Ignore UTF-8 BOM

[h4cc]

The Parser seems to fail for files with a UTF-8 BOM set:

** (CaseClauseError) no case clause matching: {:error, {1, :gpb_scan, {:illegal, [65279]}}, 1}

[tomas-abrahamsson]

Yes, BOMs are not expected. It seems Google protobuf accepts a BOM in 3.0.0-beta, but not in 2.6.1. I'll take a look.

[tomas-abrahamsson]

I just pushed 3.24.0, which skips over any UTF-8 BOMs. (hence closing this issue.)

(It also utf8-decodes the input .proto files, and utf8-encodes the generated files, but generally, this should be a nop in most cases. Only issue I can think of is non-ASCII in default values for strings, supposedly uncommon, but just for your information in any case)