W3C home > Mailing lists > Public > www-dom-ts@w3.org > April 2002

HTML L2 tests

From: Curt Arnold <carnold@houston.rr.com>
Date: Thu, 4 Apr 2002 02:06:47 -0600
Message-ID: <001001c1dbaf$ab16c340$a800a8c0@CurtMicron>
To: <www-dom-ts@w3.org>
Cc: "Rick Rivello" <richard.rivello@nist.gov>
Rick had been running into some problems running the in-development L2 HTML suite.  Running that suite currently is a minefield.

If you run "alltests.html" will IE, almost every test will fail.  With the current DOMTestCase.js, the default media type is always "text/xml", which on IE means MSXML, MSXML doesn't support the HTML interfaces almost every test will attempt to use an HTML-specific interface and fail.

To run the tests using the "text/html" DOM, select "html_alltests.html".  You should get a test report like the file that I attached as ie6html2.html.  I assume that most of the test failures in this file are because of problems with the tests, not problems with the implementation.

Running an individual test using "text/html" requires manually editing the test to add the media type (search for text/html).  I've attached HTMLAnchorElement01.html to show how that is done and how to use alert() statements as a poor substitute for a debugger.

The loader in DOMTestCase.js doesn't work for multiple HTML documents in NN6/Mozilla.  So almost the opposite, for Mozilla and the current state of DOMTestCase.js, if you want to run the whole suite, you have to run "alltests.html" which will test the XHTML implementation in NN6/Mozilla.   I've attached nn62html2.html which lists the test failures from Netscape Navigator 6.2 on the same test but using XHTML test documents.  Tests that were failed by both implementations, most probably indicates a problem with the test or the test documents.

Running the tests with Mozilla 0.9.9 or the current nightly build will crash in one of the tests between lines 18750 and 19000.  I haven't finished the half interval search to identify the specific test that causes the crash.

Before running the tests, I committed a global replace of ignoreCase="auto" to ignoreCase="false".  There seem to be only a few tests (HTMLAreaElement06, HTMLTableElement07, HTMLTableElement08, HTMLTableElement09) that were showing behavior consistent with ignoreCase="auto".  

Received on Thursday, 4 April 2002 03:07:08 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:34:04 UTC