Re: [CSS21] min-height: 100% in html, body

On Fri, 27 Jan 2012 21:41:19 +0100, Simon Sapin <simon.sapin@kozea.fr>  
wrote:

> Le 27/01/2012 18:02, Øyvind Stenhaug a écrit :
>> First of all, the term "specified explicitly" seems nonsensical. All
>> elements have a specified value. On the other hand, the containing block
>> isn't an element, and isn't necessarily even related to one. So the
>> parentheses actually contain the crucial part here (and readers must
>> figure out for themselves how exactly to determine whether the  
>> containing
>> block depends on content height).
>>
>> Now, the main issue: even if assuming the simple case where the  
>> containing
>> block corresponds to some particular dimensions of an ancestor block
>> element's sole box, exactly which value does "the height" refer to?
>>
>> - The so-called tentative used height? ("Depends on content height" in  
>> the
>> example)
>> - The used value? (Does not "depend on content height" in the example)
>> (- Something else?)
>
> Hi,
>
> I agree this should be more explicit, but I think it is the computed  
> value. A computed value of 'auto' is the only case where the used height  
> depends on the content:

I don't know if that's what the spec author was thinking, but it's not  
that simple.

a) display: table-cell; height: 0;

Computed value of 'height' is not 'auto', but used value of 'height' still  
depends on the content...

b) position: absolute; top: 0; bottom: 0; height: auto;

Computed value of 'height' is 'auto', but used value of 'height' does not  
depend on the content.

> According to 10.1, a containing block is one of :
>
> 1. The viewport
> 2. A page area
> 3. The content edge or padding edge of a block container
> 4. The bounding box around the padding boxes of ... something. I’m not  
> sure how exactly that one works. (Can an inline element generate more  
> than one inline-level boxes while *not* being split across lines?)

(Yes, e.g.
<div dir="ltr"><span>english WERBEH</span> WERBEH</span></div>
where the uppercase parts represent right-to-left letters)

> 1 and 2 always have a fixed height. I’m not sure about 4, but for 3 it  
> depends on the 'height' property. The height is "specified explicitly"  
> if and only if the 'height' property dos not compute to 'auto'.

Such a definition is not compatible with what browsers actually do, and is  
inconsistent with the "depends on content height" parenthetical, as shown  
above.

-- 
Øyvind Stenhaug
Core Norway, Opera Software ASA

Received on Monday, 30 January 2012 15:31:10 UTC