W3C home > Mailing lists > Public > public-css-testsuite@w3.org > July 2008

RE: CSS2.1 i18n and bidi tests for review

From: Richard Ishida <ishida@w3.org>
Date: Tue, 22 Jul 2008 20:29:52 +0100
To: "'Eira Monstad'" <eiram@opera.com>, "'Bert Bos'" <bert@w3.org>, "'fantasai'" <fantasai.lists@inkedblade.net>
Cc: <public-css-testsuite@w3.org>
Message-ID: <00f601c8ec31$513883b0$f3a98b10$@org>


Quite by coincidence I had been working on an update to these tests and
their results on the i18n site.  I have updated the format and made a few
small corrections.  I also added some tests related to script tags in RFC
4646 (BCP 47), and tidied up the tests related to xml:lang so that they test
whether a UA uses lang related selector values with xml:lang attribute
values when the document is text/html (they don't). 

Today I extended the tests to cover XML 1.1 served as XML, using xml:lang
with an escape (couldn't get xml|lang to work).

I haven't publicly announced the new versions yet, but they can be found at 
http://www.w3.org/International/tests/test-css-lang-1 (and -2 to -6) and the
first set of results at

Wrt the CSS tests, the files are served as XHTML 1.1 but the assertion says
"lang attribute selector with att=val in HTML...".  Is this what was
intended or should the assertion say XHTML?  Note that the lang attribute is
not part of XHTML 1.1, which is why the pages don't validate.

Hope that helps,

Richard Ishida
Internationalization Lead
W3C (World Wide Web Consortium)


> -----Original Message-----
> From: Eira Monstad [mailto:eiram@opera.com]
> Sent: 05 July 2008 23:08
> To: Bert Bos; fantasai
> Cc: public-css-testsuite@w3.org; 'Richard Ishida'
> Subject: Re: CSS2.1 i18n and bidi tests for review
> On Sat, 05 Jul 2008 11:30:04 +0200, Bert Bos <bert@w3.org> wrote:
> > fantasai wrote:
> >
> >>
> >>
> >>   Same comments about background-color etc.. but I think we need
> >>   a spec clarification here before I can review the content of
> >>   the test. I would expect :lang to be case-insensitive because
> >>   the language codes are... But that's not specified in the spec,
> >>   so.. I guess I'll have to file an issue.
> >
> > Section 5.11.4 indeed doesn't say that the argument of ':lang()' is
> > case-insensitive, but I don't think it should. The selector matches
> > whatever the document defines to be its language, and although we hope
> > and recommend that every document uses RFC 4646 (and the selector
> > clearly works best if it does), we don't require that.
> >
> > This test case uses HTML and HTTP, which do use that RFC, and thus the
> > language codes in this test case are indeed case-insensitive.
> >
> > I think it is correct to test for that, because CSS says (4.1.3) that it
> > defers to the document for the case-sensitivity of text that comes from
> > the document. I.e., a CSS UA that fails this test case is indeed
> > non-conforming.
> I agree. I spent some time tracking down justifications for either
> interpretation before deciding to test for this, but ended up with the
> same conclusion, defer to document. I also discussed briefly with Hixie,
> who agrees too.
> --
> Eira Monstad
> CoreQA
> Furthermore, it is my opinion that BTS2 must be destroyed
Received on Tuesday, 22 July 2008 19:30:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:13:18 UTC