- From: Curt Arnold <carnold@houston.rr.com>
- Date: Wed, 28 Nov 2001 01:19:50 -0600
- To: <www-dom-ts@w3.org>
- Message-ID: <013501c177dd$11500860$7600a8c0@CurtMicron>
I was able to get Mozilla (a 11/13 nightly build) to run a modified HTMLAnchorElement04.xml test running loading an XHTML document. Doesn't solve the longer term issues, but it should allow work to proceed. Eventually we would need the harness to pick the appropriate test document (.xml for an XHTML implementation and .html for a HTML implementation). Mozilla would need to be tested with both for example. implementation.load doesn't load HTML documents, but will load XHTML documents. A test that works for both XHTML and HTML would need to use lower case tagnames in calls to getElementsByTagName. Cases that use upper case tagnames would need to be protected so they only run on HTML implementations only which currently can't be done so I'll need to expand the prologue syntax. I did change a lot of uses of null as a variable name to a reference to a variable declared using <var isNull="true"/> and added some additional schema constraints to catch the problems. Using null works for Java but would be difficult to implement for bindings (like Xerces-C) that use a stack object to represent a reference to a Node. isSupported12.xml made calls to isSupported("Core","") and isSupported("Core",null). Since these tests are a bit controversial (would depend on extending errata core-14 to isSupported), I've removed those tests from isSupported12 and created isSupported13 and isSupported14.
Received on Wednesday, 28 November 2001 02:21:25 UTC