Yes, I was talking at the 24mm end of the 17-40; at the bottom range, the 17-40 has a LOT of barrel distortion--around 3.8% which many would consider "horrible", at least for architecture or city photography, but it also can give a real,true wide-angle look to its images on FF. Much more so than the numbers might indicate. The 24-105 L has around 1.8% barrel at 24mm, and a pretty fair amount of vignetting as well. Maybe when you get the new body and lens you can evaluate just how w-i-d-e 17mm on FF really is, and I think you'll see that each distinct millimeter 17-18-19-20-21-22-23-24 on FF is almost like a different lens unto itself.
It's hard to put into words the difference, since on FF the sensor is roughly 2.3x larger than a 1.6x Canon's sensor, so the depth of field, the camera-to-subject distances, and the "rendering" of the image is *significantly* different on FF than on APS-C, even with "equivalent" angles of view. The difference is not simply angle of view--you have working distance difference and therefore perspective differences between APS-C and FF, and you have a huge difference in format size as well. The trend is to look at things in terms of *ONLY* equivalent angle of view between 1.6 and FF, but the difference is much more than a simple mathematical equivalence in FOV based on taking a focal length and going "oh, I multiply by 1.6x and everything is equivalent." That's not how it works,by a long shot. Equivalent does not mean "equal photograph" in terms of several factors. Once you have the 5D-II you'll be able to judge for yourself. I hate to see you take a bath on a new lens that you'll quite possibly want to have back in the near future.
It's a shame the kayak and saltwater dunking you took with the other body and the 100-400...bummer deal on that.