- From: Ilya Grigorik <igrigorik@google.com>
- Date: Tue, 2 Jun 2015 13:19:32 -0700
- To: Boris Zbarsky <bzbarsky@mit.edu>
- Cc: Jonas Sicking <jonas@sicking.cc>, "public-web-perf@w3.org" <public-web-perf@w3.org>
- Message-ID: <CADXXVKpFhLKQpCCRR5DZTtYjJq24Z78uHzMs_Z8JGhc5n5bVtA@mail.gmail.com>
On Tue, Jun 2, 2015 at 10:51 AM, Boris Zbarsky <bzbarsky@mit.edu> wrote: > Put another way, what are the actual use cases here? Is it to allow > whoever has a reference to the worker to convert its times into their own > timebase? Yes, I think that's our primary use case. Because one plausible definition here would be to have the "workerStart" > getter take the timebase of the incumbent settings object's global and > return the offset between that and the timebase of the worker global. This > would have the salutary effect that code like this would have the > properties described in the comments in it: > > w.onmessage = function(e) { > var workerTimeStamp = e.data; // in worker's timebase > var timeStamp = workeTimeStamp + w.workerStart; > // timeStamp is now in the same timebase as performance.now() > } > > no matter where w came from to reach this code. So in my example above > the answer would be "second subframe" when examined from inside the second > subframe, but "parent page" if examined directly from the onload handler. > That's a nice property. On Tue, Jun 2, 2015 at 11:06 AM, Boris Zbarsky <bzbarsky@mit.edu> wrote: > So we _could_ usefully have all the different SharedWorker objects > returning different workerStart values as desired, as long as we define > "started it" as "called the SharedWorkerConstructor" (in some sense) as > opposed to "caused the creation of the SharedWorkerGlobalScope". > As far as speccing that, I'm not sure what the right hooks are here.. Boris, any suggestions on wording?
Received on Tuesday, 2 June 2015 20:20:40 UTC