- Making M665 compatible with repetier (see
http://reprap.org/wiki/G_code#M665:_Set_delta_configuration)

- M665 B also sets the radius for manual calibration menu

- Converting tower ajustment definitions to arrays - tower angle
corrections compatible with Esher 3D wizzard

- Only tower angles need to be adjustable with M665 and stored to EEPROM
- tower radius and diag rod can be adjusted in the FW only with #define

- On `DELTA` the `M665 H` option supplants `M206`
- On `DELTA` `NO_WORKSPACE_OFFSETS` only reverts `G92` behavior
- Spawn 4 conditionals based on `NO_WORKSPACE_OFFSETS`
- Optimize coordinate space conversion for `DELTA` workspace
- To keep EEPROM version, retain `home_offset[XYZ]`, just ignore XY

Firmware was forgetting the z_fade_height in Planner, set by M420 Z<f>,
after each reset or loading settings from EEPROM. Added the float
z_fade_height to EEPROM-Storage, now remembers the value.

Pretty much...  The code is in place.  Still more work to do.    But it
has a lot of hooks and variables in other code, so commit and merge
before I pick up a million 'Conflicts'.

- Make all `unified_bed_leveling` data/methods static
- Move some UBL-related variables into the class
- Replace `map_[xy]_index_to_bed_location` with `mesh_index_to_[xy]pos`

Pre-compute the combined position shift and home offset to save a
single float fetch-and-add per conversion. Great for delta/scara and
bed leveling.

1) modified 3 code files
Marlin.h
Marlin_main.cpp
endstops.cpp

2) modified config files so I could test on my machine

Testing was done on an AzteegX3pro based machine.

The probe was hooked to the Z_MIN endstop.

My controller doesn't have a dedicated Z_PROBE input so I couldn't test
that functionality.

Verified that a large file (without any G38 commands) executed the same
before and after the changes.

Verified that the head moves as expected when G38.2 and G38.3 commands
are issued. Single & multiple axis moves were tested along with + and -
directions.

Code was added to the main ISR. In normal operation only one extra IF
statement is evaluated. I didn't notice any performance degradation
because of the added code.

The G38 commands are expected to be issued manually by the operator
during machine setup. The G38 commands wait until the machine is idle
before proceeding. That way the other commands are minimally impacted
by the extra ISR overhead when a G38 command is in the queue.

The G38 commands are very similar to the G28 commands except 1) only the
Z_PROBE is used and movement can be in the + or - direction.

See issue 4677 for a discussion on adding G38 commands to Marlin.
Feature request: add ability to use G38.2 command (CNC)
MarlinFirmware/Marlin#4677

This is in advance of moving non-linear bed leveling to the planner
class.