Re: [RequestIdleCallback] About Fig.2 in Idle Periods section

I agree with Olli that we don't want to make the spec specific to
Chromium's behavior. Both figures 1 and 2 do both explicitly say they are
example, so the spec isn't trying to suggest that either of these are the
only possible timelines.

I would rather not have multiple diagrams showing the same thing (the spec
is already pretty long), but would be happy to make the diagrams more
browser agnostic if anyone has any suggestions, as long as they present the
same information (e.g., the user input in fig 2 will be handled once the
idle callback has returned within the 50ms capped deadline).

Cheers,
Ross


On 4 September 2015 at 10:35, Olli Pettay <olli@pettay.fi> wrote:

> On 09/04/2015 12:16 AM, Simon Hong wrote:
>
>> Hi,
>>
>> In Fig.2, there is a HandleInput period after last Idle callback finished.
>> I think vsync timeline should be inserted between last Idle callback and
>> Handle Input
>> because Fig.1 there is a vsync timeline before every Input period.
>> As I know, input handling is started at the start time of vsync in
>> chromium.
>>
>> Thanks.
>> Simon.
>>
>
>
> Input handling period isn't specified anywhere and browser engines do it
> differently, so
> perhaps better to have a few variations in the figures showing possible
> options.
> The spec should not hint chromium's behavior is the only possible one.
>
>
> -Olli
>
>