Re: Validity of test DOM 2 HTML HTMLAppletElement 05

On Feb 2, 2007, at 8:56 AM, Jo=E3o Eiras wrote (on www-dom-ts):

>
>
> The test HTMLAppletElement 05 of DOM 2 HTML
> http://www.w3.org/2004/04/ecmascript/level2/html/=20
> HTMLAppletElement05.html
>
> is not much valid, I think.
> The test checks if the value held by applet.codeBase is the same as =20=

> declared in the page markup.
> Both Opera and Mozilla resolver the src, href, data and codebase =20
> attributes for elements. Accessing these properties directly in DOM =20=

> elements returns a resolved url, while accessing using getAttribute=20
> (..) returns the original value, also for the sake of using selectors.
> Is there as part of the specification where these properties are =20
> required to be the original values ?

Probably more effective to have interpretation questions on www-dom, =20
not www-dom-ts, so moving the conversation there.

This case seems similar to errata item html-1 (http://www.w3.org/=20
2000/11/DOM-Level-2-errata#html-1), but there isn't a "ditto for =20
HTMLAppletElement.codeBase" in the errata.  The discussion that =20
resulted in the errata is http://lists.w3.org/Archives/Public/www-dom/=20=

2003JanMar/0064.html

Also, "base uri" is used in the description for codeBase and =20
according to http://www.ietf.org/rfc/rfc2396.txt a base URI must =20
include a scheme which would support a case for resolving a relative =20
URI.

I'm not sure of the process now that the DOM WG has expired.  There =20
paths forward seem to be:

1. Endorse the test and say that resolving the URI is prohibited
2. Modify the test to accept either a absolute or relative URI as the =20=

return value from HTMLAppletElement.codebase.
3. Issue an errata mandating resolution and modify the test =20
appropriately.

Probably would be good to survey the implementations to see their =20
behaviors and review the recommendation to see if there are other =20
attributes that might have the same ambiguity.