Re: [selectors4] Spec for attribute value case-sensitivity is clearly to vague

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:59 UTC

On Thu, Jan 21, 2016 at 8:34 PM, Boris Zbarsky <bzbarsky@mit.edu> wrote:
> We now have at least two implementations of
> https://drafts.csswg.org/selectors-4/#attribute-case shipping in release
> browsers: Safari and Chrome.  Sadly, they are not actually interoperable
> with each other.  Simple testcase:
>
>   <!doctype html>
>   <style type='text/css'>
>     [foo="bar" I] { color: purple; }
>   </style>
>   <div foo="BAR">I am purple in Safari but not Chrome</div>
>
> The spec says "the attribute selector may include the identifier i before
> the closing bracket" but doesn't make it clear what the case handling is.
> Given the case-insensitivity of CSS in general, clearly WebKit decided to
> also accept "I", whereas Blink read this more literally.
>
> The spec could use better wording or a more exhaustive example, or
> _something_, since clearly what's there right now is not enough to produce
> interoperable implementations.

The spec is quite clear on the matter:
<https://drafts.csswg.org/selectors-4/#case-sensitive> states plainly
that all of Selector syntax is case-insensitive except for the few
exceptions called out in that section.

This is just a silly bug on Blink's part.

~TJ