- From: Simon Pieters <simonp@opera.com>
- Date: Fri, 11 Sep 2009 10:00:09 +0200
- To: "John Cowan" <cowan@ccil.org>, public-xml-core-wg@w3.org
On Fri, 11 Sep 2009 09:43:53 +0200, John Cowan <cowan@ccil.org> wrote:
> I don't see the point of this paragraph:
>
>> A PICharRef has the same semantics as a CharRef in XML, except that
>> Legal Character is not a well-formedness constraint. User agents
>> must replace the PICharRef in the pseudo-attribute's value with the
>> character it represents according to XML had it been a CharRef, except
>> if the character being referred does not match the production for Char
>> in XML, in which case the processing instruction must be ignored. [XML]
>
> Since the effect is the same (ignore the whole PI),
No, the effect of violating a well-formedness constraint is:
http://www.w3.org/TR/REC-xml/#dt-fatal
We don't want an xml-stylesheet-level error to cause an XML 1.0-level
fatal error, so to avoid such interpretation I have avoided to use the
concept of well-formedness for xml-stylesheet.
> why not go ahead
> and make Legal Character a WFC? Then change the first paragraph to
> say "must match [...] and must satisfy the WFC", and all is done.
> Nobody really wants illegal characters in PICharRefs, and for future
> use we may as well declare them a well-formedness error.
It seems I forgot to add a requirement for document conformance about
illegal characters in PICharRefs. I'll fix that.
--
Simon Pieters
Opera Software
Received on Friday, 11 September 2009 08:01:10 UTC