I got a "caller" workflow and a "called" workflow.
Caller workflow:
name: caller
on:
workflow_dispatch:
jobs:
call-another-workflow:
uses: project/called/.github/workflows/main.yml@main
Called workflow:
name: called
on:
workflow_dispatch:
workflow_call:
jobs:
run_python_script:
runs-on: [ self-hosted ]
steps:
# Following step will be executed in the caller context (should be executed in the called context)
- uses: actions/checkout@v3
- name: Setup python
uses: actions/setup-python@v4
with:
python-version: '3.9'
# Cannot find requirements.txt (wrong repository is checked out)
- name: Install python packages
run: |
python -m pip install --upgrade pip
pip install -r requirements.txt
- name: Execute the python script
run: python some_python_code.py
Error: <path> can't open file <path> [Errno 2] No such file or directory
-> The wrong repository is checked out (caller repo is checked out instead of the called which is needed). If I trigger the called workflow manually (the 'actions/checkout@v3' is checking out the right repo) the workflow runs successfully.
The caller workflow should 'just' trigger the called workflow in a way, that the workflow runs like it was triggered manually.
Unfortunately the github actions documentation says:
If you reuse a workflow from a different repository, any actions in the called workflow run as if they were part of the caller workflow. For example, if the called workflow uses actions/checkout, the action checks out the contents of the repository that hosts the caller workflow, not the called workflow.
(reference: https://docs.github.com/en/actions/using-workflows/reusing-workflows#overview)
-> Do we have the possibility to just trigger the workflow in a way that it is running like I triggered it manually? Alternatively, I can trigger the workflow from another workflow via web interface. But then it gets messy. Some ideas?
