base-branch:
required: false
description: 'The name of the branch being compared to. Uses $GITHUB_BASE_REF if not set'
default: ${{ github.base_ref }}
command:
required: false
description: 'The command to run to get the file changes can contain {glob} and {branchName} to specify replacements'
default: 'git diff --name-status --no-renames {branchName} {glob}'
glob-template:
required: false
description: 'How to format the globs received'
default: "'{glob}'"
filter-patterns:
required: false
description: >-
A map with keys ADDED, CHANGED, and DELETED as keys and a pattern object ({"pattern": "<pattern>"}) as a value.
Any defined keys will be kept, others discarded.
The matching is done by checking if the file change starts with the pattern.
default: |
ADDED: {"pattern": "A\t"}
CHANGED: {"pattern": "M\t"}
DELETED: {"pattern": "D\t"}
change-map:
required: true
description: >-
A multi-line map of changes to find.
eg. python_files: {"globs": "*.py", "separateDeleted": false}
requirements: {"globs": "requirements/*.txt"}
py_and_requirements: {"globs": ["requirements/*.txt", "*.py"]}
The final boolean determines if we separate out floats
eg. For a Python project where you want a list of Python files changed with and without deleted files
uses: AceFire6/file_changes@v0.1.0
with:
change-map: |
python: {"globs": "*.py", "separateDeleted": true}
requirements: {"globs": "requirements/*.txt"}
migrations: {"globs": "**/migrations/*.py"}
any-matches:
description: 'Any glob matches found'
# one set of outputs for each entry in change-map
<change-map.0.0>: '<file_changed1> <file_changed2> ...'
# Value set to make boolean checks simpler - 'true' or 'false'
# 'false' if there were no changes found
any-<change-map.0.0>: 'true'
# If separate deletes key is true
deleted-<change-map.0.0>: '<file_deleted1> <file_deleted2> ...'
eg. Referring back to the inputs example of a Python project. The is what the outputs would look like assuming the following:
- Added files:
tests/test.py
- Changed files:
main.py
helpers/utils.py
requirements/api.txt
- Deleted files:
utils.py
requirements/aip.txt
any-matches: 'true'
python: 'tests/test.py main.py helpers/utils.py'
any-python: 'true'
deleted-python: 'utils.py'
requirements: 'requirements/api.txt requirements/aip.txt'
any-requirements: 'true'
migrations: ''
any-migrations: 'false'
First, you'll need to have a reasonably modern version of
node
handy. This won't work with versions older than 9, for instance.
$ npm install
$ npm run build && npm run package
$ npm test
PASS ./index.test.js
✓ throws invalid number (3ms)
✓ wait 500 ms (504ms)
✓ test runs (95ms)
...
The action.yml contains defines the inputs and output for your action.
Update the action.yml with your name, description, inputs and outputs for your action.
See the documentation
Most toolkit and CI/CD operations involve async operations so the action is run in an async function.
import * as core from '@actions/core';
...
async function run() {
try {
...
}
catch (error) {
core.setFailed(error.message);
}
}
run()
See the toolkit documentation for the various packages.
Actions are run from GitHub repos so we will checkin the packed dist folder.
Then run ncc and push the results:
$ npm run package
$ git add dist
$ git commit -a -m "prod dependencies"
$ git push origin releases/v1
Note: We recommend using the --license
option for ncc, which will create a license file for all of the production node modules used in your project.
Your action is now published! 🚀
See the versioning documentation
After testing you can create a v1 tag to reference the stable and latest V1 action