Re: RTCError, DOMError, and... what happened?

On 10/11/2013 03:40 PM, Adam Roach wrote:
> On 10/11/13 00:09, Harald Alvestrand wrote:
>>
>> FWIW, the 2004 (!) version of DOMError contains the "message" field:
>>
>> http://www.w3.org/TR/DOM-Level-3-Core/core.html#ERROR-Interfaces-DOMError
>
> Apparently, I don't understand how any of this works. What I found was 
> this less-than-a-year-old document, which defines an unusably minimal 
> DOMError:
>
> http://www.w3.org/TR/dom/#interface-domerror
>
> In any case, if we're using objects defined elsewhere as a critical 
> part of this spec, can we put a clear documentation citation in here 
> please? It's very difficult to take a stab at implementing a sensible 
> error reporting scheme when our spec obliquely says "use DOMError," 
> which can apparently mean (at least!) three very different things.

Agreed.

Suggestion: we should define

[Constructor(DOMString name, optional DOMString message = "")]
interface RTCError {
   readonly attribute DOMString name;
   readonly attribute DOMString message;
};

with the comment

"We intend to use DOMError as soon as there is a stable reference for 
DOMError that contains both the name and the message field".

We should also list the "name" values we use explicitly, and note which 
ones are currently part of the DOM specification.

It seems that given that we're publishing a W3C specification, we must 
reference the W3C version of the spec, not the "living spec" - Anne 
might be able to say what the schedule is like for that.