// This function is the meat of the conversion process. It is really long, which
// normally wouldn't be a virtue (especially as an "in-line" function), but in an
// effort to save memory fragmentation by using stack memory to buffer the input
// and output data, I'm using the alloca memory allocation function. Because of this
// it can't be allocated in some subroutine and returned to a calling program (or the
// stack will have erased them), so it has to be one big fat long function...
// The basic structure is:
//
// o Check Input Data
// o Give the sub-class (via PreConvert) the opportunity to load tables and do
// any special preprocessing it needs to ahead of the actual conversion
// o Possibly call the TECkit COM interface to convert Unicode flavors that the
// engine (for this conversion) might not support (indicated via PreConvert)
// o Normalize the input data to a byte array based on it's input EncodingForm
// o Allocate (on the stack) a buffer for the output data (min 10000 bytes)
// o Call the subclass (via DoConvert) to do the actual conversion.
// o Normalize the output data to match the requested output EncodingForm (including
// possibly calling the TECkit COM interface).
// o Return the resultant BSTR and size of items to the output pointer variables.
//
protected virtual unsafe string InternalConvertEx
(
EncodingForm eInEncodingForm,
string sInput,
int ciInput,
EncodingForm eOutEncodingForm,
NormalizeFlags eNormalizeOutput,
out int rciOutput,
bool bForward
)
{
Util.DebugWriteLine(className, "BEGIN");
Util.DebugWriteLine(className,
"eEncFormIn " + eInEncodingForm.ToString() + ", " +
"eEncFormOut " + eOutEncodingForm.ToString());
if (sInput == null)
{
EncConverters.ThrowError(ErrStatus.IncompleteChar);
}
if (sInput.Length == 0)
{
rciOutput = 0;
return("");
}
#if DEBUG && __MonoCS__
// for debugging only BEGIN
//byte[] baIn = System.Text.Encoding.UTF8.GetBytes(sInput); // works
byte[] baIn = System.Text.Encoding.BigEndianUnicode.GetBytes(sInput); // easier to read
Util.DebugWriteLine(className, Util.getDisplayBytes("Input BigEndianUnicode", baIn));
baIn = System.Text.Encoding.Unicode.GetBytes(sInput);
Util.DebugWriteLine(className, Util.getDisplayBytes("Input Unicode", baIn));
int nInLen = sInput.Length;
byte [] baIn2 = new byte[nInLen];
for (int i = 0; i < nInLen; i++)
{
baIn2[i] = (byte)(sInput[i] & 0xFF);
}
Util.DebugWriteLine(className, Util.getDisplayBytes("Input Narrowized", baIn2));
/*
* System.Text.Encoding encFrom = System.Text.Encoding.GetEncoding(12000);
* System.Text.Encoding encTo = System.Text.Encoding.UTF8;
*
* // Perform the conversion from one encoding to the other.
* Util.DebugWriteLine(className, "Starting with " + baIn.Length.ToString() + " bytes.");
* byte[] baOut2 = System.Text.Encoding.Convert(encFrom, encTo, baIn);
* Util.DebugWriteLine(className, "Converted to " + baOut2.Length.ToString() + " bytes.");
* string resultString = System.Text.Encoding.Default.GetString(baOut2, 0, baOut2.Length);
* Util.DebugWriteLine(className, "Test output '" + resultString + "'");
*/
// for debugging only END
#endif
// if the user hasn't specified, then take the default case for the ConversionType:
// if L/RHS == eLegacy, then LegacyString
// if L/RHS == eUnicode, then UTF16
CheckInitEncForms
(
bForward,
ref eInEncodingForm,
ref eOutEncodingForm
);
// allow the converter engine's (and/or its COM wrapper) to do some preprocessing.
EncodingForm eFormEngineIn = EncodingForm.Unspecified, eFormEngineOut = EncodingForm.Unspecified;
PreConvert
(
eInEncodingForm, // [in] form in the BSTR
ref eFormEngineIn, // [out] form the conversion engine wants, etc.
eOutEncodingForm,
ref eFormEngineOut,
ref eNormalizeOutput,
bForward
);
// get enough space for us to normalize the input data (6x ought to be enough)
int nBufSize = sInput.Length * 6;
byte[] abyInBuffer = new byte[nBufSize];
fixed(byte *lpInBuffer = abyInBuffer)
{
// use a helper class to normalize the data to the format needed by the engine
Util.DebugWriteLine(className, "Calling GetBytes");
ECNormalizeData.GetBytes(sInput, ciInput, eInEncodingForm,
((bForward) ? CodePageInput : CodePageOutput), eFormEngineIn, lpInBuffer,
ref nBufSize, ref m_bDebugDisplayMode);
#if DEBUG && __MonoCS__
byte[] baOut = new byte[nBufSize];
ECNormalizeData.ByteStarToByteArr(lpInBuffer, nBufSize, baOut);
Util.DebugWriteLine(className, Util.getDisplayBytes("Input Bytes", baOut));
#endif
// get some space for the converter to fill with, but since this is allocated
// on the stack, don't muck around; get 10000 bytes for it.
int nOutLen = Math.Max(10000, nBufSize * 6);
byte[] abyOutBuffer = new byte[nOutLen];
fixed(byte *lpOutBuffer = abyOutBuffer)
{
lpOutBuffer[0] = lpOutBuffer[1] = lpOutBuffer[2] = lpOutBuffer[3] = 0;
// call the wrapper sub-classes' DoConvert to let them do it.
Util.DebugWriteLine(className, "Calling DoConvert");
DoConvert(lpInBuffer, nBufSize, lpOutBuffer, ref nOutLen);
#if DEBUG && __MonoCS__
Util.DebugWriteLine(className, "Output length " + nOutLen.ToString());
byte[] baOut2 = new byte[nOutLen];
ECNormalizeData.ByteStarToByteArr(lpOutBuffer, nOutLen, baOut2);
Util.DebugWriteLine(className, Util.getDisplayBytes("Output In Bytes", baOut2));
Util.DebugWriteLine(className, "Got val '" + System.Text.Encoding.Unicode.GetString(baOut2) + "'");
#endif
string result = ECNormalizeData.GetString(lpOutBuffer, nOutLen, eOutEncodingForm,
((bForward) ? CodePageOutput : CodePageInput), eFormEngineOut, eNormalizeOutput,
out rciOutput, ref m_bDebugDisplayMode);
#if DEBUG && __MonoCS__
Util.DebugWriteLine(className, "normalized result '" + result + "'");
byte[] baResult = System.Text.Encoding.BigEndianUnicode.GetBytes(result);
Util.DebugWriteLine(className, Util.getDisplayBytes("Normalized Output in UTF16BE", baResult));
baResult = System.Text.Encoding.Unicode.GetBytes(result);
Util.DebugWriteLine(className, Util.getDisplayBytes("Normalized Output in UTF16LE", baResult));
baResult = System.Text.Encoding.UTF8.GetBytes(result);
Util.DebugWriteLine(className, Util.getDisplayBytes("Normalized Output In UTF8", baResult));
Util.DebugWriteLine(className, "Returning.");
#endif
return(result);
}
}
}
/// legacy data as a byte array as input, we need to treat it as a byte array.
/// </summary>
protected virtual unsafe string InternalConvertEx(EncodingForm eInEncodingForm,
byte[] baInput,
EncodingForm eOutEncodingForm,
NormalizeFlags eNormalizeOutput,
out int rciOutput,
bool bForward)
{
Util.DebugWriteLine(className, "(input bytes) BEGIN");
Util.DebugWriteLine(className,
"eEncFormIn " + eInEncodingForm.ToString() + ", " +
"eEncFormOut " + eOutEncodingForm.ToString());
if (baInput == null)
{
EncConverters.ThrowError(ErrStatus.IncompleteChar);
}
if (baInput.Length == 0)
{
rciOutput = 0;
return("");
}
// if the user hasn't specified, then take the default case for the ConversionType:
// if L/RHS == eLegacy, then LegacyString
// if L/RHS == eUnicode, then UTF16
CheckInitEncForms(bForward, ref eInEncodingForm, ref eOutEncodingForm);
// allow the converter engine's (and/or its COM wrapper) to do some preprocessing.
EncodingForm eFormEngineIn = EncodingForm.Unspecified, eFormEngineOut = EncodingForm.Unspecified;
PreConvert(
eInEncodingForm, // [in] form in the BSTR
ref eFormEngineIn, // [out] form the conversion engine wants, etc.
eOutEncodingForm,
ref eFormEngineOut,
ref eNormalizeOutput,
bForward);
int nBufSize = baInput.Length;
fixed(byte *lpInBuffer = baInput)
{
int nOutLen = Math.Max(10000, nBufSize * 6);
byte[] abyOutBuffer = new byte[nOutLen];
fixed(byte *lpOutBuffer = abyOutBuffer)
{
lpOutBuffer[0] = lpOutBuffer[1] = lpOutBuffer[2] = lpOutBuffer[3] = 0;
// call the wrapper sub-classes' DoConvert to let them do it.
Util.DebugWriteLine(className, "Calling DoConvert");
DoConvert(lpInBuffer, nBufSize, lpOutBuffer, ref nOutLen);
Util.DebugWriteLine(className, "Output length " + nOutLen.ToString());
byte[] baOut = new byte[nOutLen];
ECNormalizeData.ByteStarToByteArr(lpOutBuffer, nOutLen, baOut);
#if DEBUG
Util.DebugWriteLine(className, Util.getDisplayBytes("Output In Bytes", baOut));
Util.DebugWriteLine(className, "Got val '" + System.Text.Encoding.Unicode.GetString(baOut) + "'");
#endif
string result = ECNormalizeData.GetString(lpOutBuffer, nOutLen, eOutEncodingForm,
((bForward) ? CodePageOutput : CodePageInput), eFormEngineOut, eNormalizeOutput,
out rciOutput, ref m_bDebugDisplayMode);
#if DEBUG
Util.DebugWriteLine(className, "normalized result '" + result + "'");
byte[] baResult = System.Text.Encoding.BigEndianUnicode.GetBytes(result);
Util.DebugWriteLine(className, Util.getDisplayBytes("Normalized Output in UTF16BE", baResult));
baResult = System.Text.Encoding.Unicode.GetBytes(result);
Util.DebugWriteLine(className, Util.getDisplayBytes("Normalized Output in UTF16LE", baResult));
baResult = System.Text.Encoding.UTF8.GetBytes(result);
Util.DebugWriteLine(className, Util.getDisplayBytes("Normalized Output In UTF8", baResult));
Util.DebugWriteLine(className, "Returning.");
#endif
return(result);
}
}
}