Re: Need reviewers for Cameron's Web IDL tests

On 9/19/13 8:31 AM, Dominique Hazael-Massieux wrote:
> Beyond the piecemeal review of test cases, WebApps will need to
> determine what to do for the WebIDL features that have no or limited
> usage/test:
> http://w3c-test.org/web-platform-tests/submissions/271/WebIDL/tests/submissions/heycam/notes.txt

My 2 cents:

 >  * unsigned long long without [EnforceRange]

This can be tested with ProgressEvent (see its ctor and members), I 
believe.   Or did this mean _with_ [EnforceRange]?

 >   * platform array objects

Imo these are broken and no one is really implementing them so far...

 >  * operation overloading with distinguishing index > 0

FormData.append, IDBFactory.open, IDBObjectStore.createIndex, and all 
the array-taking uniform and vertexAttrib methods on 
WebGLRenderingContext (also tex(Sub)Image2D, buffer(Sub)Data).

 >  * writable attribute with an enum type (JS -> IDL conversion)

WebSocket.binaryType, XMLHttpRequest.responseType

 >  * enum type as an operation return type or attribute type (IDL -> JS 
conversion)

The attributes above?

 >  * interface with a named property setter but not a named property 
creator

It's not clear to me that we need to have this feature.  Is it used at 
all?  I'd argue named property setters in general are antipatterns; does 
anything other than DOMStringMap use them?

 >  * [ImplicitThis]

This is used by every single Window and EventTarget thing, no?  Or needs 
to be, to get the right behavior.

-Boris

Received on Thursday, 19 September 2013 14:17:01 UTC