* Recursively remove args with omit as as their value, to match suboptions in argument_spec
* Add test for the new remove_omit function
(cherry picked from commit 7404dc6767)
* Don't skip an inventory source just because it has a comma, make sure it's also doesn't exist as a path. Fixes#34931
* Add integration test for inventory path with commas
(cherry picked from commit 1ac04ba52e)
Digital Ocean changed their api to no longer accept an integer as a resource_id. Fix adds changed data type (from integer to string).
Signed-off-by: Anthony Bond <bond.anthony@outlook.com>
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
(cherry picked from commit f9b138b0ab)
* Ignore become if we shouldn't be sudoable
* More explicit passing of kwargs
* dd/piped put should be sudoable=False. Fixes#34523
(cherry picked from commit 03794864c2)
This reverts commit 8c822018db.
There's some other bugs in 2.4's import/include that aren't blockers for
2.4.3 (becaues they aren't regressions) that prevent the new integration
tests from passing. Revert the integration tests.
* First pass at include_role tests
* Reorganize test structure
Do all import and include tests in a single target.
* Build out more tests and test with linear and free strategy for each type
* import_role tests
* Set target host for play test
* Basic import_playbook tests
* Basic import/include_tasks tests
* Add recursion test
* import_playbook tests
* Add import_playbook group_vars test
Issue #33177
* Additional group_var tests for import_playbook
* Enable recursion test
* More work on import_tasks
* Run all tests via run.sh, improve conditional tests, add never tag
Add never tag to certain failining tests and explicitly skip it for now until #34104 is merged. Some tasks need to remain commented out because they cause the entire play to fail and aren't skippable by tags because they are syntax and/or variable checks.
Improve test for whether or not a role was run based on a conditional. Since the [import|include]_role does not register a variable, use a variable that is set inside a role as a canary test for whether or not the role was run.
* Use a fail task rather than trigger a failure via bogus command
This should resolve the "unstable" test results.
* Import tag tests
Fix bug it tests where validate tasks weren't run on tag tests.
Add tests for task import/include with tags.
* Remove test for playbook group_var inheritance
(cherry picked from commit 10a8c6bc25)