> I'm impressed by your analysis, Vincent. Yes, math can impress.. :).
> Also, this is the first bug analysis I saw that was > written using mathed. You asked for it. > As a poor human scientist, I can not follow your argumentation > in detail That's my fault. I didn't want to write a full four-page paper about an issue that is created by humans and which is already solved. > (the only thing that struck me is that the condition > k/2 = k/2 should always be true, no?) Note the brackets: floor(k/2) = ceil(k/2), just to show where Qt went wrong. > >However, could you limit the change to the affected Qt versions? > The fix is targeted to 4.5.x, but not yet for 4.5.3. Thus, I'm not sure yet which versions will be affected. Vincent
