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

Re: jsUnit and the DOM TS

From: Dimitris Dimitriadis <dimitris@ontologicon.com>
Date: Tue, 16 Apr 2002 00:12:53 +0200
Cc: www-dom-ts@w3.org
To: bclary@netscape.com (Robert Clary)
Message-Id: <EEC5A264-50BD-11D6-B068-000393556882@ontologicon.com>
To begin with, thanks Robert and Edward for the work. I will certainly 
try to provide as much feedback as possible.

Some comments inlined:

On Monday, April 15, 2002, at 10:07 , Robert Clary wrote:

> I would like to start a thread here in the DOM TS mailing list where we 
> can discuss the application of the upcoming jsUnit 1.3.0 to the DOM TS.
>
> Some topics include:
>
> How to incorporate the new version into the DOM TS and if any patches 
> to jsUnit will be required to do so.
>
[dd] I think the easiest approach would be that the previously released 
DOM TS is deprecated in view of the one we are going to release next, 
Level1 with HTML, where we can include the latest version of JsUnit. 
Documentation and so forth will of course be updated.

> The explicit exposure of test function names in order to increase the 
> number of supported browsers.
>
> How to distinguish the default behavior of a browser from that provided 
> by plugins which are not exposed as part of the browser's native 
> document? For example, the DOM support provided by Internet Explorer's 
> document object is quite distinct from that provided by the MSXML 
> ActiveX control. This is somewhat similar to the situation with the 
> Adobe SVG plugin.
>
[dd] On this I have no obvious idea, but it will certainly be raised, so 
we had better deal with it. I think this list's authority on this issue 
is Curt.

> The use of non-proprietary, standards based methods to load test 
> documents.
>
> For example, how can jsUnit's new asynchronous Document loading be used 
> in the DOM TS? The new version of jsUnit provides a Document Loader 
> which can load documents into an IFRAME from where they can be accessed 
> by a test function. This is probably the most cross browser solution 
> possible since it relies only on basic HTML 4, JavaScript and DOM 0 
> support.
>
> I think this approach should be the default method of loading test 
> documents for the ECMAScript based tests, however there are issues with 
> this approach. For example, Internet Explorer 5 on Windows transforms a 
> raw XML Document into a DHTML Document when loaded into a browser 
> document. Is there any way to turn this feature of Internet Explorer 
> off? This may mean that Internet Explorer will require a different XML 
> Document loading mechanism for the DOM CORE TS, but should not preclude 
> supporting other browsers which do not have this problem. I do not 
> believe this affects Internet Explorer's ability to load HTML Documents 
> and should not affect the ability of the DOM HTML TS to use this 
> approach when loading test documents.
>
[dd] Robert, first please report to the list how Mozilla behaves under 
the circustances you mention. Second, coudl the Microsoft representative 
please provide feedback here? That will bring us one step closer to 
support in both major browsers.

> Unfortunately, not all browsers support this approach. In particular, 
> IE 5.1 on Mac OS 9 and OS X do not appear to support this document 
> loading approach. It remains to be seen whether other browsers support 
> this approach as well. It is important that jsUnit 1.3.0 alpha be 
> tested in as many browsers as possible so that we can determine the 
> best document loading strategy. I encourage everyone to participate in 
> the testing of jsUnit 1.3.0.
>
[dd] I'll run the DOM TS on my machine (OS X, OS 9, both Mac, and Linux 
for PPC, Debian distribution, where I hope to be able to run it on 
Konqueror and Mozilla 0.9.9, yesterday's nightly build).

Curt, I remember that you had cc:ed this list a message to the Konqueror 
community, are you still active there? Their feedback would also be 
meaningful.

In http://lists.w3.org/Archives/Public/www-dom-ts/2002Mar/0002.html I 
raised some of the issues you touch on here and cc:ed Microsoft's 
representative for the Mac platform. I haven't seen any feedback and 
kindly invite the rep. to try the modified JsUnit and report to this 
list or to Edward and Robert on how IE on Mac behaves.

> Is there a different approach which will provide better cross browsers 
> document loading support?
>
> /Bob
>
Received on Monday, 15 April 2002 18:12:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 6 April 2009 12:58:46 GMT