[SPARK-17985][CORE] Bump commons-lang3 version to 3.5.

[ueshin]

What changes were proposed in this pull request?

SerializationUtils.clone() of commons-lang3 (<3.5) has a bug that breaks thread safety, which gets stack sometimes caused by race condition of initializing hash map.
See https://issues.apache.org/jira/browse/LANG-1251.

How was this patch tested?

Existing tests.

[SparkQA]

Test build #67107 has finished for PR 15525 at commit f318dff.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[ueshin]

Jenkins, retest this please.

[srowen]

I think this is OK. The general issue is whether it will be compatible with other transitive dependencies, but I believe it would remain so.

[SparkQA]

Test build #67111 has finished for PR 15525 at commit f318dff.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[rxin]

Merging in master.

[rxin]

Actually I'm seeing the following exceptions locally as well as on Jenkins for Hadoop 2.2:

[error] /scratch/rxin/spark/core/src/main/scala/org/apache/spark/util/Utils.scala:1489: value read is not a member of object org.apache.commons.io.IOUtils
[error]       var numBytes = IOUtils.read(gzInputStream, buf)
[error]                              ^
[error] /scratch/rxin/spark/core/src/main/scala/org/apache/spark/util/Utils.scala:1492: value read is not a member of object org.apache.commons.io.IOUtils
[error]         numBytes = IOUtils.read(gzInputStream, buf)
[error]                            ^

I'm going to revert the patch for now.

[ueshin]

@rxin I think the failure is not related to this PR.
The builds for hadoop-2.2 are still failing after reverted.
https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.2/1948/console

Looks IOUtils is introduced by 231f39e (#15285) and fixing is in progress at #15537.

[rxin]

Ah alright thanks. Can you submit a new pr so I can merge it again?

[ueshin]

I see. I'll submit soon.

[ueshin]

@rxin I submitted a new pr #15548. Thanks.