W3C home > Mailing lists > Public > www-validator-css@w3.org > December 2006

Re: BUG? text-shadow doesn't exist

From: Jukka K. Korpela <jkorpela@cs.tut.fi>
Date: Sun, 31 Dec 2006 11:08:30 +0200 (EET)
To: "R Walker (RMW Web Publishing)" <mountainash@rmwpublishing.net>
cc: www-validator-css@w3.org
Message-ID: <Pine.GSO.4.64.0612311053480.27210@mustatilhi.cs.tut.fi>

On Sun, 31 Dec 2006, R Walker (RMW Web Publishing) wrote:

> 100 h2 Invalid number : text-shadow Property text-shadow doesn't exist
> : 2px 2px 2px #dedede
> Which is strange as the words "text-shadow" link to
> http://www.w3.org/TR/REC-CSS2/text.html#propdef-text-shadow which
> don't say anything about it not been valid CSS.

This is a symptom of the problem that the "W3C CSS Validator" was changed 
to use "CSS 2.1 profile" (that's technobabble for using something like 
some CSS 2.1 draft, with unspecified modifications, as the criterion for 
"validity") by default. Moreover, the links were not modified: they 
refer to CSS 2.0 no matter which "profile" is chosen. Since the current 
CSS 2.1 draft - which should not be cited as anything but work in 
progress, according to its own status statement - does not contain the 
text-shadow property, you get an error message.

Click on the "More Options" link at http://jigsaw.w3.org/css-validator/
to get a user interface where you can see the default for "profile" and 
you can change that setting.

(Alternatively, use the utility with a CSS-disabled or 
JavaScript-disabled browser, for a better browsing experience.
The page uses CSS and JavaScript to hide some essential controls by 
default, including the control where you can see the default setting for 

> I also think this maybe an error with the new Fuji' validator as it
> comes up 'Invalid number' for what looks to be an internal error
> number.

I think such bugs have existed in the "W3C CSS Validator" from the very 
beginning. And I guess your hypothesis is correct: it ejects the "Invalid 
number" text to describe its _own_ error in error reporting, in a rather 
misleading way, since CSS style sheets might well contain constructs 
describable as "invalid number".

Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
Received on Sunday, 31 December 2006 09:08:37 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:40:42 UTC