Abdelrazak Younes wrote:
Georg Baum wrote:
The screenshot looks different than here (but the patch did not change
anything drawing related). Obviously your version is correct, while
mine is not. I have no idea why,
I think I know why, your settings use the internal bidi algorithm
(Paragraph::transformChar) while mine do not. In my case and for Arabic,
it is also a font issue. Inside RowPainter::paintFromPos() there is this
test for arabic:
bool const arabic = lang == "arabic" &&
(lyxrc.font_norm_type == LyXRC::ISO_8859_6_8 ||
lyxrc.font_norm_type == LyXRC::ISO_10646_1);
(lang == "arabic" ) returns true but none of the two following condition
are true. So this means that "arabic" is false at this point and
Paragraph::transformChar() is not used later on. So, in my case, I guess
Qt draws correctly the passed string. In your case I guess the string is
reordered...
In any case, Paragraph::transformChar() is completely wrong in an
unicode context.
Abdel.
But Abdel, what's causing the different behavior between you and Georg?
Is it a matter of which font is being used as the screen font? What is
font_norm_type? And we're seeing the problem with hebrew, not arabic...
