Closed reid-spencer closed 10 months ago
@lihaoyi - Locally, running the tests succeed. In the CI, they fail because of the return code (1 instead of 0). The only difference I can see is hardware and the JVM is 17 not 8. Do you have a "must pass all tests in CI" policy before merging?
Yes, we want CI to be green before merging. The last CI run on master was green, so I would expect any PRs to be green as well
Yes, we want CI to be green before merging.
Good, I concur.
The last CI run on master was green, so I would expect any PRs to be green as well
I've merged nothing to main, yet. See PR #297 and its various failures, if you're interested.
I'll continue to look for solutions.
The
Fail
operator appears to ignore its documentation and always produces a failure message of "fail". The documentation shows this example:I believe the intent of using
Fail
withinfailWithLabel
, based on its name, is to provide a new failure label. However, the assertion on the last line shows the issue. I believe it should say:If you look at this line of code you can see the mistake:
Changing
"fail"
(static message) tomsg
will make use of the message and provide much more significant value in reporting custom error messages in fastparse.