Eval and unresolved variables in chapter 8 "types" tests.

Many of the Test are not runnable in ECMAScript 5. 

1) I eliminate the unresolved variables

2) I, also,  modify test S8.4_A9_T2; case 6 where str=""; and strObj=new String(); are
equal see 15.5.1.1;

[christian.plesner.hansen, 2010-07-19 06:06:48]

I have some comments but the codereview tool seems to be broken.

Some of these changes assume that a primitive string is strict equals to its
corresponding wrapper.  Is that a new thing in ES5 because I'm pretty sure
that's not correct in ES3.

[pedro.delgallego, 2010-07-19 08:01:53]

On 2010/07/19 06:06:48, christian.plesner.hansen wrote:
> I have some comments but the codereview tool seems to be broken.

Yes, I have been trying to figure out what I did wrong. The first time I
uploaded  it was because the Base URL was wrong but this time I choose
sputniktest and the base url is "http://sputniktests.googlecode.com/svn/trunk".
any suggestion? 


> Some of these changes assume that a primitive string is strict equals to its
> corresponding wrapper.  Is that a new thing in ES5 because I'm pretty sure
> that's not correct in ES3.

You are right. I misunderstand this paragraph of the spec. 

15.5.1 The String Constructor Called as a Function When String is called as a
function rather than as a constructor, it performs a type conversion.

15.5.1.1 String ( [ value ] ) Returns a String value (not a String object)
computed by ToString(value). If value is not supplied, the empty String "" is
returned.

Now I see that is referring String called as a function rather than as a
constructor. 

Should I upload a new Patch or add a new Patch Set on this one?

[christian.plesner.hansen, 2010-08-08 02:53:35]

Please add a new patch set, and be sure to publish+mail so we know when it's
happened.

[pedro.delgallego, 2010-08-09 13:39:09]

Revert to base the worng tests + delete and put to non writable attributes
should fail in strict

[pedro.delgallego, 2010-08-09 13:42:36]

> Revert to base the wrong tests 

I revert the changes in those test that assume that primitive string was  strict
equals to its wrapper.

> delete and put to non writable attributes
> should fail in strict

I also add some changes : 
Test that tries to delete a property that has the attribute {
[[Configurable]]:false } should fail. (11.4.1)

Test that tries to write in a property with the attribute writable equal to
false should fail.

[pedro.delgallego, 2010-08-09 13:43:54]

I have updated the Issue 1743046: Eval and unresolved variables in chapter 8
"types" tests.

I revert the changes in those test that assume that primitive string was  strict
equals to its wrapper.

I also add some changes : 

- Test that tries to delete a property that has the attribute
{[[Configurable]]:false } should fail. (11.4.1)

- Test that tries to write in a property with the attribute writable equal to
false should fail.

[christian.plesner.hansen, 2010-08-09 23:35:05]

> 8.3_The_Boolean_Type/S8.3_A1_T1.js
> and others

Why is it necessary to add 'var'?  Don't these properties already exist on the
global object?

Otherwise LGTM.

[Jasvir, 2010-08-19 21:51:30]

LGTM.

Christian, are we looking at different files?  All the properties that I see
"var" decls added to, do not appear on the global object, even in nonstrict
mode.  Eg. in 8.3, the var decl I see being added is for "x".

On 2010/08/09 23:35:05, christian.plesner.hansen wrote:
> > 8.3_The_Boolean_Type/S8.3_A1_T1.js
> > and others
> 
> Why is it necessary to add 'var'?  Don't these properties already exist on the
> global object?
> 
> Otherwise LGTM.

[christian.plesner.hansen, 2010-08-25 04:26:03]

Sorry, I listed the wrong test.  If you look at

8.1_The_Undefined_Type/S8.1_A3.js

that adds a 'var' for the property 'undefined'.