Re: [CSS21] response to issue 115 (and 44)

On Thursday, February 19, 2004, 12:27:49 AM, Rijk wrote:

RvG> Mozilla refused to use the stylesheet for www.opera.com for a
RvG> while, because the webmaster had put a comment in it... in
RvG> Norwegian, using the a-ring character. The stylesheet was send as
RvG> Latin-1,

You mean it was served as

Content-Type: text/css;charset=iso-8859-1

or do you mean

the default for unlabelled text content over HTTP is Latin-1 in some
versions of the HTTP spec

RvG> but didn't contain charset info, and the referring page
RvG> was utf-8.

RvG> Clear rules on how to handle such a case have their use, I think.

The need for clear rules is indeed why I am participating in this
discussion.

I would say that the primary fault there was with Opera for serving
content on a text/css media type without looking at the text/* rules
for missing charset/encoding information (if they used an assumed
charset).

The secondary fault is with the CSS spec for encouraging the
indirected, referring page hueristic.

Mozilla was spec compliant in refusing the stylesheet, I would say, if
it lacked charset labelling.

However, if the stylesheet was served with an explicit charset
parameter, or had an explicit @charset, then Mozilla was incorrect to
refuse the stylesheet or to interpret it as UTF-8.



-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 Member, W3C Technical Architecture Group

Received on Wednesday, 18 February 2004 18:52:14 UTC