I would personally think of it this way: Apply the 1.6x sensor factor to the RB's normal 90mm lens, because you like a lens that has the AOV of the normal lens times 1.6. Easy. If you'd like, apply the factor to 2x50mm instead, for the theoretical 100mm RB lens that does not exist.
Here is a more detailed explanation: 50mm is the common normal lens for standard frame size on 135 format. The AOV of a 50mm on your camera is the same AOV as an 80mm lens on 35mm. Compare 80mm to 50mm. 80mm is 50mm + 3/5 of 50mm. Therefore, you like a lens that has the AOV of a lens 160% the focal length of a normal lens. Multiply the RB 90mm normal lens by 1.6, and you have your ideal FL: 144mm. So, I would look at the 150 (or 140 macro), based on numbers alone.
Since the RB frame is fatter, it appears wider, so I suggest looking at the 180mm as well. In fact, I am certain that the 180mm will indeed "feel" the most like the 50mm on APS-C.
Your figurin' is correct, though. The 2x factor is used for comparing 6x7cm to 24x36mm format, not for comparing APS-C to 6x7. Since you are using APS-C format, if you do what Keith says, and just multiply the 50mm FL by two, you don't get a similar AOV on 6x7. 90mm will be a normal AOV for 6x7, not medium long like your 50mm on APS-C. You have to 1) do what you did, and multiply 50mm by 1.6, and then by 2, or 2) condense the two factors ahead of time into one factor that goes directly from APS-C to 6x7 (a factor of 3.2).
Remember that the aspect ratio of a 35mm or digital Rebel is different than the RB, so you may want to favor either a horizontal or a vertical angle of view (probably horizontal if you shoot mostly horizontals, and vise versa), and compare lenses that way instead.