Deploy multi-platform images #1412
Conversation
| @@ -0,0 +1 @@ | |||
| #!/bin/bash | |||
There was a problem hiding this comment.
TODO: write the script to actually create the multi-platform docker manifest.
Revert later
| - name: Wait for Travis CI | ||
| uses: fountainhead/[email protected] | ||
| id: wait-for-travis | ||
| with: | ||
| token: ${{ secrets.GITHUB_TOKEN }} | ||
| checkName: Travis CI - Pull Request | ||
| ref: ${{ github.event.pull_request.head.sha || github.sha }} | ||
| intervalSeconds: 10 | ||
| timeoutSeconds: 3600 |
There was a problem hiding this comment.
Unfortunately, I expect the Travis CI check to fail every now and then (appart from the timeout being way too short for a Travis CI build).
This should really be a different workflow, possibly triggered by the checks being updated on main ?
There was a problem hiding this comment.
A different workflow should be fine, but it's harder to test since it can only run after merging, see https://docs.github.com/en/actions/using-workflows/events-that-trigger-workflows#workflow_run
You could extract it to a separate workflow afterwards.
There was a problem hiding this comment.
It's fine to leave this here at the moment as you're saying.
It shouldn't be to hard to split with the deployment script being a separate bash script.
| - name: Deploy | ||
| if: steps.wait-for-travis.outputs.conclusion == 'success' | ||
| run: ./deploy_multi_platform.sh | ||
|
|
||
| - name: Error on failure | ||
| if: steps.wait-for-travis.outputs.conclusion != 'success' | ||
| run: exit 1 |
There was a problem hiding this comment.
maybe if this was a single step, the workflow could be re-run with failing steps only. This would require a manual action though hence I still prefer to have a separate workflow triggered on check updates if possible (only need to worry about Travis CI mostly in that case).
#1306