void TextBuffer::setTextCodec (QTextCodec *codec)
{
m_textCodec = codec;
// enforce bom for some encodings
int mib = m_textCodec->mibEnum ();
if (mib == 1013 || mib == 1014 || mib == 1015) // utf16
setGenerateByteOrderMark (true);
if (mib == 1017 || mib == 1018 || mib == 1019) // utf32
setGenerateByteOrderMark (true);
}
bool TextBuffer::load (const QString &filename, bool &encodingErrors)
{
// fallback codec must exist
Q_ASSERT (m_fallbackTextCodec);
// codec must be set!
Q_ASSERT (m_textCodec);
/**
* first: clear buffer in any case!
*/
clear ();
/**
* check if this is a normal file or not, else exit
*/
KDE_struct_stat sbuf;
if (KDE::stat(filename, &sbuf) != 0 || !S_ISREG(sbuf.st_mode))
return false;
/**
* construct the file loader for the given file, with correct prober type
*/
Kate::TextLoader file (filename, m_encodingProberType);
/**
* triple play, maximal three loading rounds
* 0) use the given encoding, be done, if no encoding errors happen
* 1) use BOM to decided if unicode or if that fails, use encoding prober, if no encoding errors happen, be done
* 2) use fallback encoding, be done, if no encoding errors happen
* 3) use again given encoding, be done in any case
*/
for (int i = 0; i < 4; ++i) {
/**
* kill all blocks beside first one
*/
for (int b = 1; b < m_blocks.size(); ++b) {
TextBlock* block = m_blocks.at(b);
block->m_lines.clear ();
delete block;
}
m_blocks.resize (1);
/**
* remove lines in first block
*/
m_blocks.last()->m_lines.clear ();
m_lines = 0;
/**
* try to open file, with given encoding
* in round 0 + 3 use the given encoding from user
* in round 1 use 0, to trigger detection
* in round 2 use fallback
*/
QTextCodec *codec = m_textCodec;
if (i == 1)
codec = 0;
else if (i == 2)
codec = m_fallbackTextCodec;
if (!file.open (codec)) {
// create one dummy textline, in any case
m_blocks.last()->appendLine (TextLine (new TextLineData()));
m_lines++;
return false;
}
// read in all lines...
encodingErrors = false;
while ( !file.eof() )
{
// read line
int offset = 0, length = 0;
bool currentError = !file.readLine (offset, length);
encodingErrors = encodingErrors || currentError;
// bail out on encoding error, if not last round!
if (encodingErrors && i < 3) {
kDebug (13020) << "Failed try to load file" << filename << "with codec" <<
(file.textCodec() ? file.textCodec()->name() : "(null)");
break;
}
// get unicode data for this line
const QChar *unicodeData = file.unicode () + offset;
// construct new text line with content from file
TextLine textLine = TextLine (new TextLineData(QString (unicodeData, length)));
// ensure blocks aren't too large
if (m_blocks.last()->lines() >= m_blockSize)
m_blocks.append (new TextBlock (this, m_blocks.last()->startLine() + m_blocks.last()->lines()));
m_blocks.last()->appendLine (textLine);
m_lines++;
}
// if no encoding error, break out of reading loop
if (!encodingErrors) {
// remember used codec
m_textCodec = file.textCodec ();
break;
}
}
// remember if BOM was found
if (file.byteOrderMarkFound ())
setGenerateByteOrderMark (true);
// remember eol mode, if any found in file
if (file.eol() != eolUnknown)
setEndOfLineMode (file.eol());
// remember mime type for filter device
m_mimeTypeForFilterDev = file.mimeTypeForFilterDev ();
// assert that one line is there!
Q_ASSERT (m_lines > 0);
// report CODEC + ERRORS
kDebug (13020) << "Loaded file " << filename << "with codec" << m_textCodec->name()
<< (encodingErrors ? "with" : "without") << "encoding errors";
// report BOM
kDebug (13020) << (file.byteOrderMarkFound () ? "Found" : "Didn't find") << "byte order mark";
// report filter device mime-type
kDebug (13020) << "used filter device for mime-type" << m_mimeTypeForFilterDev;
// emit success
emit loaded (filename, encodingErrors);
// file loading worked, modulo encoding problems
return true;
}
