271545 - XML() does not create empty text nodes

Status: RESOLVED FIXED

(Rhino Graveyard :: E4X, defect)

Description

[Igor Bukanov]

Werner Sharp wrote:
> 
> In the appendChild behavior below, it looks like Rhino is treating "new
> XML()" different than "new XML("")".  The same issue appears when doing...
> 
> x = new XML();
> x.a = "foo";
> print (x.toXMLString());
> x = new XML("");
> x.a = "foo";
> print (x.toXMLString());
> 
> The first case creates "<a>foo</a>" while the second case does not.

Comment 1

[Igor Bukanov]

Here is a relevant section of ECMA 357:

13.4.2 The XML Constructor
Syntax
       new XML ( [ value ] )
...
Semantics
When the XML constructor is called with no arguments or a single argument value,
the following steps are taken:
    1. If value is null, undefined or not supplied, let value be the empty string
...

Comment 2

[Igor Bukanov]

Changing the title: the actual bug is that while XML() creates true leaf empty
text object which should ignore attempts to add xml child to it, the resulting
XML("") does allow child insertion.

Note that XML(nonEmptyStringWithoutXMLMarkup) behaves properly and creates text
node which ignore any attempt to add a child to it.

Comment 3

[Igor Bukanov]

Changing title one more time:

The bug is that while XML("") does create empty text node, XML() creates
non-leaf object.

Comment 4

[Igor Bukanov]

I added e4x/Regress/regress-271545.js to the test suite which currently generates:

Bug Number 271545
STATUS: XML("") should create empty text node
Failure messages were:
FAILED!: Section 1 of test -
FAILED!: Expected value:
FAILED!: 
FAILED!: Actual value:
FAILED!: <a>foo</a>
FAILED!: 

Comment 5

[Igor Bukanov]

Attachment: Fix: make XML() equivalent to XML("")

Comment 6

[Igor Bukanov]

I committed the fix