fix issue 1280

[chselz]

Should fix issue #1280

I compiled it using meson and ifort compiler. For me it worked then using water as an example

3

O 0.000000 0.000000  0.000000
H 0.758602 0.000000  0.504284
H 0.260455 0.000000 -0.872893

[chselz]

@foxtran do you have an idea why CI fails? I don't see how my changes could have affected the cpcm-x test.

[foxtran]

See #1325

[foxtran]

In fact, it fails not randomly. As far as I remember, that is happening because cpx uses files in API and they can be overwritten by concurrent jobs.

[foxtran]

You can try to run xtb/cpx + one of xtb/tblite or xtb/gfn2 or xtb/gfnff test and get this error manually.

[lmseidler]

in principal this is not the place to handle this but everything else requires a better design aka larger refactor so this is ok

[lmseidler]

In set_logicals you could check using getValue whether the --nfinal flag was set, as you do for setting nopt in the first place in set_docking (where you now set the new logical). This avoids introducing the new logical.

[Albkat]

LGTM