I'm running the below file in FP10 using SDK 3.4. <?xml version="1.0" encoding="utf-8"?> <mx:Application xmlns:mx="http://www.adobe.com/2006/mxml"; layout="absolute" applicationComplete="init()"> <mx:Script> <![CDATA[ var a:Array = [28700000.0,-27300000.0,-27000000.0,8700000.0,-24500000.0] var b:Array = [287000000.0,-273000000.0,-270000000.0,87000000.0,-245000000.0] public function init() { trace(a) a.sort(Array.NUMERIC) trace(a) trace(b) b.sort(Array.NUMERIC) trace(b) } ]]> </mx:Script> </mx:Application>
The array b is not sorted the second time it is traced. The output for the second trace of b is: -273000000,-270000000,87000000,-245000000,287000000 Array a (which is simply all the elements in b divided by 10) does get successfully sorted. With numbers of the order of magnitude that they are in b then weird things happen with larger arrays. They get mostly sorted except for the fact that there's a chunk of positive numbers where negative numbers should be or visa-versa, this errant chunk is correctly sorted except for the fact that it is in totally the wrong place in the array. I've logged this as a bug in JIRA but I just want to know if anyone else has seen this behaviour? We first noticed something was up when we were sorting dates-as-milliseconds and all kinds of anomalous behaviour occurred). -- Alistair Hutton
