https://bugs.kde.org/show_bug.cgi?id=448353
--- Comment #5 from michael <reeves...@gmail.com> ---
Just rewrote the offending section of code. The LF lines would not have
triggered the QTextStream::pos() so it does seem to be the culprit. Turns there
were odd EOL corner cases that the didn't get detected right
The new code should fix both issues.
case '\r':
if((FileOffset)lastOffset < mDataSize)
{
prevChar = curChar;
curChar = ts.read(1).unicode()[0];
if(curChar == '\n')
{
vOrigDataLineEndStyle.push_back(eLineEndStyleDos);
break;
}
//work around for lack of seek API in QTextStream
skipNextRead = true;
}
//old mac style ending.
vOrigDataLineEndStyle.push_back(eLineEndStyleUndefined);
break;
The speed now should only depend on the time for QTextStream::read.
--
You are receiving this mail because:
You are watching all bug changes.
