RE: OBJECT attribute for HTMLAppletElement

Rick Rivello wrote:
 >     I'm not sure I understand.
 >     The above test loads applet2.html which has an object attribute 
 >     What else is needed to make the test valid?

Sorry, it was a long night.  I didn't notice the "2" and thought it was 
trying to infer the value for object from the running applet.

I don't know how the "object" attribute was ever used with Java 
applets.  It is supposed to be the serialized form of the applet, 
however Java serialization is explicitly binary, not character.  I've 
attached the serialized form of our org.w3c.domts.DOMTSApplet.  How you 
would get that into HTML so that the right binary sequence got to the 
applet I don't know (notice the class name is in ASCII, not UTF-16).

The Microsoft documentation does not list OBJECT as a supported 
attribute for APPLET, however it appears that they are returning an 
object pointer for something.

I'm thinking that the HTMLAppletElement.object attribute might be a good 
thing to somehow "undo" from the DOM L2 HTML spec.  If it actually 
worked, DOMString would not be the appropriate type, it should be some 
kind of byte stream.  Removing it from the spec would not be more of a 
breaking change than changing return types from DOMString to long.

If not removed from the spec, I would suggest removing this test from 
the suite since I don't see a way that the feature could really be used 
as written and if this was the only test that IE failed, then I would 
not suggest they change the current behavior just to satisfy this test.

p.s. I updated build.xml to copy the DOMTSApplet.class file into the 
HTML L2 build directory like it did for the HTML L1.  I don't think it 
 changes the results of any tests, but I believe it did make the tests 
load faster.

Received on Thursday, 22 August 2002 23:56:42 UTC