- From: Yoav Weiss <yoav@yoav.ws>
- Date: Fri, 27 Feb 2015 18:48:13 +0100
- To: Rune Lillesveen <rune@opera.com>
- Cc: "www-style@w3.org" <www-style@w3.org>, Brad Kemper <brad.kemper@gmail.com>, Kenneth Rohde Christiansen <kenneth.christiansen@gmail.com>, Florian Rivoal <florian@rivoal.net>, Matt Rakow <marakow@microsoft.com>, David Bokan <bokan@chromium.org>
- Message-ID: <CACj=BEhLaL-5RA9L-OGz1EtJ=tU7yB_5kxdP4WX0fXhysYsrTw@mail.gmail.com>
On Fri, Feb 27, 2015 at 11:39 AM, Rune Lillesveen <rune@opera.com> wrote:
> On Thu, Feb 26, 2015 at 11:52 PM, Yoav Weiss <yoav@yoav.ws> wrote:
> > On Thu, Feb 26, 2015 at 10:58 PM, Brad Kemper <brad.kemper@gmail.com>
> wrote:
>
> >>> At that point, you would know which images to fetch. If you are
> >>> pre-loading based on viewpoint assumptions, then just stop once the
> >>> assumptions prove false.
> >>>
> >>> PS: the length of the vw, vh, etc. units don't need to be resolved
> until
> >>> after the cascade stage either.
> >>
> >>
> >> That is far from being true. `vw` units are used in `sizes` in order to
> >> determine the image density regardless of the external CSS downloaded.
> >>
> >>
> >> So, they don't use the initial viewport for that?
> >
> > They do now. But once @viewport is supported, they should switch to the
> > actual viewport, or they risk resulting in redundant downloads.
>
> I'm assuming you're referring to the Blink implementation of responsive
> images.
>
I'm referring to the Blink and WebKit implementations.
>
> @viewport is no different than <meta name=viewport> when it comes to
> affecting the actual viewport. Are you currently not taking meta
> viewport into account?
>
My bad. Meta viewport should be currently taken into account. (I need to
add more testing around that tho)
>
> The only differences between the meta tag and @viewport for this
> discussion are:
>
> 1. The meta tag can only appear in the document itself (not restricted
> to <head> in the implementation in Blink, afaict).
Hmm, I assumed that it was.
> If @viewport is
> allowed in external style sheets, it's easier to trigger loading extra
> resources and trigger unnecessary reflows.
>
> 2. Implementations typically wait for stylesheets to finish loading
> before applying the style, while the viewport meta is typically
> applied as encountered. There shouldn't be a problem collecting
> @viewport rules from stylesheets to calculate the viewport size for
> pre-loading images for instance.
>
I agree, assuming that @viewport is encountered before any images have
already started fetching. The further down @viewport is in the external
style, the probability for that shrinks.
>
> This is much of an implementation detail, I think. Bottom line, what
> I'm trying to say is that whatever pre-scanning, threaded html parser,
> or whatever optimizations you have today, the resource loading problem
> is exactly the same with <meta viewport> today when put in the same
> position in the document:
>
> <head>
> <style>@viewport { ... }</style>
> <meta name="viewport" content="...">
> </head>
>
>
I agree, which is why I think we should limit that position for @viewport
(and maybe also for <meta viewport>, if still feasible), to avoid/reduce
that risk.
> --
> Rune Lillesveen
>
Received on Friday, 27 February 2015 17:48:41 UTC