PJ Fanning created HADOOP-19076:
-----------------------------------
Summary: move jersey code in hadoop-common jar to a new
hadoop-jersey1-common jar
Key: HADOOP-19076
URL: https://issues.apache.org/jira/browse/HADOOP-19076
Project: Hadoop Common
Issue Type: Task
Reporter: PJ Fanning
Hadoop's Jersey dependencies are causing us real trouble.
I'm wondering if it would be a good idea to take the jersey and javax.ws code
out of hadoop-common and move it into a dedicated hadoop-jersey1-common jar. We
could later create a hadoop-jersey2-common (or hadoop-jersey3-common - because
Jersey 3 is out and maybe better to skip to Jersey 2).
hadoop-jersey1-common and hadoop-jersey2-common would have equivalent classes -
just depend on different versions of Jersey.
Example code:
*
https://github.com/apache/hadoop/blob/12498b35bbb754225b0b2ca90d5ad4f5cf628d56/hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/http/HttpServer2.java#L1030
*
https://github.com/apache/hadoop/blob/12498b35bbb754225b0b2ca90d5ad4f5cf628d56/hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/util/HttpExceptionUtils.java#L89
Hadoop modules that need access to the common jersey code could start with
depending on hadoop-jersey1-common but later be refactored to use
hadoop-jersey2-common. We could do this on a module by module basis (one a
time).
hadoop-common jar would have its jersey and jsr311-api dependencies removed.
Wdyt [~slfan1989], [~steve_l], [~ayushsaxena] ?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
