bgaborg edited a comment on issue #1077: HADOOP-16380. S3Guard tombstones can mislead about directory empty status - wip URL: https://github.com/apache/hadoop/pull/1077#issuecomment-511379438 the test `testRootTombstones` will pass if I change the return new `S3AFileStatus(Tristate.TRUE, path, username);` to return new `S3AFileStatus(Tristate.FALSE, path, username);` at `org.apache.hadoop.fs.s3a.S3AFileSystem#s3GetFileStatus (org/apache/hadoop/fs/s3a/S3AFileSystem.java:2760)`. I'll update my PR with a test against this. Although we will have the following errors after the change: ``` ITestS3GuardToolDynamoDB.testDynamoDBInitDestroyCycle:250->Assert.assertFalse:64->Assert.assertTrue:41->Assert.fail:88 s3guard.test.testDynamoDBInitDestroy-403465466 still exists; ITestS3GuardDDBRootOperations.test_400_rm_root_recursive:237->Assert.assertTrue:41->Assert.fail:88 Root directory delete failed; ITestS3AContractRootDir>AbstractContractRootDirectoryTest.testRmEmptyRootDirNonRecursive:118 ? PathIO ``` So at least we are checking for this. It would be better if we had a unit test for this imho. I'll create one if I can - I'll check the feasibility.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
