.NET Framework - C# definition for NAN (in Reflector) is not correct?

Asked By josephbubb

20-Jun-08 03:52 AM

Greetings,

Question about how the NaN constant is defined in .NET.

When I look at the C# definition for the System.Double NaN definition
in Roeder's Reflector, it looks like this:

public const double NaN = (double) 1.0 / (double) 0.0;

However, according to the .NET documentation (and the IEEE standard),
the actual definition is the result of dividing zero by zero (0/0).
Dividing 1/0 is positive infinity.

Interestingly, when I switch to the any other decompilation languages
(ex. Visual Basic, or Managed C++), the value of NaN is not
specifically defined as a division, but instead uses a constant.  For
example, Visual Basic looks like this:

Public Const NaN As Double = NaN



Is this simply a bug in Reflector for C#, or am I missing something?

Thanks in advance.

By the way, I'm using Reflector version 5.1.2.0 and pointing to
the .Net framework version 2.0.50727.1433.

18-Jun-08 04:46 PM

Looks like a reflector bug to me.

--
Web site: http://www.pobox.com/~skeet
Blog: http://www.msmvps.com/jon.skeet
C# in Depth: http://csharpindepth.com