Re: [visualViewport]: Proposal and RFC

See inline replies

On Tue, Sep 6, 2016 at 8:44 PM Marat Tanalin <mtanalin@yandex.ru> wrote:

> 10.08.2016, 19:36, "David Bokan" <bokan@google.com>:
> > We've been working on the proposal for a visual viewport API at
> https://github.com/WICG/ViewportAPI. We (Chrome) would like to ship this
> feature in the nearish future but I'd like to make sure we have consensus
> that this is a good idea and the current state of the proposal is
> acceptable before we do so.
>
> There is certainly a need for a consistent, reliable, and straightforward
> visual-viewport API for applications that dealing with physical pixels is
> crucial for, especially for webtech-based extensions like those powered by
> the WebExtensions API.
>
> As an author of the SmartUpscale [1] extension intended for disabling blur
> solely for images displayed at integer zoom levels in strong direct
> relation to _physical_ pixels, I am specifically interested in the proposed
> `window.visualViewport.scale` property (as well as an event to track its
> changes) as a legitimate way to determine pinch-zoom level instead of dirty
> unreliable workarounds like `window.outerWidth / window.innerWidth` we are
> currently forced to use for this purpose (or even more impressive
> dummy-iframe-based approach used in your polyfill).
>
> Do I understand correctly that the `window.visualViewport.scale` property:
>
> 1. is supposed to be available and functional always, regardless of
> whether there is a viewport meta element on the page?
>

Yes, this will be always be available and functional, regardless of the
viewport meta tag.


> 2. contains zoom relative to that already reflected in
> `window.devicePixelRatio`, so the total physical-pixels-per-logical-pixel
> ratio can be calculated as the _product_ of `window.visualViewport.scale`
> times `window.devicePixelRatio`?
>

By physical-pixels-per-logical-pixel, do you mean
physical-pixels-per-css-pixels? If so, then almost*.
`window.visualViewport.scale` times  `window.devicePixelRatio` doesn't take
the browser zoom into account. See this
<https://www.chromium.org/developers/design-documents/blink-coordinate-spaces>
page for more info on the coordinate spaces.


> Also, while this is somewhat implicitly clear from the "viewport's width
> or height is changed" phrase, the Visual Viewport API spec would probably
> get more clear if the description of conditions when the `resize` event
> should fire did include an explicit clarification that the `resize` event
> fires once the `window.visualViewport.scale` property is changed (due to
> e.g. pinch-zoom).
>
> You're right. Will add that to the spec. Thanks :)

> [1] http://tanalin.com/en/projects/smart-upscale/
>