Actually, it does make sense to me. The X1C is an amazing machine, in particular with PLA. I would not be surprised if an attempt with PLA would turn out perfect immediately. At least, it would confirm that the problem is not machine related.
PETG is quite a bit more challenging due to its high tack (making the nozzle a material “magnet”) and susceptibility to warping.
The latter would match your description of problem 1 both for the infill pulling in the wall and the infill closure layer (turning a soft infill into a brutally stiff sandwich structure).
However, in this case I would have expected that you should have been able to see a difference when increasing the number of walls.
Printing infill before the wall so that it has already shrunk when the wall comes into play could perhaps confirm this. It might decrease the infill-wall anchor, but at least the wall should be straight if the assumption is correct.