-
Notifications
You must be signed in to change notification settings - Fork 459
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
cmd/critest: fix empty ginkgo flag's value issue #930
cmd/critest: fix empty ginkgo flag's value issue #930
Conversation
In commit 53ad8bb[1], the ginkgo uses flags.StringVar for some flags, for instance, the --skip or --focus. But in the following release of ginkgo(>=v1.15.0[2]), the ginkgo starts to use string slice instead of the StringVar, which means we can't get the user's input by the flags.Value.String()[3]. The String() always returns empty value. Based on that, we should use ginkgo API to generate flags instead of manually handler in cri-test. Reference: [1]: https://github.com/kubernetes-sigs/cri-tools/blob/53ad8bb7f97e1b1d1c0c0634e43a3c2b8b07b718/vendor/github.com/onsi/ginkgo/config/config.go L79 [2]: https://github.com/onsi/ginkgo/blob/v1.15.0/config/config.go L84 [3]: https://github.com/onsi/ginkgo/blob/v1.15.0/config/config.go L68-L71 Signed-off-by: Wei Fu <fuweid89@gmail.com>
651ff15
to
d4fe3e7
Compare
My local test result is like
Both the skip and parallel mode are working. |
/retest |
@fuweid: Cannot trigger testing until a trusted user reviews the PR and leaves an In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
Ping @feiskyer @saschagrunert |
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.
Thank you!
/lgtm
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: fuweid, saschagrunert The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip.
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip.
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip.
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip.
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip. Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com>
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> (cherry picked from commit 7300296) As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip. Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com>
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> (cherry picked from commit 7300296) As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip. Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com>
Signed-off-by: Paul S. Schweigert <paulschw@us.ibm.com> As discussed in containerd#6903, running ginkgo tests in parallel while trying to skip wasn't working. However, now that kubernetes-sigs/cri-tools#930 has fixed the issue upstream, we can revert back to running our tests in parallel with the skip.
What type of PR is this?
/kind bug
What this PR does / why we need it:
In commit 53ad8bb 1, the ginkgo uses
flags.StringVar for some flags, for instance, the --skip or --focus.
But in the following release of ginkgo(>=v1.15.0 2), the ginkgo starts to
use string slice instead of the StringVar, which means we can't get the
user's input by the flags.Value.String() 3. The String() always returns
empty value.
Based on that, we should use ginkgo API to generate flags instead of
manually handler in cri-test.
Which issue(s) this PR fixes:
Fixes: #925
Special notes for your reviewer:
Does this PR introduce a user-facing change?