Bernhard Roider wrote:
Abdelrazak Younes schrieb:
The idea is that when the cursor is at the end of a text line, if the
next (or previous) line is shorter you don't want to remember that
position because the cursor will jump two lines below instead where
the x-target is met.
I cannot reproduce it when i remove the condition as done in the patch i
submitted. So either that bug has gone since that or i tested the wrong
thing. Can you give a detailed example where that happened?
btw. in current trunk when you press up or down at the eond of a line
the cursor jumps to the horizontal position where x-target was updated.
Yes. The behaviour is definitely changed since when the test was
committed. Current behaviour is definitely wrong, I guess your patch is
good then.
Just an interesting point: based on this thread, this may be related
to the bug of skipping lines on MAC (bug 3544) which Bennett asked
about earlier tonight...?
Maybe but if so, I don't see why it is Mac specific.
I also don't think so. It may be different (buggy?) behavior of stl or
boost. Compiler optimization does not seem to be the reason (look at
Bennetts post at the bug page).
Or maybe some difference in the font metrics on the Mac platform...
Abdel.
