-
-
Notifications
You must be signed in to change notification settings - Fork 280
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
Update ci skip documentation in infrastructure.md #2297
base: main
Are you sure you want to change the base?
Conversation
✅ Deploy Preview for conda-forge-previews ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
There have been several iterations in Azure (see microsoft/azure-pipelines-agent#1270), so we had to combine different "skipper" formulas to make sure we were not triggering CI services in large migrations (e.g. admin-migrations). So it's one of those things that might not hurt to keep around "just in case". Edit: Yes, some of those issues can be closed. There's some info worth keeping in the docs though. Can you add some text about this comment? |
Hmm, microsoft/azure-pipelines-agent#1270 is interesting. It highlights that the skipper phrases only work with Azure Pipelines for merge commits, not for PRs. For PRs, the CI is still triggered. Here is a recent PR from the bot that uses So I don't see a case where I am not sure if CI running on PRs is what conda-forge wants or just a consequence of Azure Pipelines' design decision. conda-forge could add its own rule into its Azure Pipelines templates to skip PRs to save some CI time like numpy did (numpy/numpy#21879 (comment)). |
Uhh yea, that's not great. @beckermr, what do you think about this? Does it interact with the "mergeability" checks?
I think it's mostly an AZP limitation. This logic you linked to could be added to the conda-smithy templates to mitigate this surprising behaviour, but for sure we should document this clearly now so it's not buried in an issue somewhere. |
I believe this is the intended behavior for the linter. If it sees |
Thanks @wshanks, ended up adding conda-forge/conda-smithy#2077 and conda-forge/staged-recipes#27765 for Azure via your suggestion. |
@@ -445,7 +445,7 @@ repo [README.md](https://github.com/conda-forge/webservices-dispatch-action) for | |||
|
|||
### Skipping CI builds | |||
|
|||
To skip a CI build for a given commit, put `[ci skip] ***NO_CI***` in the commit message. | |||
To skip a CI build for a given commit, put `[ci skip]` in the commit message. | |||
|
|||
:::note[Related links] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We can remove this admonition now, I think.
Originally As we have moved from the original setup of Travis, CircleCI and AppVeyor through various additions and removals we have reached the current state of Azure, Travis and GitHub Actions (leveraged directly or using some other self-hosted runner solution) Have forgotten some of the nuances of the different systems in between and how they handled skip notes. That said, Since sometimes CI providers have not consistently done what we have wanted. For a time we added logic to fast cancel jobs in certain conditions (not the latest build for a PR or commit, recipe has a |
Thanks @jakirkham. We just merged a pr to fast finish prs with skip messages for azure in smithy. To be honest it seems like we should just keep the logic around no matter what given the turnover over the years. |
PR Checklist:
docs/
orcommunity/
, you have added it to the sidebar in the corresponding_sidebar.json
fileI noticed that when you use the bot to do non-packaging changes like update maintainer it only uses
[ci skip]
. So is it safe for the documentation to suggest only[ci skip]
now?Also, are the linked issues still relevant? #629 does not seem useful to me. It talks about some nuances with Travis, Circle, and Azure, but I am not sure what to take away as actionable. All my feedstocks only use Azure now. I think conda-forge/staged-recipes#1148 is similarly linked for nuances between CI systems.
Also, I think #498 can be closed since skipping CI is already in the docs.