W3C home > Mailing lists > Public > www-svg@w3.org > April 2006

Re: SVGMobile12] SVGT12-183: Rules for handling "unsupported" values don't work when properties are syntactically correct and specified using W3C mechanisms other than SVG attributes

From: Ian Hickson <ian@hixie.ch>
Date: Wed, 26 Apr 2006 14:52:03 +0000 (UTC)
To: Nandini Ramani <Nandini.Ramani@Sun.COM>
Cc: www-svg@w3.org
Message-ID: <Pine.LNX.4.62.0604261445390.21459@dhalsim.dreamhost.com>

On Tue, 25 Apr 2006, Nandini Ramani wrote:
> http://lists.w3.org/Archives/Public/www-svg/2005Dec/0055.html
> Quoting from the CSS 2.1 spec:
> "If a UA does not support a particular value, it should /ignore/ that value
> when parsing style sheets, as if that value was an illegal value
> <http://www.w3.org/TR/CSS21/syndata.html#illegalvalues>. "

This applies to the parsing of CSS style sheets and is completely 
independent of the handling of the cascade when the value has already been 
successfully parsed. In the case of URIs, anything matching the relevant 
production in the grammar is considered "value". The quoted section 
therefore doesn't apply to this issue.

> So, an invalid IRI should be treated as an illegal value, which 
> according to the CSS 2.1 spec should be treated as if it were not 
> specified. So, the SVGT 1.2 honors the CSS rule.

Clearly the intention of the SVG working group is in the right place, but 
unfortunately your interpretation of the CSS specification is incorrect. 
The problem is that an invalid IRI value is still parsable and thus does 
not get ignored at the parsing stage. Indeed, the UA will be unable to 
determine that the IRI reference is "not valid" until significantly later 
in the process, at which point the UA would be unable to process the 
property "as if it hadn't been specified", since in the cascade, every 
property is always specified.

> The WG disagrees with your proposed solution::
> "Please correct the specification so that unsupported values of <paint> have
> implementable handling, e.g. requiring that unsupported values be treated as
> 'black'."
> as this would contradict the CSS 2.1 spec.

As an editor of the CSS 2.1 specification I assure you that this would not 
contradict that specification. It is in fact the current text that 
currently contradicts the specification.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Wednesday, 26 April 2006 14:52:13 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 8 March 2017 09:47:08 UTC