation is much greater. You will not see 75/90/135 brightlines unless a forthcoming R-D2 uses a different physical rangefinder, like the one from Zeiss Ikon (also a Cosina made body). You can see this ambiguity when you focus with the R-D1, then rock back and forward slightly, seeing no change in the RF patch. When you do this on a high-mag finder, you'll instantly see the focusing error.
I love the R-D1 a lot, but if precision focus is required with long lenses, I reach for the .85 finder.