Followup to #6940

Can't compile with PINS_DEBUGING enabled on Atmega1284p because it tries
to enumerate pins 46 and 47 (which don't exist) because `#ifdef
AVR_AT90USB1286_FAMILY ` doesn't work as expected.

Fixup length

Fixes #7150

- Added `H` to set a nozzle height adjustment to surface of the bed.
- Updated G29 P4 comments to reflect changes.

Addressing #6992 and #5851

* fix_planner_refresh_e_positioning

* Fix

Use the raw position for mesh measurement. Otherwise the `M206` and
`M92` Z offsets will get canceled out by bed leveling. The downside is
`G29` will not compensate for a poorly set small `M206` fudge value.

To elaborate on this issue, imagine you are probing with a Z home
offset of -0.1, meaning when Z homes, -0.1 is the current position,
implying the ideal bed zero for the nozzle is 0.1mm higher than the Z
endstop. Ordinarily when printing, Z would raise 0.1mm higher.

What happens when we probe is that all points are measured with that
-0.1 included. So when bed leveling is enabled the `M206 Z` offset gets
exactly canceled out by the bed readings.

- Fix: When "100.0" changes to "99.0" the LCD shows "199.0"
- Use 2 rows if needed on character LCD, (allowing longer labels…
Germany, et. al.)
- Known issue: A certain length label combined with a certain value
drawing function could, for example, display 99.0 on 1 line, but 100.0
on two lines. Workaround would be to pass a nominal value size argument.

* Add Z Fade Height to M420 Output

Echo the z-fade height when M420 is called, if it is enabled.

* Fix Whitespace

* Fix Compiler complaint

* Missing semi-colon

=====================

add test to Travis

Addressing #7057