Re: navigationStart and the actual start of the navigation from Boris Zbarsky on 2014-07-02 (public-web-perf@w3.org from July 2014)

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Wed, 02 Jul 2014 12:15:51 -0400
To: Przemysław Pietrzkiewicz <ppi@google.com>
CC: Arvind Jain <arvind@google.com>, public-web-perf <public-web-perf@w3.org>
Message-ID: <53B43037.8030604@mit.edu>

On 7/2/14, 12:07 PM, Przemysław Pietrzkiewicz wrote:
> Could you explain why? How is the time the user took to decide in the
> prompt related to performance?

beforeunload events need not involve a prompt.

Again, this comes back to what exactly we're trying to measure.  If a 
beforeunload event handler takes 5s to execute and there is no actual 
unload prompts, should those 5s be included in the load time?

What about if there is a prompt?  Does it matter whether the 5s of 
execution time came before or after the prompt?

-Boris

Received on Wednesday, 2 July 2014 16:16:21 UTC