OK, I'm trying to understand some things in this thread. Number one, why are people saying right off that a Schoeps supercardioid isn't the optimal "M" microphone for an M/S pair? The "M" microphone can be any type that would give you a good mono recording by itself; there is no pre-ordained ideal type or pattern for "M". I've seen a few on-line descriptions of M/S that specify cardioid as if it was the Only Officially Correct "M" pattern, but they are simply mistaken both historically and technically.
The main point of M/S, when Lauridsen introduced it in the 1950s, was its guaranteed compatibility between mono and stereo. The mono version of the recording came directly from the "M" microphone; you were supposed to use whatever microphone for "M" you would use for an optimal mono recording. Back then, good engineers all still knew how to make good-sounding mono recordings, and cardioid microphones certainly weren't their only choice for doing so!
As far as the match between "M" and "S" microphones is concerned, the CMC 641 and the AKG C 414 B-ULS (or equivalent) should be excellent together. Both have basically flat, neutral frequency response and clean polar patterns throughout the audio frequency range. The overall frequency response of "M" and "S" microphones ought to be generally similar--otherwise some distortion of the stereo image will result, as different frequencies from the same instrument or voice can have slightly different apparent directions of origin. But you shouldn't have this problem with these two microphones.
Unfortunately, "414" refers to any of over a dozen (!) different AKG microphone models using at least three distinct capsule types, some of which have rather different high-frequency response characteristics from others. I would recommend renting one of the types with basically neutral high-frequency response, rather than one of the types with elevated high-frequency response.
It would be great if you could record M and S directly, without dematrixing them. I don't know whether your recorder lets you monitor in L/R stereo while you do this, but many recorders and preamps do so; it's well worth finding out. If so, you can postpone the dematrixing until you're listening to the playback over your familiar loudspeaker setup at home, and you can then choose the relative levels of M and S that are going into the matrix. This will let you set the stereo soundstage width and the reverberation balance to your liking, with the recording already "in the can." That adjustability "after the fact" is a big part of what makes M/S worth using! Also, this approach cancels out all worries over the M and S microphones' relative sensitivity or noise. Just record the two tracks as usual, setting the gains so that your peaks in both channels leave you a couple of dB for safety below 0 dBFS.
The only other major concern I'd have is your miking distance. Will you have a choice about that?
--best regards