-
Notifications
You must be signed in to change notification settings - Fork 16
[DRAFT/TEST] try out CI config for testing packages #369
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Also just went through the |
Is this the same recursion issue? #333 |
I think so, didn't know that issue existed. Thanks for linking. |
@chinandrew This looks promising 👍
Is this referring to being able to run CI actions based on what is being "worked on" (ie changes to a single indicator)? That would be pretty useful, and was something I was thinking about trying to figure out for Jenkins. |
Yea, so something like changing the code for |
Alright linting is in place now (which is why checks aren't passing). See #333 for more details. |
opening a new pr in #449 |
Just testing out some github actions, don't merge. cc @korlaxxalrok @benjaminysmith @krivard.
Aside from that issue, something like this should satisfy the testing need for PRs. See checks below for summaries. I'm only running 2 indicators and the utils package at the moment for demonstration.
I've brought up the subject of costs for this before. Even though it's not super high it's still worth keeping in mind. If it is an issue, one thing we could do if we move to a gitflow branching approach is only run tests on merge with the develop branch. This way we can catch things before they hit main without needing to run 10 checks each time someone commits changes to a PR. It unfortunately gives up the ability to see all the checks passing in PR phase. There may be some intermediate approach where only checks on altered packages are run, but that's going to be a bit more complex and not foolproof (e.g. if i change the utils i can still break other packages).
I'm still trying to figure out the linter. On some packages (usafacts), I hit this weird recursion error that I can't reproduce locally and have been unable to debug.