[ https://issues.apache.org/jira/browse/HADOOP-7019?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12973354#action_12973354 ]
Chris Douglas commented on HADOOP-7019: --------------------------------------- This is losing focus. The issue is trying to shorten cross-project build times. It takes no position on the value of the work done during a {{mvn-install}}, only its relevance to the immediate, development context. I hope the following are uncontroversial: # If test-patch takes an extra minute, 10 minutes, or hour to verify the correctness of the framework, it's time well spent. That doesn't require that it happen during development cycles, where things are often partially, temporarily broken. # A patch breaking the test-patch cycle prevents that automated process from validating the framework, admitting preventable regressions. Patches that break it should not be committed, modulo reasonable exceptions. # If a patch _is_ committed that breaks the test-patch cycle, it should not halt all downstream development. I don't care if these are system tests, fault-injection tests, or unit tests. They're taking a long time to build in an inner loop of the dev cycle, and while building them _is important_, pulling it into the outer loop- validation of the patch- is an admissible optimization. > Refactor build targets to enable faster cross project dev cycles. > ----------------------------------------------------------------- > > Key: HADOOP-7019 > URL: https://issues.apache.org/jira/browse/HADOOP-7019 > Project: Hadoop Common > Issue Type: Bug > Components: build > Reporter: Owen O'Malley > Assignee: Luke Lu > Fix For: 0.22.0 > > Attachments: hadoop-7019-trunk-v1.patch, hadoop-7019-trunk-v2.patch > > > The current build always generates fault injection artifacts and pushes them > to Maven. Most developers have no need for these artifacts and no users need > them. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.