Le 29. 09. 10 12:10, Gervase Markham a écrit : > So, just to be clear: you are agreeing with me that the tip behaviour > (which is inconsistent) is broken, and you are also saying that it's a > regression? Yes