RE: prototype DOM TS Level [1|2] [Core|HTML]

Robert Clary wrote:

 > Rick,
 >
 > Are you running the version from CVS?
 >
 > I combined and compared our two different results and have attached 
as combined-results.txt.
 >
 > I noticed that the signature of assertURIEquals in DOMTestCase.js is:
 >
 > 
assertURIEquals(assertID,scheme,path,host,file,query,fragment,isAbsolute,actual)
 >
 > and the call in my HTMLAnchorElement04.html is
 >
 > 
assertURIEquals("hrefLink",null,null,null,"submit.gif",null,null,null,null,vhref); 



The DOMTestCase.js in CVS has ten parameters, one more (name) than in 
your DOMTestCase.js, which is consistent with the transforms and 
generated code.  The "name" attribute on assertURI was added sometime 
before the last modification of HTMLDocument04.xml on 30 June 2002.

I won't have any time to fully digest this till the weekend, but here 
are a few quick comments/questions.

Is the essential innovation the use of the query string used to launch 
JsUnit to configure the test framework?  Are there any other novel aspects?

Would there be any issues using a select box instead of radio buttons to 
select the suite or test to run?

The recently generated tests have had an <iframe> element in the test 
html file for each document loaded by the test.  As far as I can tell, 
your "native" loading uses the one iframe in main-loader.html and would 
not support tests that use multiple documents such as 
elementwrongdocumenterr.  Is there an issue with distinct <iframe>'s in 
the test document?

You mentioned the hc_* variant tests.  These were added after first 
release of DOM L1 Core to provide some coverage of the Core for HTML 
processors since the original NIST tests required XML documents.  The 
tests are a direct transliteration of the NIST types making tag and 
attribute name substitutions to make the test documents valid (but 
non-sensical) XHTML and HTML.  XML processors show the same results with 
these translated tests as with the original tests.  However, there has 
not been a thorough review by the team of test failures by XHTML or HTML 
processors.  There are at least 4 hc_* tests that attempt to create 
elements that have not been changed to use HTML tag names which I'll try 
to change in the next few days.

Received on Thursday, 6 March 2003 01:20:39 UTC