[DONT MERGE] Enabled ImportOrder check of Java code

[LuciferYang]

What changes were proposed in this pull request?

This pr enabled ImportOrder check of Java code

Why are the changes needed?

Does this PR introduce any user-facing change?

No

How was this patch tested?

Pass GA

[LuciferYang]

The ImportOrder check of Java is disabled now and there is a TODO:

spark/dev/checkstyle.xml

Lines 156 to 163 in 4817b01

	<!-- TODO: 11/09/15 disabled - order is currently wrong in many places -->
	<!--
	<module name="ImportOrder">
	<property name="separated" value="true"/>
	<property name="ordered" value="true"/>
	<property name="groups" value="/^javax?\./,scala,*,org.apache.spark"/>
	</module>
	-->

I try to re-enable it and change the rule as follows:

     <module name="ImportOrder">
            <!-- simple imports -->
            <property name="separated" value="true"/>
            <property name="ordered" value="true"/>
            <property name="groups" value="/^javax?\./,scala,*,org.apache.spark"/>
            <!-- static imports -->
            <property name="option" value="bottom"/>
            <property name="staticGroups" value=""/>
            <property name="sortStaticImportsAlphabetically" value="true"/>
            <property name="separatedStaticGroups" value="true"/>
        </module>

The no-static imports is consistent with scala code, the static imports is an independent group after no-static imports, and it sorted by alphabetically.

@srowen @dongjoon-hyun @HyukjinKwon I wonder if we need to complete this TODO ?

[LuciferYang]

The example of following the ImportOrder check rule:

import java.io.IOException;
import java.nio.ByteBuffer;


import scala.Tuple2;

import com.google.common.collect.ImmutableMap;
import com.google.common.collect.Sets;

import org.apache.spark.network.buffer.ManagedBuffer;
import org.apache.spark.network.buffer.NioManagedBuffer;

import static org.apache.spark.network.shuffle.RetryingBlockTransferor.BlockTransferStarter;
import static org.junit.Assert.*;
import static org.mockito.Mockito.*;

[srowen]

I think this is mostly too much noise to bother with. Most imports are nearly-correctly-ordered and it matters little anyway. If there is a weaker check we can enforce (e.g. are imports grouped, if not ordered, correctly) that might be OK to try to enforce

[LuciferYang]

So we can accept a weak rule similar to the following: imports are grouped and interior of each group can be disordered ?

[LuciferYang]

Set order to false can ignore the order check

[LuciferYang]

@srowen for static imports, would you mind dividing them into a separate groups?

[LuciferYang]

So we can accept a weak rule similar to the following: imports are grouped and interior of each group can be disordered ?

If this, still 180 files need to be changed :(

[srowen]

I see, it's still a big change then. I'm not sure it's worth it.

[LuciferYang]

So should we remove this TODO, It seems unlikely to have a chance to finish it ...

[srowen]

We could, I also don't care much about the TODO statement either

[LuciferYang]

We could, I also don't care much about the TODO statement either

ok