Le 11/11/2012 20:33, MURAKAMI Shinyu a écrit : > The large difference is that the existing spec has the rule “The sum > of the outer widths of the three boxes is equal to max box width.” Ok, keeping that behavior sounds reasonable. Especially in the default case where all three widths are auto. > […] in the following rules: > * ... the used width for A and C is the respective ‘max-content’. > * ... the used width for B is ‘max-content’. > and the rule > * ... the used width is min(max-content, max(min-content, available)), > will be > the used width is max(min-content, available) , I’m not sure that exact set of changes covers all cases, but I think I understand the general intent and I’ll update the algorithm in that direction. > unless ‘width: fit-content’ is specified explicitly. > > We have not implemented yet the ‘width: fit-content’ for margin boxes > but I think it will be useful. The definition of fit-content is: min(max-content, max(min-content, fill-available)) If we take that literally, any of the three box could (with a big enough max-content) occupy the whole containing block, without regard of the amount of content in the other margin boxes that share the same containing block. So I think we would have to specialize the definition of fit-content. Also, things can get complicated when the three boxes mix 'auto', 'fit-content' or absolute lengths for 'width'… Any suggestion? Cheers, -- Simon SapinReceived on Monday, 12 November 2012 17:54:36 UTC
This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:14:21 UTC