Subj : Re: Proper behavior of uneval(-0.0) ?
To   : brendan@meer.net
From : Igor Bukanov <igor@fastmail.fm>
Date : Wed Nov 12 2003 10:31 pm

Brendan Eich wrote:
> Igor Bukanov wrote:
> 
>>
>> Currently in SM uneval(-0.0) gives "0". Should it be "-0.0" instead?
> 
> 
> I don't think so, based on ECMA-262 Edition 3 9.8.1 step 2.  On the 
> other hand, uneval is an extension.  We could make it have a special 
> case here.  I'm hesitant to do that, on general KISS grounds.  Is there 
> a compelling use-case?

No: I have found this while implementing uneval in Rhino. I was 
surprised that SM goes great deals to implement toSource() for functions 
in a very different way compared with toString() so given

function f(i)
{
	while (i < 10) { ++i; }
	return 1;
}

f.toString() gives

function f(i) {
     while (i < 10) {
         ++i;
     }
     return 1;
}

while uneval(f) produces

function f(i) {while (i < 10) {++i;}return 1;}

with no semantic difference between forms while uneval(-0) would miss 
the semantically significant minus. It is really a pity that Ecma 
defines that String(-0) should give 0.

Regards, Igor

.