- From: Simon Sapin <simon.sapin@exyr.org>
- Date: Wed, 22 Jan 2014 13:30:39 -0800
- To: "Phillips, Addison" <addison@lab126.com>, Anne van Kesteren <annevk@annevk.nl>, Richard Ishida <ishida@w3.org>
- CC: "Tab Atkins Jr." <jackalmage@gmail.com>, Zack Weinberg <zackw@panix.com>, www-style list <www-style@w3.org>, www International <www-international@w3.org>
On 22/01/2014 12:25, Phillips, Addison wrote: > The issue here is that CSS's normal syntax allows different > whitespace and quote formation within the actual CSS of the file. But > the character encoding declaration requires SPECIFIC space and > quoting. That variation is not friendly to users of CSS, who expect > that @charset is just a normal syntactic element of CSS, even if, in > actuality, it is not. I personally don’t care about being friendly to authors who are not using UTF-8. The only thing we should teach authors is "Use UTF-8, period." It’s only implementation that need to worry about @charset, in order to not break legacy content that depend on it. > Since the byte munging involved is not > remarkably difficult to describe or implement and since it will > improve the likelihood that people "get it right" (let alone not > breaking existing stylesheets that somehow get it wrong), why not > specify @charset consistently with the rest of CSS? I'm fine with > saying it has to come first, etc. for the reasons you cite. For any change to how an existing feature behaves, we should evaluate the risk of the change breaking existing content, including content that may rely on broken rules or declarations *not* applying. I think this change is not worth it. -- Simon Sapin
Received on Wednesday, 22 January 2014 21:31:21 UTC