> On Nov. 11, 2014, 7 p.m., Qian Xu wrote: > > common/src/main/java/org/apache/sqoop/json/util/SchemaSerialization.java, > > line 96 > > <https://reviews.apache.org/r/27881/diff/5/?file=758636#file758636line96> > > > > Please keep setter chain in multiple lines
I just use the standard we have, unless we have a standardised rules, I rather not pick on such things. > On Nov. 11, 2014, 7 p.m., Qian Xu wrote: > > common/src/main/java/org/apache/sqoop/schema/NullSchema.java, line 24 > > <https://reviews.apache.org/r/27881/diff/5/?file=758637#file758637line24> > > > > I hate to add complexity, but `NullSchema` should be immutable, right? > > All setter methods should be hide. Maybe singleton implementation as well? sure. > On Nov. 11, 2014, 7 p.m., Qian Xu wrote: > > common/src/main/java/org/apache/sqoop/json/util/SchemaSerialization.java, > > line 72 > > <https://reviews.apache.org/r/27881/diff/5/?file=758636#file758636line72> > > > > The `Initializer` will return a `NullSchema` instance now, if no > > specific schema is provided. So null as schema will violate the design. As > > other methods do not have defensive check against null, is the check for > > schema necessary? I am not sure I understand the violation, part, since I changed it to be a NullSchema class, this check is not required. I will remove it. can you clarify what you mean by violation - Veena ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/27881/#review60935 ----------------------------------------------------------- On Nov. 11, 2014, 4:45 p.m., Veena Basavaraj wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/27881/ > ----------------------------------------------------------- > > (Updated Nov. 11, 2014, 4:45 p.m.) > > > Review request for Sqoop. > > > Bugs: SQOOP-1621 > https://issues.apache.org/jira/browse/SQOOP-1621 > > > Repository: sqoop-sqoop2 > > > Description > ------- > > see jira > > The point of this ticket is to guard the connector developers. > > Internally in Sqoop we can still create a empty Schema object and male sure > we dont have to worry about guarding everyy single place a NPE for null > schema can happen:) > > hence the code change is in the SchemaSerialization class so that we handle > null schema > > > Diffs > ----- > > common/src/main/java/org/apache/sqoop/json/util/SchemaSerialization.java > ecb5aa3 > common/src/main/java/org/apache/sqoop/schema/NullSchema.java PRE-CREATION > > connector/connector-hdfs/src/main/java/org/apache/sqoop/connector/hdfs/HdfsFromInitializer.java > 4c6f566 > > connector/connector-hdfs/src/main/java/org/apache/sqoop/connector/hdfs/HdfsToInitializer.java > bce72b5 > docs/src/site/sphinx/ConnectorDevelopment.rst 58b3b61 > > execution/mapreduce/src/main/java/org/apache/sqoop/job/mr/MRConfigurationUtils.java > d5f74f0 > > execution/mapreduce/src/test/java/org/apache/sqoop/job/mr/TestMRConfigurationUtils.java > fbe3e7b > spi/src/main/java/org/apache/sqoop/job/etl/Initializer.java d66b099 > > Diff: https://reviews.apache.org/r/27881/diff/ > > > Testing > ------- > > yes ( unit and integration ) > > > Thanks, > > Veena Basavaraj > >
