It seems nova is also using the crypt module but is missing from the list https://github.com/openstack/nova/blob/master/nova/virt/disk/api.py#L639-L64... AS pointed by Artem, the serious problem with crypto module is that we don't have a good replacement now. passlib is abandoned and is known to be incompatible with latest bcrypt (which we had to pin in u-c). Probably we need to consider pulling https://pypi.org/project/crypt-r/ or importing the implementation to oslo.utils . On 10/9/24 01:13, Iury Gregory wrote:
Thanks for raising this!
I've created a launchpad bug for Ironic and IPA to track the work [1]
[1] https://bugs.launchpad.net/ironic/+bug/2083955 <https://bugs.launchpad.net/ironic/+bug/2083955>
Em ter., 8 de out. de 2024 às 12:27, Artem Goncharov <artem.goncharov@gmail.com <mailto:artem.goncharov@gmail.com>> escreveu:
> > While pipes and crypt have direct replacement (subprocess and for > example the pypi module bcrypt), there's a few comments here: > > https://docs.python.org/3.13/whatsnew/3.13.html <https://docs.python.org/3.13/whatsnew/3.13.html>
BEWARE: crypt is on the radar of Keystone team and one of the biggest issues with it is that one of the replacement candidates is passlib which is unmaintained and that itself depends on crypt. bcrypt is not a replacement for crypt, but rather an alternative hashing algorithm. We went way of getting rid of passlib in Keystone and got prepared to drop crypt support (deprecated at Pike) as such once we start testing for 3.13.
> > > In the mean time, let's tackle the above. > > Cheers, > > Thomas Goirand (zigo)
Thanks for this overview
Artem
-- /Att[]'s/ /Iury Gregory Melo Ferreira //MSc in Computer Science at UFCG / /Ironic Core/ //Senior Software Engineer at Red Hat Brazil// /Social/:https://www.linkedin.com/in/iurygregory <https://www.linkedin.com/in/iurygregory> /E-mail: iurygregory@gmail.com <mailto:iurygregory@gmail.com>/