Further cleanup of UBL

[thinkyhead]

Rebase #6501 and apply more cleanup…

• Fix a bug in smart_fill_mesh
• Reduce code size of smart_fill_mesh
• Use 'PSTR' for ::debug methods' title argument
• Fix formatting, apply coding standards

Concise diff

[Bob-the-Kuhn]

In UBL_G29.cpp, please replaces all the X_MIN_POS ... travel limits with the UBL_MESH_MIN_X ... versions.

[thinkyhead]

How about in G26?

[bgort]

Need an ampersand here, I believe.

float x_tmp = pgm_read_float(ubl.mesh_index_to_xpos[i]),
should be
float x_tmp = pgm_read_float(&ubl.mesh_index_to_xpos[i]),

Same with the next line.

[thinkyhead]

!!!

[Roxy-3D]

Agreed!

[Roxy-3D]

How about in G26?

Yes... for now... that is probably the right thing to do. But I suspect G26 has value to the Bi-Linear and original Mesh bed leveling. In which case, using the UBL constants is not the 'right' thing to do...

[thinkyhead]

If it's "right" for UBL then it's ok. It should be self-consistent.

[Roxy-3D]

Wait, is this correct?
The first three are squared, but the three others are not!

Good catch! This is wrong.... These are xy, xz and yz terms in the polynomial. They are not x^2, y^2 or z^2 terms. The code will NOT work like this.

-  lsf->x2bar = lsf->x2bar / N - lsf->xbar * lsf->xbar;
 -  lsf->y2bar = lsf->y2bar / N - lsf->ybar * lsf->ybar;
 -  lsf->z2bar = lsf->z2bar / N - lsf->zbar * lsf->zbar;
 -  lsf->xybar = lsf->xybar / N - lsf->xbar * lsf->ybar;
 -  lsf->yzbar = lsf->yzbar / N - lsf->ybar * lsf->zbar;
 -  lsf->xzbar = lsf->xzbar / N - lsf->xbar * lsf->zbar;
 +  lsf->x2bar = lsf->x2bar / N - sq(lsf->xbar);
 +  lsf->y2bar = lsf->y2bar / N - sq(lsf->ybar);
 +  lsf->z2bar = lsf->z2bar / N - sq(lsf->zbar);
 +  lsf->xybar = lsf->xybar / N - sq(lsf->xbar);
 +  lsf->yzbar = lsf->yzbar / N - sq(lsf->ybar);
 +  lsf->xzbar = lsf->xzbar / N - sq(lsf->xbar);