Re: [I18N Core Response][CSS21] out of range unicode escapes

Simon Pieters wrote:
> On Wed, 27 Jun 2007 00:57:23 +0200, Addison Phillips  
> <addison@yahoo-inc.com> wrote:
>> Having a validator or parser to flag it would tell people what was wrong  
>> with their file.
> 
> Indeed.
> 
>> U+FFFD is an acceptable option, but less desirable only because it  
>> produces no sign of why the failure occurred.
> 
> I don't understand why a parse error produces any more sign of why the  
> error occured than replacing with U+FFFD.

I don't understand it either. I cannot see a reason why user agent
couldn't log (and possibly display) a warning when it sees a U+FFFD. The
U+FFFD is inserted exactly where the illegal character was so it
shouldn't be too hard to locate it, if necessary. Is there any case
where such a character would be the intented character in a CSS file?

As Steven Zilles wrote, the visitor/reader/user of the content is not
interested if the page doesn't look that good because of a parser error
or because of U+FFFD. If the page *author* wants to validate his source
document he can do so and a good validator would display a warning or at
least a notice about a U+FFFD in the source.

-- 
Mikko