Subj : Re: JSVAL_IS_NUMBER(NaN) == TRUE (!?)
To   : Rob Swindell
From : Brendan Eich
Date : Fri Feb 18 2005 02:55 pm

Rob Swindell wrote:
> SpiderMonkey oddity:
> 
> I found it odd that if jsval is NaN, a test of JSVAL_IS_NUMBER(jsval) would
> evaluate as TRUE.


Sure, it's of "number" type (see also typeof(0/0)).


> And my native methods don't consider NaN a valid number either, so I needed a
> valid test for is number, but non NaN, so I came up with:
> 
> #include <jsnum.h>      /* JSDOUBLE_IS_NaN() */
> 
> JSBool
> jsval_isNaN(JSContext *cx, jsval v)
> {
>     jsdouble d;
> 
>     if(JSVAL_IS_DOUBLE(v)) {
>             if (!JS_ValueToNumber(cx, v, &d))


No need to convert v to number if it's already a double (which is a 
subtype of number).


>                         return JS_FALSE;
>                 if(JSDOUBLE_IS_NaN(d))
>                         return JS_TRUE;
>         }
>         return JS_FALSE;
> }
> 
> // Replacement for JSVAL_IS_NUMBER():
> #define JSVAL_IS_NUM(cx,v)      (JSVAL_IS_NUMBER(v) && !jsval_isNaN(cx,v))
> 
> I'm surprised there's not something like this already in the jsapi. <shrug>


Yeah, it hasn't come up till now.  Most native methods should let NaNs 
flow through and produce NaN results.  Why not yours?

/be

.