[ https://issues.apache.org/jira/browse/HDFS-9339?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14981286#comment-14981286 ]
Zhe Zhang commented on HDFS-9339: --------------------------------- bq. The createKey method is a local method that wraps the KeyProvider method. One of the things it does is return false if there's an exception. I see, so we {{try}} for {{setup}}. But should we consider the test pass if {{setup}} fails from an exception (and therefore the {{assertTrue}} is not visited)? Maybe remove the {{try-finally}} and let the entire test fail if {{setup}} fails? bq. It would be possible to abstract it out, but it would involve another anonymous inner class I see, thanks for explaining. bq. They aren't actually repeated. My bad, spotted the differences now. +1 pending a conclusion on the first point. > Extend full test of KMS ACLs > ---------------------------- > > Key: HDFS-9339 > URL: https://issues.apache.org/jira/browse/HDFS-9339 > Project: Hadoop HDFS > Issue Type: Test > Components: HDFS > Affects Versions: 2.7.1 > Reporter: Daniel Templeton > Assignee: Daniel Templeton > Attachments: HDFS-9339.001.patch > > > HDFS-9295 adds an end-to-end test for KMS, but it is missing a dimension. > The tests added in that JIRA hold the configuration constant and test that > all operations succeed or fail as expected. More tests are needed that hold > the operation constant and test that all possible configurations cause the > operations to succeed or fail as expected. This JIRA is to add those tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332)