W3C home > Mailing lists > Public > www-style@w3.org > February 2014

Re: [selectors-nonelement] ::attr(*|localname), ::attr(ns|*), and ::attr(*)

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Tue, 25 Feb 2014 13:15:11 -0800
Message-ID: <CAAWBYDD=r79w00zrfQaphtTV3ir4ag8ZZLbammHhq8JzWOGVxw@mail.gmail.com>
To: Jirka Kosek <jirka@kosek.cz>
Cc: Simon Sapin <simon.sapin@exyr.org>, Simon Pieters <simonp@opera.com>, www-style <www-style@w3.org>
On Mon, Feb 24, 2014 at 9:20 PM, Jirka Kosek <jirka@kosek.cz> wrote:
> On 25.2.2014 1:32, Tab Atkins Jr. wrote:
>> However, there are probably use-cases for just grabbing all the
>> attributes on an element for processing, and it would be nice if this
>> was doable via something less verbose than "::attr(*|*)".  (Also, I
>> feel like that might look like something weird in ASCII-art, but I
>> can't quite tell what.)
>
> I still don't understand why ::attr(*) is not viable approach for this.
> I don't see any problem to have different behaviour here from universal
> selector when default namespace is specified because default namespace
> (on document instance) have different impact on elements and attributes.
>
>> I propose that if the argument is omitted, like "::attr()", we select
>> all attributes on the element.
>
> I can live with this, but in my opinion ::attr(*) makes much more sense.

Doing that would mean that we're not consistent with type selectors
*or* ourselves.  It's bad to me that ::attr(title) grabs only "title"
attributes in the null namespace, but you have to type ::attr(|*) to
select all the attributes in the null namespace, because ::attr(*)
instead selects all attributes in all namespaces.

~TJ
Received on Tuesday, 25 February 2014 21:15:59 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:39:19 UTC