W3C home > Mailing lists > Public > public-css-testsuite@w3.org > November 2009

Re: Submission of more tests for inclusion in CSS 2.1 testsuite

From: James Hopkins <james@idreamincode.co.uk>
Date: Sun, 29 Nov 2009 00:42:19 +0000
To: css21testsuite@gtalbot.org
Message-Id: <774641C1-C72B-4207-8798-6B043CC1C86B@idreamincode.co.uk>
Cc: public-css-testsuite@w3.org
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/list-style-position/overflow/list-style-position-003.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/list-style-position/overflow/list-style-position-004.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/list-style-position/float/list-style-position-005.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/list-style-position/float/list-style-position-006.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/list-style-position/float/list-style-position-007.html
> James,
> in that
> submitted/list-style-position/float/list-style-position-007.html
> I suggest to replace "browser window" with "page" since the tests are
> aimed at several UAs.

By "aimed at several UAs", I guess you meant to say "aimed at UAs  
which aren't media-specific".  I concur with this, along with the fact  
that 'page' is more likely a recognized term than 'browser window' by  
non-authors. That is now changed.

> Also, in that test, why not just say "square"? Why say "square/ 
> rectangle"?

This was part of a past pass criterion, where no 'width'/'height'  
values were specified and computed values were solely based on a  
single '&nbsp' character. It seems that I simply failed to update this  
criterion; the test case has now been changed to reflect the updated  
styles (specified 'width'/'height' values) used.

> Also in that same test, I think
> <title>CSS Test: Marker box position - adjacent sibling block box with
> 'float:right'</title>
> should be reviewed. And it seems that such test is actually testing 2
> distinct, separate expected results: marker box position and float
> applied to nested child.

This is an incorrect assumption. The test isn't testing two separate  
expected results; in fact, it's testing the relationship between the  
marker box (positioned inside the principal box) together with a  
floated (float:right) child. Opera 10.0 gets this wrong, both when the  
child is floated to the left or right.

>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-001.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-002.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-003.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-004.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-005.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-006.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-007.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-008.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-009.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-010.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-011.html
>> http://test.csswg.org/source/contributors/jameshopkins/submitted/text-decoration/visibility/text-decoration-012.html
> In all of your visibility/text-decoration tests, you have this sort of
> structural markup:
>  <div id="test">
>   visible
>   <span>hidden</span>
>  </div>
> I suggest to use instead:
> <div>
> Filler text
> <span>FAILED</span>
> </div>
> so that the tester would know immediately that the test failed for  
> some
> reason if visibility hidden was not honored.

I disagree. In this series of tests, I'm not specifically testing  
whether the contents of an element with 'visibility:hidden' are  
correctly hidden. I'm instead testing whether children with  
'visibility:hidden' applied affect the propagation of a text- 
decoration value to a line box.

The fact that elements with 'visibility:hidden' might be incorrectly  
visible, may - depending on implementation - have no bearing on  
whether the current test passes or fails in a particular UA. More so,  
if I were to add your proposed 'FAILED' text, thus testing for actual  
content visibility, it would essentially create "2 distinct, separate  
expected results", which aren't mutually exclusive; a) contents of  
elements with 'visibility:hidden' are in fact correctly hidden and, b)  
my original test subject - whether child element(s) with  
'visibility:hidden' applied don't affect the propagation of a 'text- 
decoration' value to a line box

> Also, id="test" is not
> needed since there is only 1 wrapping div container in all your tests.
> <div id="test"> does not add any useful meaning, helpful  
> significance to
> your tests.

I'm using this ID value as it denotes the test boundary. I don't think  
there's any harm in leaving it in?
Received on Sunday, 29 November 2009 00:43:00 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 20 September 2010 17:52:01 GMT