Reading of UTF8-encoded (w/ BOM) files fails

[evil-shrike]

This package would be much helper if it abstracts away utf8's BOM handling. Currently reading fails for files encoded in utf8 (JSON.parse throws "Unexpected symbol" at BOM).

Grunt's file module can solve this problem (see below). Why not to reimplement the similar approach here?

file.read = function(filepath, options) {
  if (!options) { options = {}; }
  var contents;
  grunt.verbose.write('Reading ' + filepath + '...');
  try {
    contents = fs.readFileSync(String(filepath));
    // If encoding is not explicitly null, convert from encoded buffer to a
    // string. If no encoding was specified, use the default.
    if (options.encoding !== null) {
      contents = iconv.decode(contents, options.encoding || file.defaultEncoding);
      // Strip any BOM that might exist.
      if (!file.preserveBOM && contents.charCodeAt(0) === 0xFEFF) {
        contents = contents.substring(1);
      }
    }
    grunt.verbose.ok();
    return contents;
  } catch(e) {
    grunt.verbose.error();
    throw grunt.util.error('Unable to read "' + filepath + '" file (Error code: ' + e.code + ').', e);
  }
};

[jprichardson]

Hmm. It's recommended that in UTF8, there shouldn't be a BOM, but I realize that's not always the case. I'd accept a PR to strip the BOM. However, in writing a JSON file, it's not necessary.

[jprichardson]

Actually, I'll just fix it. Hang tight.