I forgot to mention that we should also check the findbugs warnings, javac warnings, javadoc warnings and the build HA branch since there were patches committed without Jerkins' +1s. It should not have more warnings than trunk or break the build. Otherwise, we should justify why the warnings are unavoidable.
We did not do a good job earlier. One example is HADOOP-7119. The patch committed introduced findbugs warnigns (HADOOP-7576), javac warnings (HADOOP-8119), javadoc warnings (HADOOP-7898) and broke mvn eclipse:eclipse (HADOOP-7567). Although the javac and findbugs warnings were catched by this Jenkins report, the patch somehow got committed. Nicholas ________________________________ From: Tsz Wo Sze <szets...@yahoo.com> To: "hdfs-dev@hadoop.apache.org" <hdfs-dev@hadoop.apache.org> Sent: Thursday, March 1, 2012 2:10 PM Subject: Re: Merging the HA branch to trunk - Wednesday, February 29th -1 Below are the reasons: - We have not been able to get a stable build on the HA branch since Feb 13. The builds #93 and #94 both failed earlier today. In contrast, the today's build for trunk is stable. It indicates that there are some problems specific to the HA branch. - I also agree with Vinod that the email was not clear about voting. We should put [VOTE] in the subject. Nicholas
