Re: Selectors Tests

Lachlan Hunt wrote:
> 
> Hi,
>   While preparing the Selectors implementation report for Opera, I came 
> across a few issues with the testsuite.
> 
> Firstly, I have a few tests for cases that are not tested in the 
> testsuite for attribute selectors and the UI state pseudo classes, 
> :enabled and :disabled with hidden input controls.
> 
> http://lachy.id.au/dev/css/tests/selectors/attribute/
> 
> The attribute selector tests are written based on the previous WG 
> resolution that they match nothing.
> 
> http://lists.w3.org/Archives/Public/www-style/2008Apr/0038.html

I'd change class="foo" to class="" and use a lime background instead
of color: green (like the other tests in the test suite).

Ideally you should be testing against both <p> and <p class="">.

 > http://lachy.id.au/dev/css/tests/selectors/pseudo-classes/ui/

Interesting. I'm not convinced that :enabled and :disabled shouldn't
match <input type="hidden">. Konqueror and Mozilla both make them
match, so I'd like to check with the CSSWG if that's really the
interpretation we want before deciding whether to accept your tests.
AFAICT WF2 requires them to match, but Selectors' wording implies
that they don't.

> Secondly, the testsuite contains tests for the :indeterminate 
> pseudo-class.  However, these tests depend on a proprietary IE extension 
> for setting the state of checkboxes to an indeterminate state.  Since we 
> don't implement the HTMLInputElement.indeterminate property, nor any 
> other controls that can be in an indeterminate state, is it reasonable 
> to mark those tests as N/A in the implementation report?

You can mark them as N/A.

~fantasai

Received on Wednesday, 29 October 2008 20:00:41 UTC