s3tests_boto3/functional/test_s3.py: add test for list object versions with key marker #594
Conversation
…s with key marker The key-marker parameter in the list-object-versions operation is inclusive in RGW, while it's exclusive in S3. This commit adds the test to check the corresponding case. Issue ref: https://tracker.ceph.com/issues/68055 Fix PR: ceph/ceph#60068 Signed-off-by: Samaksh Dhingra <[email protected]>
s3tests_boto3/functional/test_s3.py
Outdated
| assert obj['VersionId'] == key_data['VersionId'] | ||
| _compare_dates(obj['LastModified'],key_data['LastModified']) | ||
|
|
||
| @pytest.mark.fails_on_dbstore |
There was a problem hiding this comment.
i don't think this is needed - there are other list_object_versions() tests that work with dbstore
@cbodley, I wanted to know can there also be scenario where just versionIdMarker is passed and keyMarker is not passed? If yes, what is the expected behaviour for the same? Or should we assume they are passed together only? |
|
@samakshd not sure, but it probably doesn't do anything without KeyMarker |
…s with key marker and version id marker, modified key marker test Added test for list object versions with key marker and version id marker. Modified the previous list object versions with key marker test to make it more stronger. Signed-off-by: Samaksh Dhingra <[email protected]>
|
Hi @cbodley, I have added the tests for list object version with key marker and version id marker. I have also modified key marker test to make it more stronger as I realized that my initial PR fix was not correct. I have added another commit to my existing fix PR. However I accidentally did git rebase main to my branch, which introduced all new commits from other contributors, and it got a bit messed up. Since I am new to use git this much extensively, this happened. I reset my branch to first commit, by git hard reset, did fresh changes to it again and merged first with updated main branch and then pushed it. This however did not revert some review requests and tags that got into PR due to the commits that came because of rebasing. Should I create another clean PR for the fix? |
2 participants
The key-marker parameter in the list-object-versions operation is inclusive in RGW, while it's exclusive in S3. This commit adds the test to check the corresponding case.
Issue ref: https://tracker.ceph.com/issues/68055
Fix PR: ceph/ceph#60068