[ https://issues.apache.org/jira/browse/HDFS-10848?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15956412#comment-15956412 ]
Huafeng Wang commented on HDFS-10848: ------------------------------------- Hi [~andrew.wang], sorry I didn't realize the cyclic dependency before submitting a patch. The original hadoop-hdfs-native-client native code is using the MiniDFSCluster for doing test, which introduces the hadoop-hdfs dependency. So once I combine the two client modules into one, there will eventually be a cyclic dependency like hadoop-hdfs-client -> hadoop-hdfs-> hadoop-hdfs-client. Do you have any advice on this one? Thanks in advance. > Move hadoop-hdfs-native-client module into hadoop-hdfs-client > ------------------------------------------------------------- > > Key: HDFS-10848 > URL: https://issues.apache.org/jira/browse/HDFS-10848 > Project: Hadoop HDFS > Issue Type: Improvement > Components: hdfs-client > Reporter: Akira Ajisaka > Assignee: Huafeng Wang > Attachments: HDFS-10848.001.patch > > > When a patch changes hadoop-hdfs-client module, Jenkins does not pick up the > tests in the native code. That way we overlooked test failure when committing > the patch. (ex. HDFS-10844) > [~aw] said in HDFS-10844, > bq. Ideally, all of this native code would be hdfs-client. Then when a change > is made to to that code, this code will also get tested. -- This message was sent by Atlassian JIRA (v6.3.15#6346) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org