Re: [WebIDL] questions

Kartikaya Gupta:
> I have a few other questions that I ran into while implementing:
>
> 3. Can [PutForwards] be chained? That is, what happens if the
> attribute B on J also has a [PutForwards] pointing to something
> else? If chaining is allowed, how are infinite loops dealt with? The
> algorithm in section 4.4.3 seems to imply that chaining is allowed,
> since step 32 ends up calling [[Put]] on J, which will then run
> through the steps again.

Chaining isn’t specifically disallowed, and, as you say, it will work
because of the way the host object [[Put]] algorithm is written.  I’ve
added a note to point this out and to disallow cycles:

  http://dev.w3.org/2006/webapi/WebIDL/#PutForwards

> 4. Section 3.8.3 says that [ImplementedOn] takes an "identifier list",
> yet there is no such thing defined in the spec. I assume this will
> look something like [TheExtendedAttribute=identifier,identifier,...].
> Rather than introducing a new syntactical construct, would
> it make more sense to simply allow the ImplementedOn xattr
> to be used multiple times on the same interface (i.e.
> [ImplementedOn=Node,ImplementedOn=Window])?

Yes, I removed the extended attribute identifier list argument
prematurely, it seems.  I’ve done as you suggested for now, to avoid
complicating the grammar.

> Also, the Java example for ImplementedOn says "EventTarget et =
> (Node)n" instead of "EventTarget et = (EventTarget)n".

Fixed.

> And finally, a typo in section 4.4, the third bullet point. I think
> the word "value" is missing at the end of the sentence "... IDL value
> to an ECMAScript."

Fixed.

Thanks for the comments,

Cameron

-- 
Cameron McCormack ≝ http://mcc.id.au/

Received on Thursday, 27 November 2008 06:15:04 UTC