In addition to Colyn's warning, which in my experience is quite correct, I'd also add that I've noticed that many Canon lenses seem to front-focus slightly when used on non-Canon cameras.
Recently I came across some information in a reprint of a Canon factory service manual which seems to suggest that Canon used slightly different estimates for film curl and thickness than Leica did; this easily could cause small focus disparities between Canon and non-Canon bodies, although these disparities probably wouldn't have shown up except with ultra-wide-aperture lenses. (This wouldn't have been an issue for the 50/0.95 'back in the day,' since its special breech-lock mount precluded its use on anything but a Canon body!)
Anyway, the upshot, as Colyn said, is that you shouldn't be surprised if you need to have your M-converted lens custom-tailored for optimum results on a specific camera body.
I had front-focusing on a 50/0.95 that I had converted for use on my Epson R-D 1 camera, and was able to match it to the body by slightly thinning down the lens' "collimation shim" -- an internal brass ring that determines the depth of the optical section in the focusing mount. I measured the amount of front-focus by taking test pictures of a metric ruler, then used standard focus-extension equations to determine how much closer to the body the lens would need to be moved for correct focus.
Then I removed the optical section from the focusing mount (easy thanks to the modular construction of Canon 50s), extracted the collimation shim, and thinned it by rubbing it against fine abrasive paper on a flat surface. To avoid overshooting the mark, I'd measure the thickness periodically with a digital caliper, and would stop frequently to reassemble the lens and make more test pictures to "creep up" on the correct adjustment. (Obviously this is easier with an R-D 1 than a film camera, since you can view the test shots immediately on the camera LCD at high magnification.)
Other than getting bloody fingertips (the edges of the collimation shim get sharp once you start rubbing it) I didn't have any problems doing this, and the payoff was significantly better full-aperture performance of my camera-and-lens combo. However, if you're not comfortable with delicate tinkering, or can't stand the sight of your own blood, it's surely a better bet to send your camera body and lens off to an optical technician and let him/her suffer, as Colyn is doing.
Incidentally, the amount of "thinning" required was on the order of a couple of hundredths of a millimeter, so you can see the level of precision required (and the level of care Canon needed during the original assembly!)
PS -- Another Canon golden oldie enjoying renewed popularity among the RFF faithful is the 50/1.2; if you try out one of these on an M-mount camera via a screw-to-bayonet adapter, don't be surprised if you find that it, too, front-focuses a bit and would benefit from custom calibration of the collimation shim. You might think you could do it by thinning down the screw-to-bayo adapter, but that won't work -- it would just move both the optical section AND the RF coupling tab together, which would change BOTH the rangefinder-indicated focus point AND the lens' optical focus. What you need to do to cure front- or back-focusing is change the optical focus point relative to the rangefinder focus point, which requires changing the relationship between the optical section and the RF coupling tab.
