0

A researcher at my university wants to run a conda module that was built with a later version of glibc (2.33) than I have (2.28). I'm working in a RedHat environment, and glibc is not upgradeable. I'd like to unpack the module and rebuild it in my environment so that the dependency is erased.

The module is "bsbolt" and the this in the directory created by 'git clone':

bsbolt  conda_recipe  docs  __init__.py  LICENSE  MANIFEST.in  mkdocs.yml  README.md  requirements.txt  setup.py  tests

The meta.yaml file is in conda_recipe. What is the command line I need to run to build a local module? From there, I can install it into our environment.

John Thompson
  • 1
  • 1
  • Can you install it via conda as opposed to installing it from source? – Kapocsi Jun 17 '22 at 19:38
  • The installation instructions might be of interest to you: https://bsbolt.readthedocs.io/en/latest/installation/ – Kapocsi Jun 17 '22 at 19:46
  • The whole point is to recompile it to use our glibc. The module, as is, installs fine, it just doesn't work due to unresolved references. – John Thompson Jun 17 '22 at 21:37
  • Have you looked at the `conda-build` documentation? Right now the question sounds like you haven't yet tried `conda build conda_recipe` – merv Jun 17 '22 at 22:20
  • I ran "conda build conda_recipe", and it stated getting dependencies, and then ended with: Collecting package metadata (repodata.json): ...working... done Solving environment: ...working... failed – John Thompson Jun 17 '22 at 23:14
  • Looked at that error and [stackoverflow](https://stackoverflow.com/questions/51266535/conda-returns-solving-environment-failed) told me to edit __init__.py and that was corrected. – John Thompson Jun 18 '22 at 00:22

0 Answers0