[ https://issues.apache.org/jira/browse/YARN-4446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15128590#comment-15128590 ]
Sangjin Lee commented on YARN-4446: ----------------------------------- {quote} We are using context to fill UID after we have retrieved entities from backend storage implementation. The storage implementation can modify the context within and if we use the same reference, the code to fill UID would rely on storage implementation not changing it in a manner where it breaks UID filling code. {quote} Understood. I'm fine with the way it is then. I was curious as to the reason for creating new instances. {quote} We can however in the javadoc ask them to refer to TimelineDataToRetrieve and TimelineEntityFilters classes as well. Where should we add it in these classes though because we have private fields ? So should we have it documented against getters of these classes. Or probably class javadoc for TimelineDataToRetrieve and TimelineEntityFilters can contain the whole info. Thoughts ? {quote} That's what I am thinking too. We can document it in the class javadoc for TimelineDataToRetrieve and TimelineEntityFilters, and in TimelineReader we can refer to those javadocs. > Refactor reader API for better extensibility > -------------------------------------------- > > Key: YARN-4446 > URL: https://issues.apache.org/jira/browse/YARN-4446 > Project: Hadoop YARN > Issue Type: Sub-task > Components: timelineserver > Affects Versions: YARN-2928 > Reporter: Varun Saxena > Assignee: Varun Saxena > Labels: yarn-2928-1st-milestone > Attachments: YARN-4446-YARN-2928.01.patch, > YARN-4446-YARN-2928.02.patch > > -- This message was sent by Atlassian JIRA (v6.3.4#6332)