411893 - num.toPrecision(undefined) !== num.toString()

Status: VERIFIED FIXED

(Core :: JavaScript Engine, defect)

Description

[Jeff Walden [:Waldo]]

Attachment: Patch

As it turns out, looking at the spec sez it makes the most sense to special-case this before calling the num_to workhorse.

Comment 1

[Jeff Walden [:Waldo]]

Attachment: Untested testcase

No idea where this goes in the test suite, and untested, but probably correct...

Comment 2

[Jeff Walden [:Waldo]]

Fixed on trunk.

Comment 3

[Bob Clary [:bc] (inactive)]

Comment on attachment 296619 [details]
Untested testcase

Go ahead and add yourself as the contributor.

...
>var gTestfile = 'toPrecision-undefined-check.js';

use the section and a numeric suffix for the file name for tests that are 
easily mapped to the ecma 262-3 spec. Normally I try to use a 2 digit suffix
but there is already an old test with a single digit, so keeping in tune with
that.

e.g. 15.7.4.7-2.js

the file should be ecma_3/Number/15.7.4.7-2.js

>//-----------------------------------------------------------------------------
>var BUGNUMBER     = "(none)";

use 411893 as the bugnumber

>var summary = "num.toPrecision(undefined) should equal num.toString()";
>var actual, expect;
>
>printBugNumber(BUGNUMBER);
>printStatus(summary);
>
>/**************
> * BEGIN TEST *
> **************/
>
>var failed = false;
>
>try
>{
>  var prec = 3.3.toPrecision(undefined);
>  var str  = 3.3.toString();

I'm a 'little' concerned that it might be possible someday for 3.3.toString
to fail in the same way as 3.3.toPrecision which this comparison wouldn't catch,
but that is probably just paranoia.

>  if (prec !== str)
>  {
>    throw "not equal!  " +
>          "3.3.toPrecision(undefined) === '" + prec + "', " +
>          "3.3.toString() === '" + str + "'";
>  }
>}
>catch (e)
>{
>  failed = e;
>}
>
>
>
only need one blank line here.

>expect = false;
>actual = failed;
>
>reportCompare(expect, actual, summary);

test passes current trunk as well as the branch. 

r+ with those nits picked. Thanks Waldo!

Comment 4

[Jeff Walden [:Waldo]]

(In reply to comment #3)
> >  var prec = 3.3.toPrecision(undefined);
> >  var str  = 3.3.toString();
> 
> I'm a 'little' concerned that it might be possible someday for 3.3.toString
> to fail in the same way as 3.3.toPrecision which this comparison wouldn't
> catch, but that is probably just paranoia.

That's not a problem, because the spec says that toPrecision with a first argument with value undefined is equivalent to ToString(this primitive number), which is what Number.prototype.toString is defined as in the spec with an undefined first argument.

Comment 5

[Bob Clary [:bc] (inactive)]

/cvsroot/mozilla/js/tests/ecma_3/Number/15.7.4.7-2.js
initial revision: 1.1

Comment 6

[Bob Clary [:bc] (inactive)]

v