- From: Robert Clary <bclary@netscape.com>
- Date: Wed, 24 Apr 2002 10:21:03 -0400
- To: Mary Brady <mbrady@nist.gov>
- CC: www-dom-ts@w3.org
Mary Brady wrote: > Does this version handle loading xml, xhtml, and html > documents? The problem we were having here at NIST > for both the html and hc_compatible tests is the need > to load xhtml and html files in a consistent manner. The > files are not consistently finished loading before the tests > are run. > > --Mary > Mary, You can download the alpha version at http://bclary.com/jsunit/jsunit-1.3.0-alpha.zip and modify either the example test which loads the data.html or the one which loads the staff.xml documents to see if the types of test files you are interested in are loaded consistently. The current alpha uses IFRAMEs with an onload handler specified via attribute and appears to load documents consistently in IE 5.5+ on Windows and in Mozilla on Windows, Mac and Linux. Mozilla currently has a bug where IFRAMEs containing documents specified by file:// URLs do not fire onload handlers but Mozilla does appear to fire the necessary onload handlers when using http:// based IFRAMEs and documents. Mozilla developers are aware of this problem and I hope to convince them to include a fix for this in Mozilla 1.0. Here is a summary of the current testing of jsUnit on other browsers: I have two different reports from IE5/Macintosh testers: 1) that IE 5+/Mac can run the alpha jsUnit but that it does not load the documents into IFRAMEs but does appear to be able to load documents into FRAMEs. 2) that IE5+/Mac can not run the alpha jsUnit at all. I need to do more work to determine the best way to support IE 5+ on Macintosh if that is possible. Advice on the best way to support IE5/mac would certainly be appreciated. There is a problem in IE5.5+ on Windows where raw XML documents are converted into DHTML when loaded into the browser rather than via the MSXML ActiveX control. I hope that there is a switch that can turn this feature off but I am not sure of that. I do not know if IE5.5+ does the same type of conversion to DHTML for XHTML documents but it is easily tested. Dimitris pointed out that the report of the number of tests that were run in a test page is currently being reported inconsistently by Mozilla. I have not determined the reason for that as of yet. Navigator 4.x will not run the tests since it does not support try/catch from ECMAScript 262-3rd. Konqueror does not run jsUnit at all and has serious problems in handling modal dialogs. It may not be possible to support Konqueror in it's current state. I have contacted several Konqueror developers and am hoping to hear from them soon. Opera 5/6 do not run jsUnit due to problems in modifying the contents of FRAMEs and IFRAMEs. Opera's support of ECMAScript and the DOM 0 does not appear to be adequate to support jsUnit at this time. As you can see, additional help in testing jsUnit in multiple browsers would be very helpful. Advice from people familiar with these browsers would also be very helpful. I would like to support as many browsers as possible however I can only do so much to work around specific browser shortcomings. I believe that coding jsUnit to use reasonable basic techniques from HTML, DOM 0 and ECMAScript and providing third party browser developers a goal to improve their support sufficiently to support jsUnit may be the best approach for supporting alternative browsers (other than IE5/Windows and Mozilla) in the future without having to include browser specific work arounds in jsUnit or the DOM TS itself. I am considering approaches which would allow a choice of configuring the test cases as proposed by Curt earlier but have not reached a decision on how to go about it as of yet. I would like to determine the acceptable method for loading external data documents into tests as soon as possible so that work can proceed on how to incorporate the new jsUnit into the DOM TS and if any modifications to the current alpha would be necessary in order to use it in the DOM TS. Any and all help will be appreciated. Bob
Received on Wednesday, 24 April 2002 10:23:44 UTC