- From: Kang-Hao (Kenny) Lu <kanghaol@oupeng.com>
- Date: Wed, 15 Aug 2012 15:30:16 +0800
- To: François REMY <fremycompany_pub@yahoo.fr>
- CC: Boris Zbarsky <bzbarsky@MIT.EDU>, WWW Style <www-style@w3.org>
(12/08/01 18:10), François REMY wrote: > My point was to say it should react the same way as auto does on a DIV. > If the way DIVs react to floats depends on other properties, > fill-available should probably do the same, to reuse the existing > algorithm. Why? You didn't provide a reason why 'fill-available' should reuse the existing algorithm. Duplicating the keyword for the same behavior is obviously not what we want to do here. In other words, in general, the bigger the difference between the behavior of 'auto' and 'fill-available' the more things Web developers can do. > If that's too complex or undesirable, I have no strong opinion on this > however. Duplicating the keyword wouldn't be complex. On the contrary, it's pretty trivial. However, it might indeed not be what authors want. I don't (yet) have an opinion to fantasai's question, but there are a bunch of questions that are similar: 1. Should 'fill-available' on a 'inline-block' fill its containing block or fill the remaining space after other contents in the same line have been placed (assuming the break before the 'inline-block' is forbidden)? 2. Should 'fill-available' on a 'inline-block' fill its containing block or the line box, which is smaller in the presence of a float. Existing implementations (Chrome 23, Firefox 14) treat 'fill-available' like 'auto' in these examples. It seems that doing the other way around would require a change to the order of various algorithms (for example, to do 1. the width of the 'inline-block' has to been determined after line breaking is done), so I am not sure if browsers are willing to do that. 3. For abs-pos 'table', I think it makes sense to have 'fill-available' match Opera and Firefox's behavior of 'auto'. See [1]. [1] http://lists.w3.org/Archives/Public/www-style/2012Aug/0395 Cheers, Kenny -- Web Specialist, Oupeng Browser, Beijing Try Oupeng: http://www.oupeng.com/
Received on Wednesday, 15 August 2012 07:30:47 UTC