CI: Checks job aborted if a step fails GH30298

[suzutomato]

• closes CI: Checks job aborted if a step fails #30298
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[pep8speaks]

Hello @suzutomato! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file pandas/io/common.py:

Line 84:6: E111 indentation is not a multiple of four

Comment last updated at 2019-12-25 07:43:48 UTC

[WillAyd]

lgtm @datapythonista

[datapythonista]

Thanks @suzutomato, looks good, if you revert the failing changes we can merge this, looks like the always() is fixing the problem.

There is a problem for the benchmarks artifact, but it's unrelated to this, I'll create a separate issue.

[suzutomato]

The intentional errors have been removed.
As per the last comment from @datapythonista, I won't take action for the problem with the benchmarks artifact at this moment.
Let me know if anything I can do to solve the problem.

[datapythonista]

I guess the CI errors are unrelated, can you merge master into your branch to see if the errors get fixed please?

git fetch ustream
git merge upsream/master
git push

[alimcmaster1]

@suzutomato mind finishing this whenever you have a second - definitely looking forward to this feature

[suzutomato]

So sorry, I overlooked @atapythonista's last message.
Working on it now.

[datapythonista]

Thanks!

I'm wondering if always means it'll always run, and ideally we want to run only if the previous step run. But let's merge this, as its surely an improvement and see if we can iterate further later.

[alimcmaster1]

LGTM - sure improvement

Side note maybe we really want if: success() || failure(). As the changes in this PR will continue to run the workflow even if it has been cancelled as per:
https://help.github.com/en/actions/automating-your-workflow-with-github-actions/contexts-and-expression-syntax-for-github-actions.

Assume we are currently using the free plan? https://help.github.com/en/actions/automating-your-workflow-with-github-actions/about-github-actions#usage-limits

@datapythonista

[datapythonista]

I'll give a try to success() || failure(), I think it's what we actually want