Alan Gresley wrote:

One of the most bazaar test I have done. I even invented my own
elements, selectors and one attribute.

http://css-class.com/test/xml/unicode.xml

Complicated tests easily lead to complication and frustration, though they can be useful and interesting too. Typically, after a failure, the test needs to be split into parts until each problem can be isolated.

The CSS validator just throws my CSS completely

It seems that any non-Ascii letter in an identifier makes the W3C CSS Validator report an error and discard the rest of the style sheet. I have just reported this bug in the www-validator-css mailing list. This seemed to be the best approach.

and the HTML validator
demands to have a Doctype to see if the document is valid.

By the old definition, a document type definition is required for markup validation, as validation means checking against a DTD.

By some
code sniffing, the HTML validator considers it to be invalid XHTML 1.0
Transitional.

It's not really sniffing. It just defaults to XHTML 1.0 Transitional once it has seen that you serve it as application/xhtml+xml and it has no doctype. And the validation naturally fails, because the markup is not XHTML. To use the validator on a document with your own tags, you need to write your own DTD for it.

IE9 shows the correct source code. Both Firefox and Opera rearrange of
the source code where it has RTL script and Safari changes the glyphs
all over.

I'm not quite sure what you mean, but then again, Hebrew and Arabic are mostly Greek to me. Anyway, the display of source code depends on the browser and on the program used as source code viewer (this can often be set in the browser settings). No requirements are imposed on it in any specification I know of, so it's a matter of practical judgement. Generally, such programs can be very simple (like Notepad), and I would not be surprised if some of them failed to apply Unicode directionality rules properly, to render Arabic letters using proper contextual forms, etc.

Yucca
______________________________________________________________________
css-discuss [css-d@lists.css-discuss.org]
http://www.css-discuss.org/mailman/listinfo/css-d
List wiki/FAQ -- http://css-discuss.incutio.com/
List policies -- http://css-discuss.org/policies.html
Supported by evolt.org -- http://www.evolt.org/help_support_evolt/

Reply via email to