Re: [ResourceTiming] Handling of "resourcetimingbufferfull" events makes no sense

Does the following sound reasonable:

onresourcetimingbufferfull attribute

The event handler for the resourcetimingbufferfull event. The
resourcetimingbufferfull event must be triggered immediately after the
buffer used to store the list of
PerformanceResourceTiming<http://www.w3c-test.org/webperf/specs/ResourceTiming/#performanceresourcetiming>
resources
becomes full.


And the addition of EventTarget seems to be a result of this:

http://lists.w3.org/Archives/Public/public-web-perf/2013Aug/0023.html

Should we remove the inheritance from EventTarget?

Arvind


On Mon, Sep 9, 2013 at 1:39 PM, Boris Zbarsky <bzbarsky@mit.edu> wrote:

> This is implementor feedback on http://www.w3c-test.org/**
> webperf/specs/ResourceTiming/#**extensions-performance-**interface<http://www.w3c-test.org/webperf/specs/ResourceTiming/#extensions-performance-interface>
>
> Note that http://www.w3.org/TR/2012/CR-**resource-timing-20120522/#**
> performanceresourcetiming-**methods<http://www.w3.org/TR/2012/CR-resource-timing-20120522/#performanceresourcetiming-methods>does not have the problem I'm about to describe.
>
> The draft says:
>
>   onresourcetimingbufferfull attribute
>   The event handler for the resourcetimingbufferfull event. The
>   resourcetimingbufferfull event must be fired at the Document
>   immediately after the buffer used to store the list of
>   PerformanceResourceTiming resources becomes full.
>
> But setting performance.**onresourcetimingbufferfull will add an event
> listener to the performance object, not to the Document.  So setting this
> attribute will have no effect on anything; I doubt that's what this spec
> means to specify.
>
> Further, "partial interface Performance : EventTarget" is not valid
> WebIDL: you can't change the inheritance of an interface in a partial
> interface (consider what would happen if two partial interfaces did this).
>
> Whatever changes were made here from the CR are pretty broken; I suggest
> reverting them and figuring out what we actually want the behavior here to
> be...
>
> This issue is a blocker for actually implementing this specification.
>
> -Boris
>
>