- From: David Dorward <david@dorward.me.uk>
- Date: Mon, 3 Oct 2005 09:29:56 +0100
- To: Rapsys|Phoenix <rapsys@free.fr>
- Cc: www-validator-css@w3.org
On Sun, Oct 02, 2005 at 02:21:15AM +0200, Rapsys|Phoenix wrote: > I get an error, while my css is in theory valid... > http://jigsaw.w3.org/css-validator/validator?uri=http%3A%2F%2Frapsys.ath.cx%2F To be valid it would have to conform to the profile that you are validating against. The W3C CSS Validator does not include any profiles for "CSS2 + Mozilla Foundation Proprietary Extensions and Experimental Implementations of Proposed Properties for Future Versions of CSS". It might be preferable to throw the "Property -moz-border-radius doesn't exist" error, but the trouble with the "-moz-" prefix that the Mozilla Foundation use to indicate their extensions (and the similar issues with other vendor extensions) is that it violates the rules for property names (which cannot start with a dash character). This causes it to throw a parse error and seemingly loose some data from the style block containing the non-standard property. It might be useful for the validator to be altered so that it can cope with such property names more gracefully. Perhaps throwing the error: "Property -moz-border-radius doesn't exist and begins with an illegal character". > Could css validator just output warning for using non css proprieties > or silently ignore the check on this propriety ??? I hope not, errors are errors and should be reported as such. I suspect that a profile such as I described above could be written, but it would probably require the parse engine to be changed to it could recognise property names beginning with a dash as being invalid property names rather than unexpected data. Of course, just because something could be written doesn't mean that anybody would feel the desire to do so. The project is open source though, so I'm sure you are welcome to make the change and contribute a patch. Even if it is rejected you can still use it on your local system for your own quality control process. -- David Dorward http://dorward.me.uk
Received on Monday, 3 October 2005 08:30:12 UTC