Add type types for extends FileInputFormat on BSONFileInputFormat

[catap]

No description provided.

[catap]

@llvtt, ping?

[llvtt]

@catap Thanks for the pull request. Each release of mongo-hadoop tries not to break API/binary compatibility with the previous release, except in the case of new major versions (i.e. mongo-hadoop 2.0). This would be a binary breaking change, so although I generally agree that BSONFileInputFormat probably shouldn't extend the raw FileInputFormat, unfortunately it must remain this way until 2.0.

[catap]

@llvtt thanks for explain the policy.

BTW without this patch mongo-hadoop doesn't work at apache-spark ;)

[llvtt]

This is news to me. Using the boilderplate on the spark-usage page, I can successfully run jobs written in in Java/Scala on Spark. If you're having problems, please file an issue on the HADOOP project here: https://jira.mongodb.org/browse/HADOOP.

[catap]

I've checked it in my code and understand that I'm wrong.

Sorry.