Open hexylena opened 3 years ago
Hi,
Thank you for your pull request! Yeah, this slipped through. Due to SARS2 our focus changed and we do not have the time to pick this up properly, which is also why I didn't reply promptly.
I'll leave this pull request open for now since the method you suggest causes problems due to nested environments on our work-environment. A similar solution as for the Jovian_master
environment is required where first an empty env is created using conda
which contains just the conda
and mamba
packages, then activate the conda env and update it via mamba
as shown here: https://github.com/DennisSchmitz/Jovian/blob/77422ff0083309b24f6b2633112f1d03eb46c02a/bin/includes/Install_jovian-master#L27.
Again, thank you very much for bringing this to our attention! Although the changes are minimal the validation on our production setup will take some time, it will be picked up and included in the next (minor) version release so it can be validated in one go.
Kind regards, Dennis
No worries @DennisSchmitz, completely understand the validation issue. No rush from us to merge, we'll just keep using this patch locally until you're all ready!
Hey y'all,
I was helping our two Avans students (Bob van Rijthoven, Diemo Perre) get Jovian setup and they'd been stuck on conda resolution. On a previous installation it'd gone very well, but, we needed to reconfigure and reinstall and it stuck for days at a time on creating the Jovian_helper_env. I didn't see anything obvious slowing down the resolution, but on running a manual
conda env create -f ...
with-vvv
it was just stuck in the SAT solver.Creating this environment via mamba was instantaneous in comparison. It seems you're using mamba already in some of the per tool environments, I think it may make sense for the helper environment (and the main env possibly?) to be created via mamba as well.