- From: Arnold, Curt <Curt.Arnold@hyprotech.com>
- Date: Thu, 14 Feb 2002 14:33:25 -0700
- To: "'www-dom-ts@w3.org'" <www-dom-ts@w3.org>
Been a long road to this point. Probably would be good to tag the CVS at this point and continue on with the next milestones. Any W3C convention for tagging releases in the CVS? If not, how about calling it L1Core-R1.0? Probably next step would be to identify the open issues with the L2 Core suite. Obviously, the getElementsByTagName("*","whatever") issue needs to be resolved. I don't remember any other open issues off the top of my head. The Mozilla HTML loader in DOMTestSuite.js was failing, but it might have just been a nighly build issue. It had worked with previous versions, I haven't tried it with the 0.9.8 milestone release. What I had misinterpreted as a problem with getElementsByTagName()'s case-sensitivity appears to be a failure to load the <body> element into the DOM (which causes massive numbers of test failures). The Mozilla XML implementation will crash in one of the DOM L2 Core tests (about 2/3 of the way through). There had been a similar situation earlier with the DOM L1 tests. It would be nice if someone could do a find the cases that are failing. With the DOM L1, basically had to do a half interval search, splitting alltests.xml into smaller fragments until the failure was located. I don't know if the Venkman JS debugger will let you debug a script that large (MS Interdev 6 had problems with massive scripts). If you have access to Mozilla's talkback data, if you can find a recent log for carnold@houston.rr.com, it should point to the method that failed. I'll probably change test-to-jsunit.xsl to add an unload() call at the end of any test that loads and modifies a document. Closing the corresponding window in that call would minimize the desktop clutter from running the HTML tests. It would do nothing in XML or SVG loaders.
Received on Thursday, 14 February 2002 16:32:50 UTC