478905 - E4X: XML parsing of string with Unicode character U2028 fails with "illegal XML character" message

Status: RESOLVED WONTFIX

(Core :: JavaScript Engine, defect)

Description

[Martin Honnen]

When using 
  new XML(xmlString)
with certain xmlStrings containing the Unicode character U2028 Spidermonkey fails to parse the string and gives an "illegal XML character" error. The same string however can be parsed just fine with DOMParser for instance so the error message is not correct. The problem does not occur with any XML string containing that character but only with certain strings.

Tested with Firefox 3.0.6 as well as with a current nightly trunk Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090217 Minefield/3.2a1pre

I will upload a test case next.

Comment 1

[Martin Honnen]

Attachment: JavaScript test case

Comment 2

[Martin Honnen]

Attachment: HTML test case to be loaded in the browser

The HTML test case references the JavaScript test case https://bugzilla.mozilla.org/attachment.cgi?id=362736 and shows the error in Mozilla's error console.
Furthermore it uses DOMParser to parse the string that new XML() fails to parse to show that it is well-formed XML.

Comment 3

[Brendan Eich [:brendan]]

\u2028 is LINE_SEPARATOR (2029 is PARA_SEPARATOR). ECMA-262 Edition 3 defines these as line break characters. I wonder if that's relevant; the E4X XML lexer is integrated with the JS lexer.

/be

Comment 4

[Leni]

\u2029 triggers the bug too.

Comment 5

[Nicholas Nethercote [inactive]]

e4x support has been removed (bug 788293).