Re: [css-syntax] ISSUE-329: @charset has no effect on stylesheet??

On Wed, Jan 15, 2014 at 8:08 AM, Zack Weinberg <zackw@panix.com> wrote:
> On Wed, Jan 15, 2014 at 10:39 AM, Anne van Kesteren <annevk@annevk.nl> wrote:
>> On Wed, Jan 15, 2014 at 3:22 PM, Richard Ishida <ishida@w3.org> wrote:
>> > I don't expect that the ordinary user of CSS will see any sophistication
>> > here - for them this is just an @-rule with some special characteristics
>> > that they need to get right if it's going to be effective. Why don't we just
>> > frame it that way in the spec?  I think the benefit would be that (a) it's
>> > more straightforward, and (b) that it doesn't make it look like we suck
>> > (instead, those people who don't follow the special rules would just make
>> > pages that suck).
>>
>> The suckage cannot be removed. For compatibility it has to show up in
>> the object model. That's the whole problem. If that was not a
>> requirement we could skip the @charset bytes before parsing and treat
>> @charset rules as errors during parsing.
>
> That's true but I don't think it affects Richard's point regarding the
> exposition.  Concretely, I think that section 8.2 would be better if
> it read something like this:
>
> The @charset rule is a reflection in the object model of the
> stylesheet's fallback encoding as determined by the algorithm in
> section 3.2.

But it's not!  A preceding BOM will override it, for example.  The
@charset rule really, *truly* has no effect on the document, and
should not be looked at for any reason.

>  For compatibility's sake, an @charset rule will still
> appear in the object model, as-if parsed as a normal @-rule, even if
> it was not recognized as an encoding annotation by the algorithm in
> section 3.2 (for instance, if the @-sign was not the very first byte
> of the file, or if the word 'charset' was not entirely lowercase).
> However, @charset is invalid if it is not the very first top-level
> rule of a stylesheet.  Moreover, adding, modifying, or deleting
> @charset rules via the CSSOM does not change the stylesheet's
> encoding.

I've attempted to further clarify the explanatory text.

~TJ

Received on Wednesday, 15 January 2014 21:49:05 UTC