Re: ECMA harness

As you mentioned, the <assertDOMException> construct and equivalents are
designed to support binding specific means of reporting failure and so, for
example, the code-gen for a COM binding would be responsible treating the
HRESULT return value appropriately.

A ECMAScript binding that did not use exceptions to report failure would
probably best be considered a distinct binding from the standard ECMAScript
binding.  Again, it would be good to know the specific implementation that
is the cause for the concern.



----- Original Message -----
From: "Joseph Kesselman" <keshlam@us.ibm.com>
To: <www-dom-ts@w3.org>
Sent: Wednesday, November 21, 2001 2:14 PM
Subject: Re: ECMA harness


>
> Maybe I'm missing something, but...
>
> The DOM API assumes that exceptions or some equivalent calling protocol
> will be supported. It's probably reasonable to say that if you've got a
> binding which uses one of the equivalents, it's your responsibility as a
> test harness developer to establish a variant of the code-gen stylesheets
> which understands that equivalent and handles it appropriately, reporting
> when it's triggered correctly versus inappropriately.
>
> Shouldn't be noticably harder than any of the other
> compilation-to-specific-binding work that's already required...
>
> ______________________________________
> Joe Kesselman  / IBM Research
>

Received on Wednesday, 21 November 2001 15:38:13 UTC