Re: [css3-multicol] unknown available width and shrink-to-fit

On 09/25/2012 10:03 AM, Simon Sapin wrote:
> See the previous message for details:
> http://lists.w3.org/Archives/Public/www-style/2012Jul/0598.html
>
> Summary of this issue:
>
> §3.4 of css3-multicol uses a "shrink-to-fit" variable that is used when the available width is not know. This is in direct
> contradiction, if only in the naming, with §10.3.5 of CSS 2.1 that defines in "shrink-to-fit" based on a known available width.
>
> The obvious contradiction could be avoided by using another word in in the multicol algorithm, maybe "preferred width" or
> "preferred minimum width". But the result does not make much more sense.
>
> It seems this part of the algorithm is trying to define the preferred/intrinsic widths of a multicol element. I think that it
> should just be removed, and the preferred widths left undefined so that css3-multicol can progress. They can be defined later,
> maybe in css3-sizing.

Simon, you suggested removing lines 03-10 of the pseudo-algorithm.
I've reviewed this part of the spec, and I agree. The Multi-column
module should remove the concept of 'available-width' and 'shrink-to-fit'
from its pseudo-algorithm and just use the term 'used width'.

The 'shrink-to-fit' calculations should just be the same as in CSS2.1,
i.e. the 'fit-content' expression. Then for multi-col only 'min-content'
and 'max-content' need to be defined. And since multi-col is in CR, and
we don't have interop on a good spec for that, they should be marked as
explicitly undefined for now.

~fantasai

Received on Tuesday, 9 October 2012 20:35:03 UTC