- From: Yoav Weiss <yoavweiss@google.com>
- Date: Fri, 30 Aug 2019 18:37:51 +0200
- To: Roger Pack <rogerpack2005@gmail.com>
- Cc: public-web-perf <public-web-perf@w3.org>
Received on Friday, 30 August 2019 16:38:41 UTC
Hi Roger, In the latest spec, when the buffer is full, the remaining entries are not dropped on the floor unless the event handler doesn't make room for them. See https://github.com/w3c/resource-timing/pull/168 for the specification those changes. If you still believe the specification is sub-optimal, filing an issue <https://github.com/w3c/resource-timing/issues> would be the best way to go. If you think what's currently specified is not implemented in some browsers, your best route would be to file an issue with them. AFAICT from wpt.fyi results <https://wpt.fyi/results/resource-timing?label=experimental&label=master&aligned>, this part is implemented fairly well (even if not perfectly) across the board. Cheers :) Yoav On Fri, Aug 30, 2019 at 6:28 PM Roger Pack <rogerpack2005@gmail.com> wrote: > Hello. > As a note, I would have hoped that the resourcetimingbufferfull event > would have had, as part of the event, access to the new performance entry > that "is being dropped on the floor" so that it can be processed etc. > > As it is today, if people listen for that event, they are basically > guaranteed to lose at least one Entry before they clear them all or raise > the buffer size, etc. > Cheers! >
Received on Friday, 30 August 2019 16:38:41 UTC