[SPARK-5038] Add explicit return type for implicit functions.

[rxin]

As we learned in #3580, not explicitly typing implicit functions can lead to compiler bugs and potentially unexpected runtime behavior.

This is a follow up PR for rest of Spark (outside Spark SQL). The original PR for Spark SQL can be found at #3859

[SparkQA]

Test build #24967 has started for PR 3860 at commit 73702f9.

• This patch merges cleanly.

[SparkQA]

Test build #24967 has finished for PR 3860 at commit 73702f9.

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

[AmplabJenkins]

Test FAILed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24967/
Test FAILed.

[rxin]

Jenkins, retest this please.

[SparkQA]

Test build #24974 has started for PR 3860 at commit 73702f9.

• This patch merges cleanly.

[JoshRosen]

It looks like the org.apache.spark.scheduler.TaskResultGetterSuite.task retried if result missing from block manager test has somehow become non-deterministically flaky...

[SparkQA]

Test build #24974 has finished for PR 3860 at commit 73702f9.

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

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24974/
Test PASSed.

[JoshRosen]

LGTM. I guess the only way that this could introduce problems would be if you changed the explicit type to be wider than the implicit type that was being inferred before, but it doesn't look like that's the case here.

[rxin]

Thanks. Merging in master.