W3C home > Mailing lists > Public > public-css-testsuite@w3.org > June 2007

Re: Test cases for 'applies to'

From: fantasai <fantasai.lists@inkedblade.net>
Date: Thu, 28 Jun 2007 14:06:07 -0400
Message-ID: <4683F88F.3050808@inkedblade.net>
To: "public-css-testsuite@w3.org" <public-css-testsuite@w3.org>

Arron Eicholz wrote:
> When creating cases that verify that the ‘applies to’ rules are working 
> correctly for the property should we create cases for all 91 html 
> elements when the spec states ‘all elements’? I think we should.
> 
> I say this because there are properties that say they apply only to 
> ‘block-level elements’, in this case, since it is listing a display type 
> I think it is ok to just test a single element from each display type. 
> Its only when the spec lists ‘all elements’ do we really have to create 
> a lot of tests to test thoroughly.

I don't think it's necessary to test for all elements. First off, all
form-related and frame-related elements are exempt due to the exception in
   http://www.w3.org/TR/CSS21/conform.html#conformance
Secondly, when the line says "all elements" it's not saying "all HTML elements",
it's saying "elements with any display type". It means

   - all 'display' values, including table display values and 'run-in' under
     both inline display conditions and block display conditions
   - both positioned (/floated) and non-positioned elements
   - both replaced and non-replaced elements
   - generated content (:before/:after) as well as normal content

Testing the interaction of CSS with various HTML elements is also useful,
and maybe we want to add a category of tests for that, but it is not what
that line of the property definition is really about.

~fantasai
Received on Thursday, 28 June 2007 18:06:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 20 September 2010 17:51:55 GMT