On Sat, 09 Feb 2008 12:38:16 +0100, Charles McCathieNevile <[EMAIL PROTECTED]> wrote:
I suggest we declare this test invalid, and either replace it and declare it valid only after such date as it is fixed, or simply add a new test that gets the correct file.
Fixed.
The test checks one aspect of step 3 in the part of section 4 that deals with "XML response entity body", which requires that content which fails XML well-formedness should have a responseXML of null - in this case the responseXML should be null because the file loaded contains an unescaped ampersand.
Right.
I could not find any issue raised and resolved for this in the issue tracker, but it is mentioned in a thread at http://lists.w3.org/Archives/Public/public-webapi/2007Feb/thread.html#msg088 followed by an apparent summary at http://lists.w3.org/Archives/Public/public-webapi/2007Mar/0090 (I am assuming, since the summary there matches what I understand the document to say on this point).
That thread is is about responseText.
If we make the change, what is the conformance status? Given that http://tc.labs.opera.com/apis/XMLHttpRequest/responseXML/006.htm (which also tests XML well-formedness errors giving null responseXML) is not one of the tests failed by a lot of browsers, I suspect the issue will go away, but I haven't actually run the tests on other browsers yet. So I suggest there is no real issue here, just an incorrect test.
Some browsers still fail, but given that they all treat invalid XML documents in different ways that was sort of expected.
-- Anne van Kesteren <http://annevankesteren.nl/> <http://www.opera.com/>
