On Tue, Jan 12, 2016 at 12:36 PM, Allen Wittenauer <[email protected]> wrote: > >> On Jan 12, 2016, at 10:20 AM, Ted Yu <[email protected]> wrote: >> >> For HBASE-15075 , I looked at: >> >> https://builds.apache.org/job/PreCommit-HBASE-Build/69/artifact/patchprocess/patch-unit-hbase-server-jdk1.8.0_66.txt >> >> [ERROR] COMPILATION ERROR : >> [INFO] ------------------------------------------------------------- >> [ERROR] >> /testptch/hbase/hbase-server/src/main/java/org/apache/hadoop/hbase/master/normalizer/SplitNormalizationPlan.java:[95,12] >> no suitable method found for splitRegion(byte[],byte[],java.util.UUID) >> method org.apache.hadoop.hbase.client.Admin.splitRegion(byte[]) is >> not applicable >> (actual and formal argument lists differ in length) >> method org.apache.hadoop.hbase.client.Admin.splitRegion(byte[],byte[]) >> is not applicable >> (actual and formal argument lists differ in length) >> >> Looks like the following addition to Admin.java was not effective: >> >> + void splitRegion(final byte[] regionName, final byte[] splitPoint, >> final UUID id) >> >> Does someone have idea why the compilation failed ? >> >> BTW I successfully compiled using jdk-8u65 locally. > > Without looking at the patch, this usually means the local maven > cache doesn’t have a dependency in place, typically because the order the > personality queued the modules isn’t correct. By default, Yetus does maven > modules in fs glob order. It doesn’t have the smarts yet to order these > based upon the Maven dependencies. So for now, personalities need to know > how to do this.
For a very extensive example, see the example personality based on Apache Hadoop: https://github.com/apache/yetus/blob/master/precommit/personality/hadoop.sh#L78
