* UBL G29 -P3.1 mesh fill with distance-weighted least squares fit.

* Back to original -O0 on G29 for now.

* Misc. Clean Up

Mostly UBL related clean up.
- But fixed a bug in the thermistortables.
- Made G26 more responsive to user aborts.
- Added sanity checks for older name for UBL_MESH_VALIDATION.
- Made G29 P4 able to edit invalidated mesh points
- Restore a reasonable Fade Height for UBL when creating new state information
- Get UBL's Topology Map to look a little bit better
- Make sure the user doesn't see a blank screen when doing Mesh Editing.

* Huh???   GitHub Desktop screwed up!

* get the planner object in scope

* Fix out of scope z_fade_height

* Travis timed out...

I need a change so I can force a new commit and sync.

Reference: #6804

Plus some other UBL tweaks.

- fix broken `M421` due to less-than-careful optimization
- add HOME_AFTER_DEACTIVATE define to advanced config so not everyone has to rehome after steppers are deactivated
- misc. cleanups (remove unused label, unused variables)

 - Add crc16 utility function
 - Implement CRC16 for config store, remove old checksum, increment layout version
 - Move UBL mesh store/load to MarlinSettings; increment UBL_VERSION
 - Begin to lay out MAT structure, prototype functions, etc.
 - Rename ubl.state.eeprom_storage_slot to .storage_slot
 - Misc. optimization
 - Cleanup/standardize/improve some messages

This is a work in progress!

* Fix M48 for Delta's  There was a scope issue: DELTA_PROBEABLE_RADIUS wasn't getting defined.

* Check if the specified G29 command requires homing

Things should be pretty stable for a while.   
But it wouldn't surprise me if the Delta people have a need for a few special commands.

Since we already used 'T' for `G29` Topology Report in the past, and since 'T' is available…

Also clean up ubl_motion.cpp debug info and fix declaration of cx & cy

UBL on Delta's....     Should be close!    Should not affect any Cartesian printer.

- allow H code value to override probing height
- minor cleanups

- Split M421 into separate versions for bilinear and ubl
- Fix minor issue in G26

* M421 Mesh Point Offset and misc. UBL clean up

Allow M421 to accept an offset as well as absolute value for a specified
mesh point.
And misc. UBL clean up to reduce redundent code.

* Better error checking for M421

* Fix M421 Y index bug

I just noticed....  We've had a Y index bug for who knows how long?

    /**
     * UBL System submenu
     *
     *  Prepare
     * - Unified Bed Leveling
     *   - Activate UBL
     *   - Deactivate UBL
     *   - Mesh Storage
     *       Memory Slot:
     *       Load Bed Mesh
     *       Save Bed Mesh
     *   - Output Map
     *       Map Type:
     *       Output Bed Mesh Host / Output Bed Mesh CSV
     *   - UBL Tools
     *     - Build Mesh
     *         Build PLA Mesh
     *         Build ABS Mesh
     *       - Build Custom Mesh
     *           Hotend Temp:
     *           Bed Temp:
     *           Build Custom Mesh
     *         Info Screen
     *       - Build Cold Mesh
     *       - Fill-in Mesh
     *           Fill-in Mesh
     *           Smart Fill-in
     *           Manual Fill-in
     *           Info Screen
     *         Continue Bed Mesh
     *         Invalidate All
     *         Invalidate Closest
     *     - Validate Mesh
     *         PLA Mesh Validation
     *         ABS Mesh Validation
     *       - Custom Mesh Validation
     *           Hotend Temp:
     *           Bed Temp:
     *           Validate Mesh
     *         Info Screen
     *     - Edit Mesh
     *         Fine Tune All
     *         Fine Tune Closest
     *       - Adjust Mesh Height
     *           Height Amount:
     *           Adjust Mesh Height
     *         Info Screen
     *     - Mesh Leveling
     *         3-Point Mesh Leveling
     *       - Grid Mesh Leveling
     *           Side points:
     *           Level Mesh
     *         Info Screen
     *   - Output UBL Info
     */

UBL Clean up.

add BLTouch-related messages in english and (rusty) french;
add missing endstops.h in ultralcd.cpp;
fix misc. compiler warnings;
fix lsf_reset - ZERO macro can't handle a pointer as it would only memset the size of the pointer, not the size of the entire struct

Also...   The memory corruption issue may be fixed.   The GCC compiler
was inlining static functions and this caused the G29() stack frame to
become much larger than the AVR could handle.