Remove incorrect help text suggesting that the example option can be a r... #769

Merged
merged 1 commit into from Jan 16, 2013

Projects

None yet

2 participants

Contributor

...egexp

Of special interest is that the Cucumber features have been passing, even though this option hasn't supported a Regexp for a long time. Perhaps all the specs are being filtered out and then all the tests are run? Could indicate the Cucumber tests are fragile...

Owner

Thanks. We should look into the issue you mentioned...my guess is that the implementation of the "the examples should all pass" step looks only for "0 failures", which could be "0 examples, 0 failures"--so it might be good to have it verify that some examples were actually run.

@myronmarston myronmarston merged commit 79e10c3 into rspec:master Jan 16, 2013

1 check passed

default The Travis build passed
Details
Owner

Thanks, @shepmaster!

@myronmarston myronmarston added a commit that referenced this pull request Jan 16, 2013
@myronmarston myronmarston Prevent `the examples should all pass` cuke step false positives.
It did not actually verify that any examples were run, and lead
to a false positive cucumber scenario discovered by @shepmaster
in #769. The underlying assumption of "the examples should all pass"
is that there were actually examples.

I think "the process should succeed even though no examples were run"
is a better description of the case when there are no examples.
27a89f6
Owner

@shepmaster -- you were right about the cucumber assertion being a bit messed up. I corrected it in 27a89f6 ; let me know what you think of the solution I came up with.

@shepmaster shepmaster deleted the shepmaster:remove-regexp-example branch Jan 16, 2013
Contributor

@myronmarston Your changes look great; the new step name is very clear about what it will test for. 😃

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment