I’m running the following code on the FEZ Panda II:
uint total = foo(); //Can return 0
uint highTotal = bar(); //Can also return 0
double p = (double)highTotal / (double)total; //Throws an exception when both are 0, instead of setting p to Double.NaN
Do I have to turn on the NaN behaviour?
I do not think this is implemented in NETMF. If you see the use then then this should be requested directly from Microsoft www.netmf.com
Why would this not be implemented? Isn’t NaN upon division by zero part of the IEEE floating point spec?
Why mightnt something be implemented in the micro framework? Usually to keep code size down, to give more useful functions the extra code space. Micro framework is a set of tradeoffs, some of them particular people would argue whether they’re correct; as Gus said you can go request that over on the official site.