A couple of points. Square aluminum tubing isn't guaranteed flat anymore than steel is. Extruded T slot might have specs on it but it isn't absolutely flat either. This is why some suppliers offer to machine rail mounting surfaces flat. Even Aluminum can move after machining, it is something we have dealt with at work. As for steel tubing yes there is a chance of movement after flattening the surface. If it makes you feel better you can have it normalized.

As for your grinder I didn't even imagine anything like that at all. Rather I was thinking of a vertical spindle that would somewhat approximate an end mill. Well at least for the part of the spindle the grinding belt wraps around. The obvious problem here is that in a conventional 3 axis machine you would only address one side of the piece. This you would need a rotary axis setup to revolve perpendicular to the X. So your part could be machined similarly to the way and endmill might cut it. Given that the part to be machined would sit on a pedestal you should be able to machine the top, sides and maybe even the ends. The radius of the grinding spindle working roller would dictate feature size.

This may be hard to imagine. However I see it as fairly simple to implement because the bulk motions would be just like milling with the side of a mill cutter.