After program abort, show whether G61/G61.1/G64 settings are
consistent with machine state at last executing motion prior to abort.

This test runs and aborts a test .ngc program several times with
various motion mode and blend/naive CAM tolerance settings.  Checks
are performed at various times relative to program execution:

1) Motion mode:  Does stat buffer agree with programmed mode?

2) G64 tolerance settings:  Does stat buffer agree with programmed P/Q
   args?  (Only for G64, and only when these params are available in
   stat buffer)

3) G64 tolerance settings:  Do programmed tolerance settings produce
   expected motion, as evidenced by samples taken during program
   execution?

The .ngc program is aborted before ever reaching a G64 programmed at
the end.  This demonstrates a bug caused by motion and interp running
out of synch, affecting LinuxCNC v. 2.7 and reported in Github issue
#177.

The fix will be introduced with Rob Ellenberg's state tags work,
discussed in Github issue #134.
