Fix for encoding

[K0den]

• parseZip now takes encoding parameter and parses it to parseDbf. defaults to "utf-8"

I've no idea what to do about the .cpg code though...

[calvinmetcalf]

I don't think this fix is quite right, I think we actually need a fix here which checks if it's a buffer and if so converts it into a string

[DistChen]

I changed the parseDBF implementation. If no cpg file, I can get encode from dbf file.

1. add ldid(Language driver ID) in dbfHeader：

   function dbfHeader(data) {
       var out = {};
       out.lastUpdated = new Date(data.readUInt8(1) + 1900, data.readUInt8(2), data.readUInt8(3));
       out.records = data.readUInt32LE(4);
       out.headerLen = data.readUInt16LE(8);
       out.recLen = data.readUInt16LE(10);
       // add Language driver ID
       out.ldid = data.readUInt8(29); 
       return out;
   }

   in dbf file, the 29th byte represents Language driver ID，see dbf_file_fmt.

2. modify module.exports:

   module.exports = function(buffer, encoding) {
       var actualEncode = encoding;
       var header = dbfHeader(buffer);
       if(!actualEncode){
           // need provider a ldid map, here is a example.
           var ldid = header.ldid;
           if(ldid === 77){
               actualEncode = "GBK";
           }else if(ldid === 120){
               actualEncode = "big5";
           }
           /*else if (ldid === value){
               actualEncode = "encode";
           }
          ......  */
       }
       var decoder = createDecoder(actualEncode);
       var rowHeaders = dbfRowHeader(buffer, header.headerLen - 1, decoder);
   
       var offset = ((rowHeaders.length + 1) << 5) + 2;
       var recLen = header.recLen;
       var records = header.records;
       var out = [];
       while (records) {
           out.push(parseRow(buffer, offset, rowHeaders, decoder));
           offset += recLen;
           records--;
       }
       return out;
   };

   About the mapping relationship between Language driver ID and code page, see File code page identifiers.

Hope it works for you.

[calvinmetcalf]

this has been fixed for a while