W3C home > Mailing lists > Public > public-web-perf@w3.org > October 2013

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

From: Jatinder Mann <jmann@microsoft.com>
Date: Tue, 1 Oct 2013 22:30:41 +0000
To: Boris Zbarsky <bzbarsky@MIT.EDU>, Arvind Jain <arvind@google.com>
CC: public-web-perf <public-web-perf@w3.org>
Message-ID: <d808fdbbb4eb4d93a5605e9e4c435bd2@BLUPR03MB065.namprd03.prod.outlook.com>
On 9/9/13 3:53 PM, Boris Zbarsky wrote:
> You need to say what target the event is dispatched on, what sort of event 
> object it is (Event, or some other interface) and what its properties are: .target,
> .bubbles, .cancelable.  I recommend looking at other specifications that dispatch 
> events to see how to do it.

What do you think of this proposed definition of the onresourcetimingbufferfull attribute:

The event handler for the resourcetimingbufferfull event. Immediately after the buffer used to store the list of PerformanceResourceTiming resources becomes full, the User Agent must fire a simple event named resourcetimingbufferfull that bubbles, isn't cancelable, has no default action, at the Performance object. 

On 9/9/13 3:53 PM, Boris Zbarsky wrote:
> On 9/9/13 5:17 PM, Arvind Jain wrote:
>> 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?
>
> It should move into whatever specification defines the core Performance interface.

I see that there is a mistake in Section 4.4 of the latest Resource Timing editor's draft: a partial interface can't extend the interface. If we want to make Performance an Event Target we will have to update Section 4.4 of the Navigation Timing L1 spec, where the Performance interface has been defined. However, that spec is currently at Recommendation. Philippe, would we add an addendum to the Navigation Timing L1 spec to accomplish this?

On 10/1/13 9:46 AM, Juan Carlos Estibariz wrote:
> I just noticed this thread and I wanted to point out that since the handler of 
> resourcetimingbufferfull is supposed to clear the buffer (so new entries can 
> be added), having more than one handler for this event (by using addEventListener 
> from EventTarget) is probably going to produce incorrect results for all but one of them.

The clearResourceTimings method will clear the buffer. If multiple handlers of the resourcetimingbufferfull event call clearResourceTimings, you will still get a cleared buffer. I'm not sure how there will be incorrect results.

Thanks,
Jatinder
Received on Tuesday, 1 October 2013 22:31:28 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:04:37 UTC