* Correct over optimization of pre-initialized variables

NEAR_ZERO() works most of the time...  But it leaves corner cases
uncovered.  We need to explicitly check if these variables are +/-
infinity.

* Correct over optimization of pre-initialized variables

This is what I did yesterday:

- basicly gave the tests more comprehensive names; put all the
declarations at the top; got rid of the magic negative C-value (renamed
to P + A, O, T)

- "cos(RADIANS(180 + 30 * axis)) * (1 + circles * 0.1 * ((zig_zag) ? 1 :
-1)) * delta_calibration_radius" compiles wrong is zig_zag statement is
without brackets

- DELTA_TOWER_ANGLE_TRIM reset to 3 values (the calcs use the 3th value
to normalize will not compile otherwise)

-Wrote 3 dummies to keep EEPROM lenght the same

-Reset the configs to the 'original' with autocal + menu disabled (but
can be enabled of course)

Last three Spanish language strings

Updates as per @thinkyhead request.

As proffered in #6252

Update language_de.h

* Update language_cz.h for new translations

* Correction translate

* Revert correction back to hotend

Adding missing strings to Spanish language, as per changes recently
done to the RCBugFix branch.

updated the RCbugfix release with additional Dutch translations

The first character in the string wasn't blanked if the number went from >= 100 to <100, resulting in artefacts in the displayed string.

M666

+ tower radians absolute value

The Configuration.h file entries for BL-Touch have been updated to:
```cpp
//#define BLTOUCH
//#define BLTOUCH_DELAY 375      // (ms) Enable and increase if needed
//#define BLTOUCH_HEATERS_OFF // if defined the printer's heaters are
turned off during probe event
```

The electro-magnetic interference from the bed and nozzle are affecting
the BL-Touch repeatability for some users.   This problem can be helped
by shutting down the heaters during the actual probe event and then
quickly turning them back on.

Because this code is messing with the heaters, it is written in a
paranoid manner.  It only turns the heaters back on if everything is
EXACTLY as it expects things to be.  The BL-Touch probe must have been
put into a deployed state less than 20 seconds prior, or the stow()
function will NOT turn the heaters on.

This code has been tested and works for both G28 and probing functions.

Giving a negative number of probe points disables the tower angle
correction calibration ('4point' instead of '7point' solution)

EEPROM version updated

Also setting up Z-Probe on the machine.

Minor cleanups, work-in-progress

Change to support new name GRID_MAX_POINTS_XY
Start UBL at version 1.0 in the CUSTOM_MACHINE_NAME

Avoids to have to recalibrate the delta_height after a z_offset change

These are my two 'Nice' machines.  Both are good example machines.  The
FolgerTech i3-2020 is probably the best value kit on the market right
now.  And it has developed a 'Cult Following' of hobbiests.  I'll always
have an up to date Configuration.h file set for it, so it may as well be
an example configuration.

And the gCreate gMax 1.5+ is the best printer I've ever seen that runs
Marlin.  For sure, I'll always have up to the minute Configuration.h
files for this printer also.   I've moved over to using the gMax as my
primary development machine for Marlin.   So...  Besides being a good
machine, it is a nice convienience for me to have this Configuration.h
file as an 'Example'.

Fixes #6393

There are some quite tricky ones which are hardly translateable.
Maybe @Sebastianv650 can review this

Fixes #6426

Fixes #6445

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.

Patch fastio.h with Timer 2 interrupt support