[ https://issues.apache.org/jira/browse/HDFS-7769?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14329890#comment-14329890 ]
Chris Douglas commented on HDFS-7769: ------------------------------------- As written, the bylaws require a +1 from a committer. It's been enforced fairly strictly in MapReduce/YARN. The language in the bylaws is pretty direct: bq. Consensus approval of active committers, but with a minimum of one +1 The ambiguity is whether the contributor can self-approve patches, or (unwritten) +1 a trivial patch with any reviewer's approval? I'd say no, if only to avoid arguments over what constitutes a "trivial" patch. bq. it does not quite make sense for non-committers reviewing patches. As least, no one has incentive to ask a non-committers to reviewing a patch. Non-committers review patches to provide feedback on them and signal that the patch has already had a pair of eyes on it. I'm more likely to +1 a patch if someone else has done a pass on it and provided good feedback. Requiring another committer to review also blocks development on "dead" areas of the codebase with only a single maintainer. Those either need to move out of the project or find new committers to help maintain it. > TestHDFSCLI create files in hdfs project root dir > ------------------------------------------------- > > Key: HDFS-7769 > URL: https://issues.apache.org/jira/browse/HDFS-7769 > Project: Hadoop HDFS > Issue Type: Bug > Components: test > Reporter: Tsz Wo Nicholas Sze > Assignee: Tsz Wo Nicholas Sze > Priority: Trivial > Fix For: 2.7.0 > > Attachments: h7769_20150210.patch, h7769_20150210b.patch > > > After running TestHDFSCLI, two files (data and .data.crc) remain in hdfs > project root dir. -- This message was sent by Atlassian JIRA (v6.3.4#6332)