`working_directory` has been set but not used when resolving cache

Feedback & Bug Reports
1

I’m setting up CircleCI for a new repository and have the following config:

version: 2.1
executors:
  dev:
    docker:
      - image: cimg/python:3.9
    working_directory: ~/new_project
orbs:
  python: circleci/python@1.0.0
  aws-cli: circleci/aws-cli@1.4.1
  kubernetes: circleci/kubernetes@0.11.2
jobs:
  test:
    executor: dev
    steps:
      - checkout
      - python/install-packages:
          pkg-manager: poetry
      - run:
          name: run tests
          command: echo "hello"
workflows:
  version: 2
  test:
    jobs:
      - test

However, i’m seeing this error after the checkout step:

error computing cache key: template: cacheKey:1:26: executing "cacheKey" at <checksum "~/project/poetry.lock">: error calling checksum: open /home/circleci/project/poetry.lock: no such file or directory

It seems that when running the next step for installing packages, CircleCI is looking for a cached poetry.lock file at ~/project. Why isn’t it looking at the working directory specified in the executor (~/new_project)? I thought if a job referenced the config for an executor that specified the working_directory key, the job would also use that working directory.

Am I missing anything here? I appreciate any insight.

2

It turns out the orb circleci/python@1.0.0 was the issue. i bumped the version and the issue of looking in the wrong working directory has been resolved

3

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.