Re: box-sizing + intrinsic dimensions

On Tue, Aug 7, 2012 at 2:12 PM, Boris Zbarsky <bzbarsky@mit.edu> wrote:

> On 8/7/12 3:39 PM, Ojan Vafai wrote:
>
>> How should box-sizing affect things that are sized based off intrinsic
>> dimensions?
>>
>
> For what it's worth, in Gecko -moz-available (which is the fill-available
> value) subtracts off the elements own border and padding in content-box
> sizing but not in border-box sizing, so that the actual content width is
> whatever is actually available.  This seems like it has significantly
> better ergonomics than the alternative.
>
> Similar for min-content and max-content, actually: I believe they always
> set the content width, not the overall width.


Yeah. That's fine with me. In fact, I think that might coincidentally
already be what WebKit's implementation does.

 I suppose you're just getting what you ask for here?
>>
>
> A fundamental problem with that is that it's impossible to ask for
> different box sizing in different dimensions.  If a page could do
> content-box sizing horizontally while doing border-box vertically, it would
> be a lot more reasonable to say "don't do that" when using border-box
> horizontally with intrinsic widths.
>
>
>  I'm not sure it's worth adding magic to fix these cases.
>>
>
> In case it matters, it doesn't take much magic: you just have to have the
> computations for the intrinsic widths add in the padding and/or border as
> needed, and then everything else works like it normally would.
>
> -Boris
>