WIP Ensure vector +z is nondimensional or uses unit q only

[PaulWessel]

See #7039 for discussion. The coding of modifier +z was expecting plot units but that makes no sense. This PR checks that either no unit or q is appended and errors otherwise.

Hi @joa-quim, if you have time, please note that vector2d_mods.sh and vector3d_mods.sh now fail (left column row 4) but I think those are not failures but confusion about what units are used etc. We are doing

Input: 2.20 1.27 Vector: -Sv20p+e+z1q

So with q that means the input is user units to be converted via -R -J to plot units, then scaled by 1. So given the previous plot row has

Input: 2.20 1.27 Vector: -Sv20p+e+z0.394q+c --PROJ_LENGTH_UNIT=inch

it should be longer by 2.54. The --PROJ_LENGTH_UNIT should not have any implication here since input is not in plot units but in user units. Seems to me the original PS is wrong due to the bug we are trying to fix. Let me know what you think.

[joa-quim]

I have implemented patches Julia plot to workaround this bug and I think what I get is the right arrow. So, yes those tests have the wrong plots.

arrows([0 0 2.2 1.27], show=1, uv=true, Vd=1)
psxy -R-0.1/2.3/-0.04/1.32 -JX14c/9.5c -Baf -BWSen -Sv8p+e+h0.5+z5.833333333333333iq -P -K > C:\TEMP\GMTjl_tmp.ps

[PaulWessel]

OK, then I need a review t approve and I can update the PS later.

[joa-quim]

But the 2.54 factor is still there, as well as the problem when the axes are not isometric. See, I need to multiply by 2.54.

 echo 0 0 2.2 1.27 | gmt plot -R-0.1/2.3/-0.04/1.32 -JX14c/9.5c -Baf -BWSen -N -Sv8p+e+h0.5+z2.54q -png lixo

Regarding the non-isometry, I found that multiplying the yy* by a scale factor given by

(map_height / map_width) / ((y_max - y_min) / (x_max - x_min))

(9.5/14) / ((1.32+0.04)/(2.3+0.1)) * 1.27 = 1.5207983193277308

and a zscale given by map_width / (x_max - x_min)

14 / (2.3+0.1) / 2.54 = 2.2966
solves it.

echo 0 0 2.2 1.52 | gmt plot -R-0.1/2.3/-0.04/1.32 -JX14c/9.5c -Baf -BWSen -N -Sv8p+e+h0.5+z2.2966q -png lixo

[PaulWessel]

Added WIP back on. Maybe have time to look at this tomorrow, but times are busy.