W3C home > Mailing lists > Public > www-validator@w3.org > March 2010

Re: Custom attributes in HTML elements

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Wed, 10 Mar 2010 04:48:35 +0100
To: "Jukka K. Korpela" <jkorpela@cs.tut.fi>
Cc: Claudia Murialdo <cmurialdo@gmail.com>, www-validator@w3.org
Message-ID: <20100310044835908261.2aaa8ae0@xn--mlform-iua.no>
I keep coming back to Claudia's question ...

> Claudia Murialdo wrote:
>> I need to add a few custom attributes in some HTML elements, which is
>> the best way to do it in order to keep validating with w3c validation
>> service?.

... and to Jukka's answer (internal subsets). And to his regret ...

Jukka K. Korpela, Fri, 12 Feb 2010 06:17:49 +0200:
> […] now I vaguely remember why I haven't used this nicer approach […]
> The problem here is that browsers don't even _parse_ DOCTYPE 
> declarations properly: they'll take "] >" as document content and 
> display them at the start of the page. This happens even in IE 8 and 
> Firefox 3.5, so it won't ever change. [...]

To solve the parsing issue for HTML4 pages, I've already shown a HTML4 
syntax compatible solution which gets parsed without the "]>" in the 
body. [1] 

But a XHTML syntax compatible solution for text/html parsers seemed 
impossible.  Until I learned that processing instructions (PIs) are 
allowed inside the internal subset, which made the solution even 
simpler than the HTML4 solution:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" 
[ 
<!ATTLIST html class CDATA #IMPLIED>
<?parser-hack ><!--?>
]> 
<!--><?!-->

	Explanation of how it works, for the interested:

The "<?parser-hack" line is an XHTML PI, and it is thus terminated by 
the "?>". However, Web (text/html) parsers still parse it as an HTML4 
PI, thinking that it is terminated by the first ">". This allows us to 
fool the text/html parsers in a cross browser compatible way by 
sticking inside the PI a ">" followed by the beginning of an HTML, 
which we in turn place before the "]>". 

The last, funny looking comment line catches and "breaks up" the 
comment that most browsers perceived as starting inside the PI and 
MUST/SHOULD - by and large - look exactly like above, in order to be 
cross browser compatible.

I also tested serving the above as "application/xhtml+xml". It worked 
flawlessly in Opera and Firefox. But  Webkit apparently had a parsing 
bug which were causing "yellow screen of death". Seems like Webkit's 
XML parser applies text/html parsing inside the DTD ... (It works 
without a hitch if you place the same PI anywhere inside the body of 
the document.)

[1] http://målform.no/html4-or-html5/take2_workaround

-- 
leif halvard silli
Received on Wednesday, 10 March 2010 03:49:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:40 GMT