Open
Description
Documentation for Roslyn should specify what encodings of source files are recognized by Roslyn (and what heuritics/defaults are used, if any, when there are no BOM), e.g. UTF8 w/o BOM, UTF8, UTF16 little-endian, UTF16 big-endian, UTF32 little-endian, UTF32 big-endian, etc...
There should be tests for all supported encodings.