Bruno Fassino wrote: > On Jan 8, 2008 10:53 AM wrote: >> Bruno Fassino wrote: >>> [EMAIL PROTECTED] wrote: >>> >>>> Hmmm, I always thought that "auto" kind of carried the idea of >>>> "shrink-to-fit"? >>> width:auto has much different meanings, depending on the value of other >>> properties. >> So, basically, 'auto' means "make it the full available width, but if >> you don't know what width that is, make it as wide as the content". > > I don't think that the cases where width:auto means "fit the content" > are related to difficulties in knowing what the "available width" in > the containing block is (and anyway the 'available' width takes part > in the shrink-to-fit computation, whose algorithm is intentionally > 'not exactly' defined in the CSS 2.1 specs.)
How does the available width take part in that computation beyond functioning as the maximum width? > I think they are simply cases where a different behavior has been > deemed more useful and natural. For example floats that would take > all the available width would defeat their 'floating' behavior, making > width:auto on them practically useless. I don't think it would make width:auto on floats useless if 'auto' meant shrink-to-fit. I would think the shrink-to-fit algorithm would then function as "figure out the size of the content inside the floated element, then make the floated element big enough to hold it." -- David [EMAIL PROTECTED] authenticity, honesty, community ______________________________________________________________________ css-discuss [EMAIL PROTECTED] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ List policies -- http://css-discuss.org/policies.html Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
