Math.PI accuracy

Try comparing it to a 64 bit float (the "d" at the end)

if(Math.PI==3.14159265358979323846d) {

Sys.println("they are the same");

}

I think what you're seeing is that the is the format() in your println is displaying it incorrectly.

Thanks, Jim.

I'm converting Java to Monkey C, and I'm tracing variable values through different functions.

I'm therefore comparing the output - this has just been made a lot more complicated.

If there any method I can use to print the value (other than formatting)?

I've tried assigning to a var and forcing that a double, but I still get the "wrong" version of PI!

Jim, I used your idea, and can confirm I get an incorrect value for PI:

if(Math.PI!=3.14159265358979323846d) {
           System.println("PI is incorrect");
}

If you do this, do you get the same?

If I do this though,

if(Math.PI!=3.14159265358979323846d) {
           System.println("PI is incorrect");
Sys.println(""+(Math.PI-3.14159265358979323846d));	          
} 

I get:

PI is incorrect
0.000000

So it looks in some things it uses it as a 32bit float

Spot on Jim.

//Math.PI==3.14159265358979323846 <-Correct
//Math.PI==3.14159265358979323846f <-Correct
//Math.PI==3.14159265358979323846d <- Incorrect

if(Math.PI==3.14159265358979323846) {
    System.println("PI is Correct");
} else {
   System.println("PI is Incorrect");
}

on line 5 you are comparing against a 32 bit float, and on lines 1 and 2, you are assigning a 32 bit float.

I think this is one for Travis.ConnectIQ!

Yes - Seems to be OK when 32 bit (as you said). Who is Travis?

Travis is one of the Garmin CIQ folks.  Look around the forum, you'll see him quite a bit.  People who's names end in ".Connectiq" are Garmin

I've PM'd him. Thanks for your help, Jim.

I'd already "at'd" him so he'll see this thread.  Answers in a thread help others that have the same question.