-
-
Notifications
You must be signed in to change notification settings - Fork 25.9k
Travis does not run make test-doc
when pytest
fails
#12237
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
Comments
At the beginning of the bash script, it contains What is the desired behavior for script:
- bash build_tools/travis/test_script.sh
- bash build_tools/travis/test_docs.sh And add |
I don't really see a big problem with the current behaviour. Running
test-doc in all cases would make it even harder to find the errors in the
logs. And I think if both were Travis script entries, it would still stop
at the first failure.
|
I think I agree with Joel here. We usually get a long log from Travis and this will make users hard to find the errors. And @thomasjpfan has figured out that this is the expected behavior of the script. |
The only problem is about frustration: when you fix all errors in the travis log, you expect it to go green the next time. Also a passing |
I guess we're talking about situations when normal tests and doc tests fail. Will it be difficult to locate the failures in normal tests then? |
I don't think so, there would still be about 50 lines between the end of the failed normal tests and the beginning of failed test-doc. |
OK, I think I'm persuaded by you at this point. Do you have any response to Joel's comment : "And I think if both were Travis script entries, it would still stop at the first failure." (I guess Travis will continue?) |
From Travis docs,
So I believe @thomasjpfan PR is a correct solution |
Fine, this is exactly what I've found. I'm persuaded by you (though I'll leave the PR to someone more familiar with Travis to ensure the correctness of it). ping @jnothman |
The file that travis reads for testing is test_script.sh, and it is supposed to run
pytest
and thenmake test-doc
However, it seems that
make test-doc
isn't run if the firstpytest
command fails. This happened to me on #11705:make test-doc
isn't called: https://travis-ci.org/scikit-learn/scikit-learn/jobs/435669785make test-doc
is run (and fails): https://travis-ci.org/scikit-learn/scikit-learn/jobs/435711770The issues in the second log could have been caught the first time, avoiding a reviewing round.
I have no idea why
make test-doc
isn't run though, looking attest_script.sh
there's no reason to stop testing if the firstpytest
failsThe text was updated successfully, but these errors were encountered: