For parity with bugfix-2.0.x

Followup to 13a3fb1acc

This PR will address two issues reported in #8492. Just one is left because I don't know if it is an issue or not

1) when no home is done ("no_move" is set) tool change should never move X & Y because... I don't know where I am...only Z should be adjusted to avoid bed collision if other nozzle is lower than current.
2) Since no move no bed level matrix should be applied
    When SWITCHING_NOZZLE is enabled and system is homed there are two movements to go to destination, but first can move Z to older position and then crash on the bed

This is a non-intuitive change.   @oldmcg says in this post https://github.com/MarlinFirmware/Marlin/issues/7572#issuecomment-340641207 that using DELTA_PRINTABLE_RADIUS is better than DELTA_PROBABLE_RADIUS for setting the probe limits.   This is because the automatic probe code checks all locations for being reachable by the probe prior to doing the probe.

Fixes #7978

* Fix up Travis CI and compile warnings/errors

* No ULTRA_LCD with REPRAP_DISCOUNT_SMART_CONTROLLER

* No point in setting all the endstop plugs

This fixes the old implementation of M428 which was broken, did not
match the website, and made no sense.

Thanks @RowanMeara!