RE: Delaying PerformanceEntry dispatch

Tim,
Out of curiosity, is there a public issue that discusses what led to this decision?

-Todd

From: Timothy Dresser [mailto:tdresser@chromium.org]
Sent: Tuesday, December 12, 2017 5:19 AM
To: Nic Jansma <nic@nicj.net>
Cc: Todd Reifsteck <toddreif@microsoft.com>; Ben Kelly <bkelly@mozilla.com>; Nic Jansma <njansma@soasta.com>; Yoav Weiss <yoav@yoav.ws>; Nathan Schloss <n8s@fb.com>; public-web-perf@w3.org
Subject: Re: Delaying PerformanceEntry dispatch

On our side, we've decided to temporarily block this on shipping takeRecords<https://github.com/w3c/performance-timeline/issues/74>, spec pull request here<https://github.com/w3c/performance-timeline/pull/98>.

We're hoping to target Q1 for this change.

On Mon, Dec 11, 2017 at 10:10 PM Nic Jansma <nic@nicj.net<mailto:nic@nicj.net>> wrote:
For our use case, in Boomerang, we currently hook into the 'load' event handlers of things like images, then find the resource's ResourceTiming entry -- via the PerformanceTimeline -- to get timing details.

We're not (yet) using PerformanceObserver to do this, and it seems reasonable to me that a PerformanceObserver would queue entries at idle priority.


- Nic

http://nicj.net/


@NicJ
On 11/27/2017 1:39 PM, Todd Reifsteck wrote:
My personal take is that this is good for performance, but I know there are use-cases on the web for pulling records from the Performance Timeline immediately.

I believe it was one of Nic, Yoav or Nate that previously mentioned this.

-Todd

From: Ben Kelly [mailto:bkelly@mozilla.com]
Sent: Monday, November 27, 2017 7:24 AM
To: Timothy Dresser <tdresser@chromium.org><mailto:tdresser@chromium.org>
Cc: public-web-perf@w3.org<mailto:public-web-perf@w3.org>
Subject: Re: Delaying PerformanceEntry dispatch

On Mon, Nov 27, 2017 at 8:33 AM, Timothy Dresser <tdresser@chromium.org<mailto:tdresser@chromium.org>> wrote:
Any thoughts on whether this makes sense, and whether you're likely to add this delay?

I filed a bug to discuss it for firefox:

https://bugzilla.mozilla.org/show_bug.cgi?id=1420913


On face value it sounds reasonable to me as long as sites have not come to rely on immediate notification.  I'm not the primary owner of this feature in gecko, though.

Ben