Although the SEGV when coordinates are NaN has already been
bypassed, I refactored this function when trying to understand
what it did. I find this version more readable, it also happens
to be more efficient (although that probably doesn't matter
with 3-5 substituents).
Although the SEGV when coordinates are NaN has already been bypassed, I refactored this function when trying to understand what it did. I find this version more readable, it also happens to be more efficient (although that probably doesn't matter with 3-5 substituents).
See @greglandrum 's https://github.com/schrodinger/coordgenlibs/pull/107