W3C home > Mailing lists > Public > public-css-testsuite@w3.org > November 2010

Re: user-stylesheet-015 invalid because it assumes other rules not present at user level

From: fantasai <fantasai.lists@inkedblade.net>
Date: Tue, 23 Nov 2010 11:15:07 -0800
Message-ID: <4CEC12BB.2060204@inkedblade.net>
To: Arron Eicholz <Arron.Eicholz@microsoft.com>
CC: public-css-testsuite@w3.org
On 10/22/2010 07:33 PM, fantasai wrote:
> On 10/19/2010 12:07 PM, L. David Baron wrote:
>> These tests:
>> http://test.csswg.org/suites/css2.1/20101001/xhtml1/user-stylesheet-015.xht
>> http://test.csswg.org/suites/css2.1/20101001/html4/user-stylesheet-015.htm
>> are invalid because they assume other rules are not present at the
>> user style sheet level. (Technically a bunch of the other tests are
>> as well, but these are the only ones that actually fail as a
>> result.)
>>
>> As recommended by CSS 2.1 (section 6.4), Gecko treats a number of
>> user preferences as user style sheet rules. In particular, the
>> preference for active link colors is represented using the
>> equivalent of:
>> :link:active, :visited:active { color:<preference> }
>>
>> The test user-stylesheet-015 tests that a rule with selector
>> a:active in a user style sheet styles a link. However, a:active has
>> lower specificity than the rule above, so it doesn't work.
>>
>> If the rule's selector were a:link:active, a:visited:active it would
>> work for us.
>
> Even better would be to use an ID to qualify the link.
> This would pull its specificity above anything reasonably
> likely to be used in a user stylesheet.

Arron, I saw your update to this test using #test.
I would suggest using a selector that is very specific to
this test, e.g. #user-stylesheet-015-test, so that if the
user stylesheet is not cleared out after the test it doesn't
randomly affect other things.

~fantasai
Received on Tuesday, 23 November 2010 19:15:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 23 November 2010 19:15:53 GMT