the detection of bonds is fragile

[nbehrnd]

As an information: while gradually adding colors to the .pov the script generate
in a separate branch, I noticed (short) bonds for C-H, C-C, C-N, C-O, and C-F
seem to be less problematic for xyz2mol than the longer ones of C-S, or C-Cl.
The approach by openbabel seems to work differently as e.g., for the depiction
of furfural is reasonable there. Though it is not a road block for me, it
possibly is worth to share the this observation with you. The self assigned
issue contains some examples.

[nbehrnd]

An early version which relies on average covalent radii compiled in
Wikipedia (1/H up and including 96/Cm) entered a stage of "testing",
here.