Segfault in an extreme case about matscipy HOT 9 CLOSED

Hi Christoph,

This looks like a bug and should be fixed.

Can you Post an atomic configuration (e.g. extxyz file) that reproduces this. I will then fix it.

Best,
Lars

from matscipy.

sorry for the delay - I will post something. (eventually)

from matscipy.

here is a Python code that reproduces the crash. It is kind of a ridiculous example...

import ase.build
at = ase.build.bulk("Al", cubic=True)
at.cell = at.cell * 8000
import matscipy.neighbours
matscipy.neighbours.neighbour_list("ijdDS", at, 4.5);

from matscipy.

I just pushed a fix for this. It was related to an integer overflow in the total number of cells used for the neighbor list search. Your example now runs, but it having lots of vacuum will be super-slow at some point.

from matscipy.

Maybe print a warning or even a useful error message?

from matscipy.

Will add a warning. Did you check whether it fixes the problem that occured in your Julia code?

from matscipy.

Code is now issuing a warning in that situation.

from matscipy.

yep - all working as expected.

from matscipy.

thank you!

from matscipy.