QString KEncodingDetector::decodeWithBuffering(const char *data, int len)
{
#ifdef DECODE_DEBUG
kWarning() << "KEncodingDetector: decoding "<<len<<" bytes";
#endif
if (d->m_writtingHappened)
{
#ifdef DECODE_DEBUG
kWarning() << "KEncodingDetector: d->m_writtingHappened "<< d->m_codec->name();
#endif
processNull(const_cast<char *>(data),len);
return d->m_decoder->toUnicode(data, len);
}
else
{
if (d->m_bufferForDefferedEncDetection.isEmpty())
{
// If encoding detection produced something, and we either got to the body or
// actually saw the encoding explicitly, we're done.
if (analyze(data,len) && (d->m_seenBody || d->isExplicitlySpecifiedEncoding()))
{
#ifdef DECODE_DEBUG
kWarning() << "KEncodingDetector: m_writtingHappened first time "<< d->m_codec->name();
#endif
processNull(const_cast<char *>(data),len);
d->m_writtingHappened=true;
return d->m_decoder->toUnicode(data, len);
}
else
{
#ifdef DECODE_DEBUG
kWarning() << "KEncodingDetector: begin deffer";
#endif
d->m_bufferForDefferedEncDetection=data;
}
}
else
{
d->m_bufferForDefferedEncDetection+=data;
// As above, but also limit the buffer size. We must use the entire buffer here,
// since the boundaries might split the meta tag, etc.
bool detected = analyze(d->m_bufferForDefferedEncDetection.constData(), d->m_bufferForDefferedEncDetection.length());
if ((detected && (d->m_seenBody || d->isExplicitlySpecifiedEncoding())) ||
d->m_bufferForDefferedEncDetection.length() > MAX_BUFFER)
{
d->m_writtingHappened=true;
d->m_bufferForDefferedEncDetection.replace('\0',' ');
QString result(d->m_decoder->toUnicode(d->m_bufferForDefferedEncDetection));
d->m_bufferForDefferedEncDetection.clear();
#ifdef DECODE_DEBUG
kWarning() << "KEncodingDetector: m_writtingHappened in the middle " << d->m_codec->name();
#endif
return result;
}
}
}
return QString();
}
QString KEncodingDetector::decode(const QByteArray &data)
{
processNull(const_cast<char *>(data.data()), data.size());
if (!d->m_analyzeCalled) {
analyze(data.data(), data.size());
d->m_analyzeCalled = true;
}
return d->m_decoder->toUnicode(data);
}
QString KEncodingDetector::decode(const char *data, int len)
{
processNull(const_cast<char *>(data), len);
if (!d->m_analyzeCalled) {
analyze(data, len);
d->m_analyzeCalled = true;
}
return d->m_decoder->toUnicode(data, len);
}
/**
* open file, read first chunk of data, detect eol
*/
bool open ()
{
if (m_file.open (IO_ReadOnly))
{
int c = m_file.readBlock (m_buffer.data(), m_buffer.size());
if (c > 0)
{
// fix utf16 LE, stolen from tdehtml ;)
if ((c >= 2) && (m_codec->mibEnum() == 1000) && (m_buffer[1] == 0x00))
{
// utf16LE, we need to put the decoder in LE mode
char reverseUtf16[3] = {0xFF, 0xFE, 0x00};
m_decoder->toUnicode(reverseUtf16, 2);
}
processNull (c);
m_text = m_decoder->toUnicode (m_buffer, c);
}
m_eof = (c == -1) || (c == 0) || (m_text.length() == 0) || m_file.atEnd();
for (uint i=0; i < m_text.length(); i++)
{
if (m_text[i] == '\n')
{
m_eol = KateDocumentConfig::eolUnix;
break;
}
else if ((m_text[i] == '\r'))
{
if (((i+1) < m_text.length()) && (m_text[i+1] == '\n'))
{
m_eol = KateDocumentConfig::eolDos;
break;
}
else
{
m_eol = KateDocumentConfig::eolMac;
break;
}
}
}
return true;
}
return false;
}
// read a line, return length + offset in unicode data
void readLine (uint &offset, uint &length)
{
length = 0;
offset = 0;
while (m_position <= m_text.length())
{
if (m_position == m_text.length())
{
// try to load more text if something is around
if (!m_eof)
{
int c = m_file.readBlock (m_buffer.data(), m_buffer.size());
uint readString = 0;
if (c > 0)
{
processNull (c);
TQString str (m_decoder->toUnicode (m_buffer, c));
readString = str.length();
m_text = m_text.mid (m_lastLineStart, m_position-m_lastLineStart)
+ str;
}
else
m_text = m_text.mid (m_lastLineStart, m_position-m_lastLineStart);
// is file completly read ?
m_eof = (c == -1) || (c == 0) || (readString == 0) || m_file.atEnd();
// recalc current pos and last pos
m_position -= m_lastLineStart;
m_lastLineStart = 0;
}
// oh oh, end of file, escape !
if (m_eof && (m_position == m_text.length()))
{
lastWasEndOfLine = false;
// line data
offset = m_lastLineStart;
length = m_position-m_lastLineStart;
m_lastLineStart = m_position;
return;
}
}
if (m_text[m_position] == '\n')
{
lastWasEndOfLine = true;
if (lastWasR)
{
m_lastLineStart++;
lastWasR = false;
}
else
{
// line data
offset = m_lastLineStart;
length = m_position-m_lastLineStart;
m_lastLineStart = m_position+1;
m_position++;
return;
}
}
else if (m_text[m_position] == '\r')
{
lastWasEndOfLine = true;
lastWasR = true;
// line data
offset = m_lastLineStart;
length = m_position-m_lastLineStart;
m_lastLineStart = m_position+1;
m_position++;
return;
}
else
{
lastWasEndOfLine = false;
lastWasR = false;
}
m_position++;
}
}
