Csv Column Inference ignores newline character

[bonjune]

Changed the string reader in the parseHeaders function to erase newline characters before it runs inferences using the input schema.

This will resolve the issue: https://github.com/fsprojects/FSharp.Data/issues/952 .

A test to check this change is added. The test passes.

[nightroman]

I wonder, what about "\r\n"?

            use reader = new StringReader(schema.Replace("\n", ""))

[bonjune]

Oh, that's a valid point. I will make a patch.