Re: assertEquals("NaN", x, y, 0.0) where x=y=Double.NaN

The test I have that's doing what I want looks like this:

    public void assertEqualsNaN(String str, double exp, double val, double
tol){
        if(Double.isNaN(exp))
            assertTrue(str, Double.isNaN(val));
        else
            assertEquals(str, exp, val, tol);
    }

It's only problem is that I'm adding it to every TestCase thats using it and
when it fails it gives me the code line of the assert statments here and not
the line in the test itself.

(I could encapsulated JUnits TestCase into my own MortensTestCase and add
the assertEqualsNaN there but I haven't. All my test inherit from TestCase
directly.)

Morten

> -----Original Message-----
> From: kentlbeck [mailto:kentbeck@xxxxxxx]
> Sent: 24 June 2002 20:33
> To: junit@xxxxxxxxxxxxxxx
> Subject: [junit] Re: assertEquals("NaN", x, y, 0.0) where x=y=Double.NaN
>
>
> --- In junit@xxxx, tumb@xxxx wrote:
> >
> > The way things currently work is the way the double class works.
> > This 'patch' would seem to violate the basic contract of double.
> >
> > Does assertEquals return true when both objects are null?
>
> From junit.tests.framework.AssertTest
>       public void testAssertEqualsNull() {
>               assertEquals(null, null);
>       }
>
> >
> > I don't know your procedure for deciding on changes or fixes. Is
> there
> > some way to discuss and determine whether or not this fix is proper
> or
> > that the current behavior is better?
>
> Submit a test (like the one above) and we can discuss it. Without the
> test it's hard to know precisely what you mean.
>
> Kent
>
>
>
> To unsubscribe from this group, send an email to:
> junit-unsubscribe@xxxxxxxxxxxxxxx
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>