Fix: inference on multiline csv schema

[bonjune]

This resolves the point raised at https://github.com/fsprojects/FSharp.Data/pull/1483#issuecomment-1529633129 by @nightroman

• Defined a dedicated function to read schema string to break it down into strings, separated by comma, skipping newline characters (CR/LF).
• Used ResizeArray to build an array of strings, so .Length changed to .Count when used.