Re: [css3-background] background-clip-009.html has 1 issue

On 14/05/2013 8:56 AM, "Gérard Talbot" wrote:
>
> Le Dim 5 mai 2013 21:18, Julien Chaffraix a écrit :
>> Hi Gérard,
>>
>>> http://test.csswg.org/source/contributors/google/submitted/background-clip-009.html
>>>
>>> There is still an issue with that test.
>>>
>>> You added a margin-bottom and wrote
>>> "
>>> Added a margin-bottom to this test as it doesn't change the output but
>>> make it fail if the browser clips at the margin box.
>>> "
>>> http://lists.w3.org/Archives/Public/public-css-testsuite/2013Feb/0011.html
>>>
>>> but the thing is the background (background-color and/or
>>> background-image) of an element never paints its own margin area (with
>>> only the root element as the sole exception to this). Your comment
>>> suggests that the margin area would be painted with black color (the
>>> background-color of the element) if the clipping would not be executed
>>> or would be wrongly executed. But the margin area should not be
>>> painted
>>> by the background-color to begin with.
>>
>> That wasn't the intent of my reply actually. I expect the browser to
>> clip at the border box. The extra margin was designed to catch the
>> case where the browser actually paints the margin box with the
>> background color.
>
> What if it would have painted the margin box and then clip at the border
> box? Then we would not know that the background painting implementation
> is wrong.
>
> We have at least one test checking, verifying that background-color does
> not (never!) paint the margin area of an element:
>
> [RC6]
> http://test.csswg.org/suites/css2.1/20110323/html4/margin-backgrounds-001.htm
>
> [nightly-build]
> http://test.csswg.org/suites/css2.1/nightly-unstable/html4/margin-backgrounds-001.htm
>
> ... albeit the text assert of those tests seems awkward and perfectible
> to me...

At least one test is required to make sure that backgrounds do not paint 
in the margin box.


>> In retrospect it's probably not that much of an
>> issue but I thought it was nice to have.
>>
>>> So, I believe we should remove
>>>
>>> line 23:     margin-bottom: 50px;
>>>      /*
>>>      Added a margin-bottom to this test as it doesn't change
>>>      the output but make it fail if the browser clips at the
>>>      margin box. Julien
>>> line 27: */
>>
>> I removed these lines as they seem to have confused you
>
> They have confused me. In my mind, in CSS 2.1, background painting is
> one distinct, separate thing and then there is background clipping
> (cisaillement, rogner, découper) which is another distinct, separate
> thing. CSS3 seems to be merging both things together under one single
> property: background-clip.
>
> Maybe (really not sure of this as it would decisively affect all of
> margin collapsing spec) CSS3 background-clip should have been split into
> 2 other properties: background-painting and background-clip ..

Gérard, please see the 'background-origin' property [1]? Note the part 
about 'background painting area'.

Also read up on the 'background painting area' that is reference in 
'background-clip' [2], 'backgrounds of special elements' [3] and also in 
3.4., 3.9., 3.6. and 7.1.


1. http://dev.w3.org/csswg/css-backgrounds/#the-background-origin
2. http://dev.w3.org/csswg/css-backgrounds/#the-background-clip
3. http://dev.w3.org/csswg/css-backgrounds/#background-painting-area


Alan


-- 
Alan Gresley
http://css-3d.org/
http://css-class.com/

Received on Tuesday, 14 May 2013 02:47:04 UTC