Re: HTML5 parser integrated in W3C Markup Validator

olivier Thereaux wrote:
> 
> Dear all,
> 
> as part of an effort to promote the current work on HTML to web 
> developers, I've been working lately on integrating the W3C Markup 
> Validator with the HTML5 parser and checker developed over the past few 
> years by Henri Sivonen. Thanks to validator.nu's well documented API and 
> Henri's helpful cooperation I managed to integrate the two tools fairly 
> seamlessly. The result of that work is in CVS, and testable on the dev 
> instance of the validator:
> http://qa-dev.w3.org/wmvs/HEAD/
> 
> Test with a "random" HTML5 page:
> http://qa-dev.w3.org/wmvs/HEAD/check?uri=http%3A%2F%2Fintertwingly.net%2Fblog%2F 

Random, eh?  :-)

The warning puzzles me.  I *think* that document is valid XHTML5, served 
with the MIME type that draft HTML5 standard suggests for such documents.

http://html5.validator.nu/?doc=http%3A%2F%2Fintertwingly.net%2Fblog%2F

Perhaps Ian or Henri could confirm?

Here's HTML5 served as text/html:

http://rails.intertwingly.net/blog/index.html

And it produces a different warning:

http://qa-dev.w3.org/wmvs/HEAD/check?uri=http%3A%2F%2Frails.intertwingly.net%2Fblog%2Findex.html&charset=(detect+automatically)&doctype=Inline&group=0

Which, again, is not produced by Henri's service:

http://html5.validator.nu/?doc=http%3A%2F%2Frails.intertwingly.net%2Fblog%2Findex.html

> Test of error display:
> http://qa-dev.w3.org/wmvs/HEAD/check?uri=http://qa-dev.w3.org/wmvs/HEAD/dev/tests/html5-error-explanation.html;ss 
> 
> (Note - the "revalidate" feature is broken on that dev server, 
> apparently due to a bug in the latest version of libwww-perl. We're 
> looking into that.)
> 
> The HTML5 parsing is triggered (for now) by these two cases:
> * if the user selects "HTML5" in the doctype dropdown in the options
>  http://qa-dev.w3.org/wmvs/HEAD/#validate_by_uri+with_options
> 
> * if the document contains the pseudo doctype declaration
>  <!DOCTYPE html> -> http://www.w3.org/TR/html5/syntax.html#the-doctype
> 
> More triggers are likely to be added later.
> 
> This integration should hopefully provide more visibility to HTML5 and 
> to the validator.nu tool: validator.w3.org gets about a million 
> validations a day and if only a fraction of it looks into the HTML5 
> option that will provide some useful additional testing grounds. In the 
> meantime, bug reports (or patches!) on how the integration of the two 
> tools have been made would be welcome, either on the www-validator list, 
> or on the bugzilla:
> http://www.w3.org/Bugs/Public/
> Please do check for existing reports before you do:
> http://www.w3.org/Bugs/Public/buglist.cgi?query_format=advanced&product=Validator&component=HTML5 
> 
> Thank you,

- Sam Ruby

Received on Monday, 25 August 2008 22:38:16 UTC