String TextCodecGtk::decode(const char* bytes, size_t length, bool flush, bool stopOnError, bool& sawError)
{
// Get a converter for the passed-in encoding.
if (m_iconvDecoder == reinterpret_cast<GIConv>(-1)) {
createIConvDecoder();
ASSERT(m_iconvDecoder != reinterpret_cast<GIConv>(-1));
if (m_iconvDecoder == reinterpret_cast<GIConv>(-1)) {
LOG_ERROR("Error creating IConv encoder even though encoding was in table.");
return String();
}
}
size_t countWritten, countRead, conversionLength;
const char* conversionBytes;
char* prefixedBytes = 0;
if (m_numBufferedBytes) {
conversionLength = length + m_numBufferedBytes;
prefixedBytes = static_cast<char*>(fastMalloc(conversionLength));
memcpy(prefixedBytes, m_bufferedBytes, m_numBufferedBytes);
memcpy(prefixedBytes + m_numBufferedBytes, bytes, length);
conversionBytes = prefixedBytes;
// all buffered bytes are consumed now
m_numBufferedBytes = 0;
} else {
// no previously buffered partial data,
// just convert the data that was passed in
conversionBytes = bytes;
conversionLength = length;
}
GOwnPtr<GError> err;
GOwnPtr<UChar> buffer;
buffer.outPtr() = reinterpret_cast<UChar*>(g_convert_with_iconv(conversionBytes, conversionLength, m_iconvDecoder, &countRead, &countWritten, &err.outPtr()));
if (err) {
LOG_ERROR("GIConv conversion error, Code %d: \"%s\"", err->code, err->message);
m_numBufferedBytes = 0; // reset state for subsequent calls to decode
fastFree(prefixedBytes);
sawError = true;
return String();
}
// Partial input at the end of the string may not result in an error being raised.
// From the gnome library documentation on g_convert_with_iconv:
// "Even if the conversion was successful, this may be less than len if there were partial characters at the end of the input."
// That's why we need to compare conversionLength against countRead
m_numBufferedBytes = conversionLength - countRead;
if (m_numBufferedBytes > 0) {
if (flush) {
LOG_ERROR("Partial bytes at end of input while flush requested.");
m_numBufferedBytes = 0; // reset state for subsequent calls to decode
fastFree(prefixedBytes);
sawError = true;
return String();
}
memcpy(m_bufferedBytes, conversionBytes + countRead, m_numBufferedBytes);
}
fastFree(prefixedBytes);
Vector<UChar> result;
result.append(buffer.get(), countWritten / sizeof(UChar));
return String::adopt(result);
}
String TextCodecGtk::decode(const char* bytes, size_t length, bool flush, bool stopOnError, bool& sawError)
{
// Get a converter for the passed-in encoding.
if (!m_iconvDecoder)
createIConvDecoder();
if (!m_iconvDecoder) {
LOG_ERROR("Error creating IConv encoder even though encoding was in table.");
return String();
}
Vector<UChar> result;
gsize bytesRead = 0;
gsize bytesWritten = 0;
const gchar* input = bytes;
gsize inputLength = length;
gchar buffer[ConversionBufferSize];
int flags = !length ? G_CONVERTER_INPUT_AT_END : G_CONVERTER_NO_FLAGS;
if (flush)
flags |= G_CONVERTER_FLUSH;
bool bufferWasFull = false;
char* prefixedBytes = 0;
if (m_numBufferedBytes) {
inputLength = length + m_numBufferedBytes;
prefixedBytes = static_cast<char*>(fastMalloc(inputLength));
memcpy(prefixedBytes, m_bufferedBytes, m_numBufferedBytes);
memcpy(prefixedBytes + m_numBufferedBytes, bytes, length);
input = prefixedBytes;
// all buffered bytes are consumed now
m_numBufferedBytes = 0;
}
do {
GOwnPtr<GError> error;
GConverterResult res = g_converter_convert(G_CONVERTER(m_iconvDecoder.get()),
input, inputLength,
buffer, sizeof(buffer),
static_cast<GConverterFlags>(flags),
&bytesRead, &bytesWritten,
&error.outPtr());
input += bytesRead;
inputLength -= bytesRead;
if (res == G_CONVERTER_ERROR) {
if (g_error_matches(error.get(), G_IO_ERROR, G_IO_ERROR_PARTIAL_INPUT)) {
// There is not enough input to fully determine what the conversion should produce,
// save it to a buffer to prepend it to the next input.
memcpy(m_bufferedBytes, input, inputLength);
m_numBufferedBytes = inputLength;
inputLength = 0;
} else if (g_error_matches(error.get(), G_IO_ERROR, G_IO_ERROR_NO_SPACE))
bufferWasFull = true;
else if (g_error_matches(error.get(), G_IO_ERROR, G_IO_ERROR_INVALID_DATA)) {
if (stopOnError)
sawError = true;
if (inputLength) {
// Ignore invalid character.
input += 1;
inputLength -= 1;
}
} else {
sawError = true;
LOG_ERROR("GIConv conversion error, Code %d: \"%s\"", error->code, error->message);
m_numBufferedBytes = 0; // Reset state for subsequent calls to decode.
fastFree(prefixedBytes);
return String();
}
}
result.append(reinterpret_cast<UChar*>(buffer), bytesWritten / sizeof(UChar));
} while ((inputLength || bufferWasFull) && !sawError);
fastFree(prefixedBytes);
return String::adopt(result);
}