>>>>> "Abdelrazak" == Abdelrazak Younes <[EMAIL PROTECTED]> writes:
Abdelrazak> Jean-Marc Lasgouttes wrote: >>>>>>> "Richard" == Richard Heck <[EMAIL PROTECTED]> writes: >> Richard> I suppose that's possible, but the problem seems to me to be Richard> the opposite. The way the code was, editXY was being called Richard> in BufferView::workAreaDispatch for the sole purpose of Richard> finding the deepest inset over which the mouse pointer is Richard> located. But editXY does a lot else---it resets the cursor Richard> and calls LyXText::setCurrentFont, in particular. So it's not Richard> so much the lack of such a call as the presence of one in Richard> this method that was causing the problem. >> What I do not understand is why this bug is not present in 1.4 (or >> is it?). Abdelrazak> Because mouse hover is new to 1.5. I see, thanks. Another example of why eye candy can be expensive (when it leads to rewriting inner parts of the kernel). Abdelrazak> Here is an alternative patch that do the correct thing Abdelrazak> without touching editXY(). In the future (i.e. post 1.5.0) Abdelrazak> we could clean up editXY() to use that instead of Abdelrazak> checkInsetHit(). I will commit soon. I see nothing worong with the code, except that it should be called getCoveringInset and that 'inset' and 'cell_number' should probably be const. But we are adding yet another way to find insets, and the previous patches that lead us to this situation did not look bad either. JMarc