[SPARK-8278] Remove non-streaming JSON reader.

[rxin]

No description provided.

[rxin]

This is basically #6728 but brought up to date.

[rxin]

cc @yhuai and @NathanHowell I noticed the streaming jackson implementation doesn't support COLUMN_NAME_OF_CORRUPT_RECORD anymore. Is this going to be a problem?

[SparkQA]

Test build #37731 has finished for PR 7501 at commit 2863883.

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

[SparkQA]

Test build #37733 has finished for PR 7501 at commit 2098345.

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

[SparkQA]

Test build #37739 has finished for PR 7501 at commit 767ec55.

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

[yhuai]

@rxin Streaming jackson implementation does handle corrupt record (see https://github.com/apache/spark/blob/master/sql/core/src/main/scala/org/apache/spark/sql/json/JacksonParser.scala#L39 and https://github.com/apache/spark/blob/master/sql/core/src/main/scala/org/apache/spark/sql/json/InferSchema.scala#L37).

[rxin]

Ah ok. It just reads from the config directly. I'm going to merge this then.