Imported From: https://issues.scala-lang.org/browse/SI-1708?orig=1
Reporter: @SethTisue

@dragos said:
Should be easy to implement, here's a short test (taken from http://java.sun.com/developer/JDCTechTips/2001/tt0410.html#using):

public strictfp class FPTest {
    public static void main(String[] args) {
        double d = 8e+307;
        System.out.println(4.0 * d * 0.5);
        System.out.println(2.0 * d);
    }
}

There seems to be no equivalent on .NET, so this will have an effect only on jvm targets.

@SethTisue said:
link is now dead, but archived here: http://web.archive.org/web/20080702220525/http://java.sun.com/developer/JDCTechTips/2001/tt0410.html#using

@dragos said:
I need someone else to take care of this.

@paulp said:
Every supposed @Strictfp illustrating java example I try on OSX java6 or openjdk, I get the same results either way. Not likely I can implement this without being able to see something happen. Requesting assistance: closing NEEDINFO, but anyone who wants to implement it feel free to reopen.

@SethTisue said:
the way we would "see something happen" if Scala had strictfp support is by looking in the class files output by the compiler and seeing whether the strictfp bit is being set, the same way the Java compiler sets it. it's not necessary to actually test whether you get different answers from math operations; we can assume the JVM will do the right thing there.

I think this ticket should remain open as a requested enhancement. I think assigning it to scala_community would be reasonable.

@paulp said:
Replying to [comment:6 SethTisue]:

the way we would "see something happen" if Scala had strictfp support is by looking in the class files output by the compiler and seeing whether the strictfp bit is being set, the same way the Java compiler sets it. it's not necessary to actually test whether you get different answers from math operations; we can assume the JVM will do the right thing there.

I think this ticket should remain open as a requested enhancement. I think assigning it to scala_community would be reasonable.

That's fine if the ticket feels slighted, but you're going the wrong way: "closed with needinfo status" means "I will work on this if I am supplied with info" whereas "open and assigned to scala_community" means "nobody is likely to work on this."

@paulp said:
(In r23427) Created @Strictfp annotation in scala.annotation. If you prefer
your fp with no tolerance for tomfoolery, this is the commit for you.
Closes #1708, review by community.

@SethTisue said:
awesome. I won't have time to look at this until early December, but I'll definitely look then. (I'm putting it my calendar.)

@SethTisue said:
I tried this out, using javap to verify the results, and it seems to be working great. thanks so much.