Re: Null change proposal for ISSUE-88 (mark II): proposed note

Ian Hickson, Fri, 9 Apr 2010 00:55:01 +0000 (UTC):
> On Fri, 9 Apr 2010, Leif Halvard Silli wrote:
>> Ian Hickson, Thu, 8 Apr 2010 20:37:29 +0000 (UTC):
>>> On Thu, 8 Apr 2010, Richard Ishida wrote:
>>> We can make the pragma entirely non-conforming instead of conforming 
>>> with a warning, if that would help. [...]
>> 
>> As surprise to the Mozilla community? What about the just mentioned 
>> "matching implementations" motto? [1]
> 
> Making it non-conforming doesn't affect implementations, see the change 
> proposal. It only affects authors and validators.

Right. And users and authors are covered by the term "Mozilla 
community". It affects them.

>> In zero change proposal (Mark IV) you claim: [2]
>> 
>> ]]
>> POSITIVE EFFECTS
>> * Encourages authoring behaviour compatible with both legacy user 
>> agents 
>> and with conforming user agents.
>> [[
>> 
>> But here is an example to prove that it is not true:
>> 
>>         div[lang=""]:lang(en){background:red}
> 
> That's got nothing to do with Content-Language processing; the same bug 
> can be shown purely with lang="" attributes.

Not in Gecko, which is why I only mentioned Mozilla in my argument:

http://malform.no/testing/html5/attr-lang/mozilla-lang-lottery/lang-inherit


For KHTML, Chrome and Webkit, then yes.  But Gecko implements the 
semantics of empty lang="" and empty xml:lang="". As you can verify by 
looking at the above page, with the said browsers.

> Gecko doesn't implement the 
> part of HTML5 that says that "" is to be treated as a language code.

As told above: Incorrect. If we take the page I pointed to above, which 
was entirely without any content-language header or meta declaration, 
and add a content-language declaration to it, then we see that the bug 
in Gecko (which is not a bug according to HTML4, AFAICT, but instead an 
implementation of HTML4’s generous fallback functionality), is 
*specifically* related to content-language:

http://malform.no/testing/html5/attr-lang/mozilla-lang-lottery/lang-inherit-cl


But of course, if I had been speaking about KHTML, Chrome and Webkit, 
then you would have been right. They do not implement the semantics of 
an empty lang="". At the same time they treat META content-language 
just like they treat a @lang, and so - from looking at them, we cannot 
draw any conclusions from the fact that they sometimes look as if they 
treat META content-language the same way as Mozilla:

http://malform.no/testing/html5/attr-lang/mozilla-lang-lottery/kwc

-- 
leif halvard silli

Received on Friday, 9 April 2010 02:03:13 UTC