I've put a bit of thought into CNCing a shaper. I wouldn't be in a terrible hurry to control the tool angle. You can cut angles and shapes using the cross and downfeed, you will likely need to use multiple tools to do complicated geometry anyway. Here are the requirements that I came up with:
1. No need for ballscrews. My plan was for a retrofit, so finding room for ballscrews would be a problem anyway, but there is no need. You are not cutting circles or complicated shapes such as in a mill, and you are always approaching the cut from the same direction, backlash compensation will work fine.
2. The drives need only be fast enough to make the maximum move during the highest stroke speed. In other words you will only be moving the the motors before and after the cut, not during.
3. There is no need for a clapper. You can retract the tool away from the cut at an angle before the backstroke starts and advance it before the cutting stroke starts. How fast you can do this will determine your maximum cut/stroke (also the lead in/lead out of the tool is part of this equation).
4. As with any CNC machine, changing tools quickly will be key to this machine being useful or a novelty. A quick change tool system will be needed.
5. Since you will need to be able to figure this out in some sort of CAD/CAM, standardized tools will be needed. This is one are that begs for inserted carbide. Not often used on shapers and planers, it doesn't mean they can't be. Grade would need to be as soft as possible to withstand the interrupted cut.
6. A fancy custom controller is not needed. A Beaglebone Black running machinekit was what I was thinking of using. You would need to put some sensors on the machine to give positional feedback of the ram in addition to the normal limit switches.
7. The real problem is the CAD/CAM. You only really need the G0 movement code, as you will always be moving the tool at the max the drives can move it. Figuring out that movement and writing the G code efficiently is the problem. With my limited software skill/desire/patients for programming I was thinking of doing something in Excel that would allow you to string together facing at angles, slots, curves and dovetails to build the code for the machine. I'm sure someone else out there has the skills to do something much better than that.
I already have the test subject to try this on, an 8 inch Lewis kit shaper. But there are many more projects ahead of this one, so I don't expect anything to happen on it anytime soon. If it worked on well on the Lewis, I was going to put it on my 16" G&E and really move some metal.