[
https://issues.apache.org/jira/browse/HDFS-7888?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14391168#comment-14391168
]
Jing Zhao commented on HDFS-7888:
---------------------------------
Thanks for working on this, Bo. The patch looks good to me. Some minors:
# how about using {{getSimpleName}} here?
{code}
+ String path = this.getClass().getName();
+ String className = path.substring(path.lastIndexOf('.') + 1);
TraceScope scope =
- dfsClient.getPathTraceScope("DFSOutputStream#writeChunk", src);
+ dfsClient.getPathTraceScope(className + "#writeChunk", src);
{code}
# {{encounterBlockBoundary}} may not be an appropriate method name. maybe we
can rename it to something like {{endBlock}}?
> Change DataStreamer/DFSOutputStream/DFSPacket for convenience of subclassing
> ----------------------------------------------------------------------------
>
> Key: HDFS-7888
> URL: https://issues.apache.org/jira/browse/HDFS-7888
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Li Bo
> Assignee: Li Bo
> Attachments: HDFS-7888-001.patch, HDFS-7888-trunk-001.patch,
> HDFS-7888-trunk-002.patch
>
>
> HDFS-7793 refactors class {{DFSOutputStream}} on trunk which makes
> {{DFSOutputStream}} a class without any inner classes. We want to subclass
> {{DFSOutputStream}} to support striping layout writing. This JIRA depends
> upon HDFS-7793 and tries to change DataStreamer/DFSOutputStream/DFSPacket for
> convenience of subclassing.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
