W3C home > Mailing lists > Public > www-style@w3.org > July 2013

Re: [selectors] Finding a way to run "complete" profile selectors in CSS

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Fri, 12 Jul 2013 09:38:16 -0700
Message-ID: <CAAWBYDBL2JnkquOK4oRaV_AZfYtoZqYLHG6d1ENra4NNcnQ4kw@mail.gmail.com>
To: Timmy <timmywillisn@gmail.com>
Cc: www-style list <www-style@w3.org>
Whoops, didn't catch that there was more text later in the message.
(Please use text-based formatting, not things like colors and
indentation, to indicate quotes.  Some of us always use text-based
email, and the list archives definitely do, so emails like yours are
harder to understand later.)

On Fri, Jul 12, 2013 at 7:16 AM, Timmy <timmywillisn@gmail.com> wrote:
> This doesn't seem like the best API to me. If the problem is that we'd like
> to be able to use selectors in the complete profile in CSS (rather than
> through QSA), you're still required to call a javascript function (i.e. the
> proposed .CSS) to use them? I say let's have a solution that is purely CSS.
>
> Perhaps this was already suggested, but what about taking a queue from
> javascript strict mode and adding some sort of CSS statement to effectively
> "turn on" selectors in the complete profile? It would be a way for the
> developer to say that she is aware they are slower, but wants to use them
> anyway. Perhaps this flag could be scoped so that it could be applied only
> to a style element or only to a particular stylesheet.
>
> This seems simpler to me than the deferred at rule (as styles from the
> complete profile can be anywhere in a user's CSS rather than separated into
> at rules). If any of the selectors from the complete profile get fast enough
> to warrant addition to the fast profile, no changes to code are necessary
> (they just get faster).

This would just be a "please make my entire page run slower" switch,
which isn't acceptable.  What's worse, it would probably work fine on
the developer's high-end desktop or beefy laptop that they're coding
on, but then be super-slow on a user's mobile phone.

~TJ
Received on Friday, 12 July 2013 16:39:02 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:32 UTC