tests: ported authselect tests from downstream#349
Conversation
e00d242 to
bb2b6c0
Compare
bb2b6c0 to
5483fad
Compare
|
RHEL8 and RHEL9 are stuck with |
|
Is this ready for review? I think you said that you are still working on it. |
|
These tests should work for 1.2, I thought we discussed major releases we will use git branches, and minor releases, use pytest markers. I need to finish the PAM PR before this can be merged. |
If these will work for 1.2 then we can backport it to 1.2.x and run it for 8 and 9 from this branch. RHEL10 can use master.
Thanks, let me know when it is ready for review. |
5483fad to
1fc6ec9
Compare
1fc6ec9 to
c76673c
Compare
|
@pbrezina This is ready for review. Also, I do not have developer access to authselect, so I'm unable to assign this issue to myself. |
|
Hi, I invited you to the organization. I am not sure if I will be able to review this before the Christmas break (I'll try). Can you please also make the tests run during PR CI? See SSSD's ci.yaml for reference. |
pbrezina
left a comment
There was a problem hiding this comment.
Hi Dan,
please, include mhc.yaml, polarion.yaml .
All tests here are for authselect, so test_authselect.py and test_authselect prefix does not make sense. It would be better to organize the test differently eg by profile.
Included test may be organized as:
- test_sssd
- test_minimal
- test_distribution (eg test_authselect__when_removed_the_uninstall_is_clean)
Tests "enable_and_disable" are not sufficient. Presence of specific pam module is irrelevant, what matters is functionality. You must test that the functionality is working correctly.
test_authselect__when_removed_the_uninstall_is_clean - the problem is not RHEL vs Fedora but behavior change between authselect version. This test belongs only to 1.2.x branch.
|
@pbrezina ack, changing a lot of things and removing the non-functional tests, and making the placeholder/sanity into functional before the next review. |
4a5218b to
39356d4
Compare
|
@pbrezina This is ready for review now. It depends on the following PRs. SSSD/sssd-test-framework#86 |
39356d4 to
87e648a
Compare
pbrezina
left a comment
There was a problem hiding this comment.
Thanks, this is much better and I think we can go with this for now.
In long term, however, I would like to rework it to be less sssd-like and more focused on authselect. We can not test winbind or nis stuff with the sssd framework. Lots of features are shared between profiles so we could use topology parametrization here, with the topology representing a profile. We are not that much interested in the sssd providers since we rely on the posix interface.
But its good enough for the moment. Thank you.
87e648a to
3808d32
Compare
pbrezina
left a comment
There was a problem hiding this comment.
Hi Dan, see the inline comment. Also please, rebase it properly without the merge commit and we can push. Thank you.
911f7b5 to
405c7cd
Compare
405c7cd to
a621f77
Compare

Uh oh!
There was an error while loading. Please reload this page.