interp: split loaded vs applied tool offset params (#5401-#5409, #5081-#5089)

[grandixximo]

Forward-port of #4219 (merged to 2.9) to master, fixing #4216.

#5401-#5409 again hold the loaded tool's stored offset (refreshed on M6 / G10 / init), restoring the behavior probing routines in the wild rely on. The offset actually applied to motion by G43/G43.1/G43.2 moves to the new parameters #5081-#5089, and is zeroed by G49.

Two commits:

• Cherry-pick of the runtime + docs fix from interp: split loaded vs applied tool offset params (#5401-#5409, #5081-#5089) #4219.
• Revert of the test-assertion commit from interp: make #5401-#5409 reflect the applied tool length offset (fixes #2994) #3996 (49ddfe6), which had rewritten the tool-offset assertions for the applied semantics. The restored assertions match the loaded-offset behavior.

Full test suite passes locally (291 tests, 0 failed).

[wucke13]

@grandixximo I think this did miss to update the absolute Z position formula example on the G38.n gcode?

[grandixximo]

I did not look at the absolute Z position formula, because I had a vague memory of having separated the formula out from #5403, but reading your message gave me a bit of a scare that I might have missed something, fortunately I did not miss, it was related before my changes in 6da626e
the new official formula is

G38.2 Z-100 F100
#<zmachine> = [#5063 + #5023 - #5422]

In both 2.9 and 2.10

The new one above is simpler than this following initial proposal we had

#<zmachine> = [#5063 + #[5203 + #5220 * 20] + #5213 * #5210 + #5403]

Which is not working well anymore, has to be changed to either

#<zmachine> = [#5063 + #[5203 + #5220 * 20] + #5213 * #5210 + #5083]

or the new one from the new docs

G38.2 Z-100 F100
#<zmachine> = [#5063 + #5023 - #5422]

So it should all be good, nothing missed, just gave me a heart attack lol

[wucke13]

Ah, sorry for the scare --- I didn't read careful enough. Two additional notes:

• At first I thought #5422 was a typo, but indeed the #542x parameter block starts with 0, not 1, so Z is 2, not 3.

• In the current docs, it reads:

  5420-5428 - Current relative position in the active coordinate system including all offsets and in the current program units for X, Y, Z, A, B, C, U, V & W, volatile.

  This is indeed the only parameter that is described as relative. I wonder what the word conveys there, I would assume that this is an absolute (but work-, as opposed to machine-) coordinate system? If indeed it is relative, then, relative to what?

[grandixximo]

I think it's in contrast to what is referred to as "absolute machine position" which is what you reference with G53. So it is relative to the active coordinate system if I interpret it correctly