[jira] Updated: (HADOOP-2653) [HQL] Unnecessary HQLClient Object creation in a shell loop.

     [ https://issues.apache.org/jira/browse/HADOOP-2653?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack updated HADOOP-2653:
--------------------------

    Priority: Trivial  (was: Minor)

Thanks for adding a description (Considered description make the life of a reviewer easier).  IIUC,  this patch is 'optimizing' shell.

If so, this issue is trivial -- I set it so -- and we all should be working on more important issues.

What does the below test?

{code}
+  public void test() {
+    HQLClient hql = new HQLClient(new HBaseConfiguration());
+    ReturnMsg rm = hql.executeQuery("non-command;");
+    assertTrue(rm.getType() == -9);
+  }
{code}

Should be a javadoc. or comment because I cannot from reading the code figure what it does or what the -9 expected result means (These -9s are in a few places.  They should be defines if all the -9s mean the same thing).

Why the HQLClient constructor change using setters instead of passing params to the Constructor?  It seemed better the way it was previous.

Is the offset of 9 safe in the below?

{code}
+    conf.set("hbase.master", masterAddress.substring(9, masterAddress.length()));
{code}

If so, should be accompanied by a coment saying why (Why can't passed masterAddress by null or not match the expected pattern)?

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.