// TODO: remove method if not needed
public void Test(int skip, IDCMapping mapping, IDCFileInfo fileInfo)
{
if (m_traceSwitch.TraceVerbose)
{
Debug.WriteLine("DataConverter Test");
Debug.WriteLine("\tDCMapping");
Debug.WriteLine("\t\tBeginMarker =" + mapping.BeginMarker + ".");
Debug.WriteLine("\t\tDomain =" + mapping.Domain + ".");
Debug.WriteLine("\t\tEndMarker =" + mapping.EndMarker + ".");
Debug.WriteLine("\tIDCFileInfo");
Debug.WriteLine("\t\tFileEncoding =" + fileInfo.FileEncoding + ".");
// Debug.WriteLine("\t\tFileEncodingSource =" + fileInfo.FileEncodingSource + ".");
Debug.WriteLine("\t\tInputFileName =" + fileInfo.InputFileName + ".");
Debug.WriteLine("\t\tOutputFileName =" + fileInfo.OutputFileName + ".");
}
}
/// <summary>
/// A Factory method to create the correct type of MarkerSpec object.
/// </summary>
/// <param name="mapping"></param>
/// <returns></returns>
public static MarkerSpec CreateMarkerSpec(IDCMapping mapping)
{
MarkerSpec ms;
ms = CreateMarkerSpec(mapping.BeginMarker, mapping.MarkerEncoding, mapping.DataEncoding,
mapping.IsInline, mapping.NewBeginMarker, mapping.EndMarker, mapping.NewEndMarker);
return ms;
}
public void ConvertNew(IDCMapping[] mappings, IDCFileInfo[] fileInfos)
{
// Test(1, mappings[0], fileInfos[0]);
m_tokenizer = new Tokenizer();
foreach (IDCMapping mapping in mappings)
{
if (mapping.MarkerEncoding.Length <= 0)
mapping.MarkerEncoding = DefaultMarkerMap;
if (mapping.DataEncoding.Length <= 0)
mapping.DataEncoding = DefaultDataMap;
MarkerSpec ms = MarkerSpec.CreateMarkerSpec(mapping);
m_tokenizer.Tri.Add(ms);
}
Token token;
string output;
FileStream stream;
StreamReader streamReader;
Stream outputStream;
bool fBOM;
StreamWriter outputWriter = null;
// Do for each input file in fileInfo
System.Text.Encoding encoding;
foreach (IDCFileInfo fileInfo in fileInfos)
{
stream = new FileStream(fileInfo.InputFileName,
FileMode.Open, FileAccess.Read);
bool fAlreadyUnicode = true;
switch (fileInfo.FileEncoding)
{
case DCFileEncoding.DC_FE_BYTES:
case DCFileEncoding.DC_FE_Unknown:
encoding = ReversibleEncoding;
fAlreadyUnicode = false;
break;
case DCFileEncoding.DC_FE_UTF16BE:
encoding = System.Text.Encoding.BigEndianUnicode;
break;
case DCFileEncoding.DC_FE_UTF16LE:
encoding = System.Text.Encoding.Unicode;
break;
case DCFileEncoding.DC_FE_UTF8:
encoding = System.Text.Encoding.UTF8;
break;
default:
Debug.Fail("Requested input file encoding not implemented.");
encoding = ReversibleEncoding;
fAlreadyUnicode = false;
break;
}
streamReader = new StreamReader(stream, encoding);
m_tokenizer.Input = streamReader;
outputStream = new FileStream(fileInfo.OutputFileName, FileMode.Create,
FileAccess.Write);
fBOM = fileInfo.HasBOM;
if (fBOM)
{
// Use StreamWriter if BOM needed.
outputWriter = new StreamWriter(outputStream, System.Text.Encoding.UTF8);
}
do
{
// Enhance (BobbyD): It seems that all the StreamWriters output a BOM,
// even if we don't want one. One solution is below, that is, write the data
// using Streams instead of StreamWriters. Except this is sort of messy,
// a cleaner solution may be to subclass Encoding.UTF8, override the method
// GetPreamble() to return a zero length byte array, instead of the UTF8 BOM,
// and then the nice and clean StreamWriter can be used. More information on
// this is under the help topic of Encoding.GetPreamble Method.
token = m_tokenizer.Next();
output = token.Output(fAlreadyUnicode);
byte[] bytes = System.Text.Encoding.UTF8.GetBytes(output);
if (token is NewlineToken)
{
if (fBOM)
outputWriter.WriteLine(output);
else
{
outputStream.Write(bytes, 0, bytes.Length);
outputStream.WriteByte((byte)'\r');
outputStream.WriteByte((byte)'\n');
}
}
else
{
if (fBOM)
outputWriter.Write(output);
else
outputStream.Write(bytes, 0, bytes.Length);
}
} while (!(token is EndOfFileToken));
m_tokenizer.Input.Close(); // close the input stream
if (fBOM)
outputWriter.Close();
else
outputStream.Close();
}
}