Trigger GitHub actions if push in another repo - github

In one of my repos (repo A), there is a script that gets the content of a file on another repo (repo B, not one of mine). Therefore, I would like to trigger a GitHub action on repo A each time there's a push on repo B. I didn't see this case on GitHub actions trigger documentation, is this possible?

You can execute each day a script with cron in order to start or not your workflow. A script can read easily the date of last commit of another repository.
Detect when the last commit is less than a day old
With this script, the exit code is 0, only if the last commit is less than a day old.
#!/bin/bash
exit $(curl -s https://api.github.com/repos/wikimedia/mediawiki/commits/REL1_36 | jq -r "((now - (.commit.author.date | fromdateiso8601) ) / (60*60*24) | trunc)")
Execute the script with cron each day
name: trigger
on:
schedule:
- cron: '30 5 * * *'
workflow_dispatch:
jobs:
check:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout#v2
- run: ./isUpdated.sh
Start your workflow when a new commit exists since yesterday
name: Your workflow
on:
workflow_run:
workflows: ["trigger"]
types: [completed]
workflow_dispatch:
jobs:
check:
runs-on: ubuntu-latest
if: ${{ github.event.workflow_run.conclusion == 'success' }}
steps:
- name: Checkout
uses: actions/checkout#v2
- run: echo ok

edit: this won't work since OP doesn't have push access to repo B
I think that you might be able to get what you want by using the repository_dispatch event. This essentially triggers the workflow when an endpoint is pinged with the correct payload.
From the documentation:
To trigger the custom repository_dispatch webhook event, you must send a POST request to a GitHub API endpoint and provide an event_type name to describe the activity type. To trigger a workflow run, you must also configure your workflow to use the repository_dispatch event.
My best guess is that your desired solution can be achieved by creating 2 workflows: 1 in Repo A that runs on: repository_dispatch, and 1 in Repo B that runs on: push and contains a step that hits Repo A's API (e.g. using curl). This would give you the following events:
You push to Repo B.
Repo B's on: push workflow is triggered, and the step that hits Repo A's API endpoint (e.g. using curl) is run.
Repo A's API endpoint receives the payload and the on: repository_dispatch workflow is triggered.

Related

github actions | get origin branch on push

I want that a workflow that starts on push will print the origin (base) branch name and name of the one who did the push.
for example if my name is Tom and I merged 'test' branch to master, I will be able to get 'Tom' and 'test' as values on the workflow that starts on push (right after the pull request).
I did find something that works on pull request to get the base branch name: ${{ github.event.pull_request.head.ref }}, but when replacing 'pull_request' with push it doesn't work anymore.
Is there a way to do it?
I've been searching for hours now and can't find a proper solution, for each type of event there is a different solution.
Thank you for reading!
You can use the following properties of the github context:
github.ref_name
github.event.pusher.name (see: push event payload)
Here is an example of a complete workflow:
name: show-push-info
on:
push:
jobs:
print-info:
runs-on: ubuntu-latest
steps:
- name: Print info
run: echo "'${{ github.ref_name }}' was pushed by '${{ github.event.pusher.name }}'"

Create Pull Request after deleting files using GitHub Actions

I'm making an action that runs a script that deletes some files from a github repository and I want to create a pull request with those changes.
So far the structure is like this:
#myactionscript.yml
name: test
on: push
jobs:
delete:
runs-on: ubuntu-latest
steps:
#setup steps
...
- name: delete files
run: |
#script that deletes files
- name: Create Pull Request
uses: peter-evans/create-pull-request#v4.0.4
I published the action so I could test it on a private repo. If I print the folder contents, I can see that the file deletion script is working, but the pull request is not being created. However, it does work when I use it externally on my private repo, like this:
#privaterepoworkflow.yml
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout#v2
- name: File deletion
uses: myuser/deletion-action#test5
- name: Create Pull Request
uses: peter-evans/create-pull-request#v4.0.4
This creates the Pull Request with the deleted files, but it requires me to add the code snippet to my private repo, so for every repository that I wanna run this action on, I will also have to paste the Create Pull Request snippet. Is there any way to make it so the Pull Request is created while adding this snippet only in the delete action itself?

Merge by github action does not trigger CI that listen by branch

I had two github action, one of merge-staging-in-to-master, the other one is CI.
The merge-staging-in-to-master was triggered by manually.
And CI was triggered by push to master.
But, when I triggered the merge-staging-in-to-master manually, the CI does nothing...
The expectation I want is invoke the CI action after merge-staging-in-to-master was done.
There is my code below.
merge-staging-in-to-master.yml
# This is a basic workflow to help you get started with Actions
name: merge-staging-to-master
# Controls when the workflow will run
on:
# Allows you to run this workflow manually from the Actions tab
workflow_dispatch:
# Allows external webhook trigger
repository_dispatch:
types:
- merge-staging-to-master
# A workflow run is made up of one or more jobs that can run sequentially or in parallel
jobs:
# This workflow contains a single job called "build"
merge-staging-to-master:
# The type of runner that the job will run on
runs-on: ubuntu-latest
# Steps represent a sequence of tasks that will be executed as part of the job
steps:
# Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
- uses: actions/checkout#v2
# Runs a set of commands using the runners shell
- name: Run a multi-line script
run: |
git fetch origin --unshallow
git checkout -t remotes/origin/master
git pull origin master
git config --global user.name "user"
git config --global user.email "user#mail.com"
git merge staging -m "[AUTO] merge staging back to master"
git push
CI.yml
# This is a basic workflow to help you get started with Actions
name: CI
# Controls when the workflow will run
on:
# Triggers the workflow on push or pull request events but only for the main branch
push:
branches: [ master ]
# Allows you to run this workflow manually from the Actions tab
workflow_dispatch:
# A workflow run is made up of one or more jobs that can run sequentially or in parallel
jobs:
# This workflow contains a single job called "build"
build:
# The type of runner that the job will run on
runs-on: ubuntu-18.04
# Steps represent a sequence of tasks that will be executed as part of the job
steps:
# Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
- uses: actions/checkout#v2
- uses: actions/setup-node#v2
with:
node-version: '14'
- id: 'auth'
name: 'Authenticate to Google Cloud'
uses: 'google-github-actions/auth#v0'
with:
credentials_json: '${{ secrets.GOOGLE_CREDENTIALS }}'
- name: Install Dependencies
run: |
sudo apt update
sudo apt install -y libcairo2-dev libjpeg-dev libpango1.0-dev libgif-dev librsvg2-dev
- name: 'Set up Cloud SDK'
uses: 'google-github-actions/setup-gcloud#v0'
- name: Build
run: sudo npm install --build-from-source && npx prisma generate dev && npm run build
- name: Deploy
run: npm run deploy
This is a builtin feature of GitHub actions to prevent jobs from triggering more jobs and potentially burning through your action minutes in a very short time (all the while blocking hosted runners for others as well).
The GitHub actions tokens used to authenticate have a special flag on them by which GitHub knows the token is from an Actions workflow.
If you know what you're doing you can use a PAT or Oauth App token to authenticate your git push instead, that will trigger further workflows. If you search the GitHub marketplace you'll find a few actions that can retrieve an OAuth App token.

Automating a PR merge followed by opening a new PR

I have a branch called stage and currently have a workflow that runs when a PR is labeled (to merge into the stage branch), and I'm currently trying to create a workflow that happens if the branch is successfully automerge.
For example -- here's a working example of the workflow to the stage branch:
name: PR opened to stage
on:
pull_request:
types: [labeled]
branches: [ stage ]
jobs:
audit-checks:
runs-on: self-hosted
steps:
- uses: actions/checkout#v2
if: contains(github.event.pull_request.labels.*.name, 'ready')
- name: Enable Auto Merge
uses: peter-evans/enable-pull-request-automerge#v1
if: contains(github.event.pull_request.labels.*.name, 'ready')
with:
token: ${{ secrets.GITHUB_TOKEN }}
pull-request-number: ${{ github.event.pull_request.number }}
merge-method: merge
And here's another workflow that, when a PR to stage is closed, is supposed to run. I have a step as well that checks to see if the label includes ready and is supposed to create another PR to another branch:
name: ready for prod test
on:
pull_request:
types: [closed]
branches: [ stage ]
jobs:
audit-checks:
runs-on: self-hosted
steps:
- uses: actions/checkout#v2
- name: automate PR creation for prod
if: |
${{ github.event.pull_request.merged }} && contains(github.event.pull_request.labels.*.name, 'ready')
uses: peter-evans/create-pull-request#v3
with:
token: ${{ secrets.GITHUB_TOKEN }}
branch: stage
base: prod
labels: ready
title: Automated PR opened
It seems that when the PR is initially opened and labeled ready, the first workflow runs just fine. However, once that workflow enables auto merge and it actually auto-merges (upon completion of the workflow), the second workflow doesn't ever get triggered.
Is it possible that the second workflow is somehow running before the first one completes and doesn't actually do anything because of it? I'm not sure it's anything in my 2nd job (at least not yet) since it hasn't even run (or run and fail).
Just found my answer here: https://docs.github.com/en/actions/security-guides/automatic-token-authentication
When you use the repository's GITHUB_TOKEN to perform tasks on behalf
of the GitHub Actions app, events triggered by the GITHUB_TOKEN will
not create a new workflow run. This prevents you from accidentally
creating recursive workflow runs. For example, if a workflow run
pushes code using the repository's GITHUB_TOKEN, a new workflow will
not run even when the repository contains a workflow configured to run
when push events occur.

Only run job on specific branch with GitHub Actions

I'm relatively new to GitHub Actions and I have 2 jobs–one that runs my tests, and one that deploys my project onto a server.
Obviously I want the tests to run on every branch, but deploying should only happen when something gets pushed to master.
I'm struggling to find a way to run a job on a specific branch. I know it's possible to only run entire workflows on a specific branch, however that would mean I would have a "test" workflow and a "deploy" workflow.
This sounds like a solution, however they would run parallel. In an ideal world, the tests would run first, and only if they succeed, then the deploy job would start. This isn't the case when using 2 separate workflows.
How would I be able to achieve this? Is it possible to run jobs on a specific branch?
In a recent update you can now put if conditionals at job level. See the documentation here. https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#jobsjob_idif
I tested this workflow which runs the job test on every push, but only runs deploy on the master branch.
name: my workflow
on: push
jobs:
test:
runs-on: ubuntu-latest
steps:
- name: Execute tests
run: exit 0
deploy:
runs-on: ubuntu-latest
needs: test
if: github.ref == 'refs/heads/master'
steps:
- name: Deploy app
run: exit 0
What follows is my original answer, and an alternative solution if you prefer to have separate workflows.
The first workflow runs for every branch except master. In this workflow you run tests only.
on:
push:
branches:
- '*'
- '!master'
The second workflow runs for just master and runs both your tests and deploys if the tests were successfully passed.
on:
push:
branches:
- master
Most answers provide a solution for one single branch. To restrict the job to run on any specific set of branches, you can do it using the if conditional with multiple disjunction (||) operators; but this is too verbose and doesn't respect the DRY principle.
The same can be archived with less repetition using the contains function.
Using contains:
contains('
refs/heads/dev
refs/heads/staging
refs/heads/production
', github.ref)
compared to using multiple ||:
github.ref == 'refs/heads/dev' || github.ref == 'refs/heads/staging' || github.ref == 'refs/heads/production' || …
Full example:
---
on: push
jobs:
test:
name: Test
runs-on: ubuntu-latest
steps:
- uses: actions/checkout#v2
- name: Run tests
run: …
deployment:
name: Deployment
runs-on: ubuntu-latest
needs: [test]
if:
contains('
refs/heads/dev
refs/heads/staging
refs/heads/production
', github.ref)
steps:
- uses: actions/checkout#v2
- name: Deploy
run: …
Here is what I've done for steps that should only run on a specific branch.
- name: Publish Image
# Develop branch only
if: github.ref == 'refs/heads/develop'
run: |
... publish commands ...
2021 update
I know it's possible to only run entire workflows on a specific
branch, however that would mean I would have a "test" workflow and a
"deploy" workflow.
This sounds like a solution, however they would run parallel. In an
ideal world, the tests would run first, and only if they succeed, then
the deploy job would start. This isn't the case when using 2 separate
workflows.
You can now use the event workflow_run to achieve the part the tests would run first, and only if they succeed, then the deploy job would start (read on to see how):
Documentation page of workflow_run
https://docs.github.com/en/actions/learn-github-actions/events-that-trigger-workflows#workflow_run
This event occurs when a workflow run is requested or completed, and
allows you to execute a workflow based on the finished result of
another workflow. A workflow run is triggered regardless of the result
of the previous workflow.
For example, if your pull_request workflow generates build artifacts,
you can create a new workflow that uses workflow_run to analyze the
results and add a comment to the original pull request.
Now, considering OP's initial issue:
I want the tests to run on every branch, but deploying should only happen when something gets pushed to master
This now could be solved like this:
The following setup is working, a few minutes ago I just implemented the same logic in one of my repository
Workflow <your_repo>/.github/workflows/tests.yml
name: My tests workflow
on:
push:
branches:
- master
pull_request: {}
jobs:
test:
# ... your implementation to run your tests
Workflow <your_repo>/.github/workflows/deploy.yml
name: My deploy workflow
on:
workflow_run:
workflows: My tests workflow # Reuse the name of your tests workflow
branches: master
types: completed
jobs:
deploy:
# `if` required because a workflow run is triggered regardless of
# the result of the previous workflow (see the documentation page)
if: ${{ github.event.workflow_run.conclusion == 'success' }}
# ... your implementation to deploy your project
Although this discussion is very old but recently I came across the same problem with a slight addition. if condition to check whether the branch is main works but what if someone pushes their branch and updates the workflow yml file to remove the if condition? The deploy job will get triggered without their branch even being reviewed or merged in main and could potentially break the production environment. This could be a concern in open source projects.
I couldn't find an answer for this anywhere so wanted to share my findings. I hope this is the right thread for it.
To ensure that there's no way a job is triggered except for in specific branches, environments can be used. A deploy job will most likely have some api keys to connect to the destination server that might be stored in secrets. Instead of storing them in repository secrets which are accessible globally within the repository, we should store them in respective environments.
The official docs for environments contain detailed explanation with sample scripts but sharing a simple example here. Let's say we want to run production deployment only when main is updated
From repository Settings, create a production environment
Choose Selected Branches in Deployment Branches dropdown and add main in the pattern
Add the api keys in production environment secrets
In the workflow yml we just need to add environment information environment: production as follows (using script from #peterevans's answer)
name: my workflow
on: push
jobs:
test:
runs-on: ubuntu-latest
steps:
- name: Execute tests
run: exit 0
deploy:
runs-on: ubuntu-latest
needs: test
if: github.ref == 'refs/heads/main'
environment: production
steps:
- name: Deploy app
run: exit 0
The environment information indicates where the secrets have to be read from. If the current branch name doesn't match the pattern provided in Selected Branches then the job will immediately fail with an error. Since we have a condition to only run this on main, normally that won't bother us because this job will be skipped on other branches anyways. But if someone, mistakenly or with ill intentions, modifies the yml file and removes the condition before pushing their branch, they'll get an error. So, our system remains secure at least from the threat here.
Hope this helps anyone wondering the same.
The following worked for me inside a job:
Where a PR target branch is one of the following:
if: contains(github.base_ref, 'staging')
|| contains(github.base_ref, 'production')
When the source branch of the pull request is one of them below:
if: contains(github.head_ref, 'feature')
|| contains(github.head_ref, 'release')
name: CI
on: push
jobs:
prod-check:
if: ${{ github.ref == 'refs/heads/main' }}
runs-on: ubuntu-latest
steps:
- run: echo "Deploying to production server on branch $GITHUB_REF"
According to the documentation, the if condition is wrapped like so
if: ${{ github.ref == 'refs/heads/main' }}
While you can't have conditions at job level at this moment, you can have conditions at step level - see Contexts and expression syntax for GitHub Actions.
For getting a branch name, current solution is to check GITHUB_REF environment variable - see Default environment variables and this question for details.
Putting it all together - if you decide to go with accepted answer in last link, your workflow might look like this:
jobs:
test:
runs-on: ubuntu-latest
steps:
- name: Run tests
run: ./my-tests.sh
deploy:
runs-on: ubuntu-latest
needs: test
steps:
- name: Extract branch name
shell: bash
run: echo "##[set-output name=branch;]$(echo ${GITHUB_REF##*/})"
id: extract_branch
- name: Deploy
run: ./deploy.sh
if: steps.extract_branch.outputs.branch == 'master'
If you'd rather keep everything in workflow file instead of separated script, you can always add if to each step in given job.
I hope it's just a temporary solution/workaround, and job conditions are going to be added before end of beta.
For steps or jobs you can also use github.ref_name which is the branch or tag name that triggered the workflow run.
name: my workflow
on: push
jobs:
if: github.ref_name == 'main'
test:
runs-on: ubuntu-latest
steps:
- name: Execute tests
run: exit 0
For more information about the github context check here.
You need to add these two script to the workflows
//main.yml
name: main
on:
push:
branches:
- main
//size.yml
name: size
on:
pull_request:
branches:
- main
There are already some pretty good answers. Here's a way to specify multiple branches in your conditional:
name: my workflow
on: push
jobs:
deploy:
if: contains(fromJson('["refs/heads/master", "refs/heads/main"]'), github.ref)
runs-on: ubuntu-latest
steps:
- name: Execute tests
run: exit 0
2022
Below workflow runs only for push in main branch:
name: Distribute to Firebase
on:
push:
branches: [ main ]
jobs:
...