- From: Boris Zbarsky <bzbarsky@mit.edu>
- Date: Wed, 27 May 2015 14:53:48 -0400
- To: Ilya Grigorik <igrigorik@google.com>
- CC: public-web-perf <public-web-perf@w3.org>
On 5/27/15 2:41 PM, Ilya Grigorik wrote: > Ah, good catch, missed that in the earlier update. Opened a bug to track > this: > https://github.com/w3c/resource-timing/issues/23 Thanks! > Right, as written its SW specific.. I think we can make it work in both > cases though. Can we just have a property that works in all cases (workers, windows, etc) so people can just write their code without worrying about where it's running. > FWIW, there is no "navigationStart" attribute in RT We're not adding it as part of RT. For that matter, performance.now() is not part of RT. RT proper has no way, in a window, to convert its timestamps into something like wall-clock times. Nor can high-res timing do it on its own. Both of them need the 0 time value as wall-clock time for various things people want to do (e.g. comparing timestamps across globals). > The intent was to have "workerStart" reflect the "zero time" within a worker. Given that the zero time within a dedicated workers is the zero time of its parent page, calling it "workerStart" is pretty odd. > I do believe that latest draft of Perf Timeline should address both of > the above: > - http://w3c.github.io/performance-timeline/#the-performance-interface > - http://w3c.github.io/performance-timeline/#the-performanceentry-interface Ah, yes, that looks much better. Unfortunately, that document doesn't seem to be linked anywhere from <http://w3c.github.io/resource-timing/>... -Boris
Received on Wednesday, 27 May 2015 18:54:17 UTC