- From: Shelby Moore <shelby@coolpage.com>
- Date: Tue, 26 Oct 2010 21:10:56 -0400
- To: "Tab Atkins Jr." <jackalmage@gmail.com>
- Cc: "Håkon Wium Lie" <howcome@opera.com>, "Alex Mogilevsky" <alexmog@microsoft.com>, "www-style@w3.org" <www-style@w3.org>
Tab Atkins Jr. wrote: > On Tue, Oct 26, 2010 at 5:17 PM, Shelby Moore <shelby@coolpage.com> wrote: >> Now assume bar is before baz, and both bar and qux are half as tall as >> foo, so then: >> >> Inline overflow: >> f | f | f | bar >> baz-----------------> >> oo | oo | oo | qux > > This is neither sane, nor consistent with existing thought on how > column spanners would work. Column spanning elements go *below* any > preceding content; they *definitely* don't split the content of > previous columns based on where they natively sit in their original > column Then the spec needs to say that. I am unable to find where the spec currently says "no spanning prior columns within the same row of columns" nor the equivalent. Assuming you specify that restriction, then we can still get sane results and the following applied to non-block content (contrary to your original example apologies): Inline overflow: f | f | f | baz-----> oo | oo | oo | bar | qux The block overflow case remains as I have written in my corrections. More examples of inline overflow follow. f | f | f | baz-----------------> oo | oo | oo | bbb | arr | qqq | uxx If you put bar before baz, then: f | f | f | bar | qux oo | oo | oo | baz-----> And: f | f | f | bbb | baz-----> oo | oo | oo | arr | qqq | uxx Let me go back to your original exact example and apply your desired restriction from above: http://lists.w3.org/Archives/Public/www-style/2010Oct/0674.html Afaics, you didn't specify the constraints, so let me answer with all cases. Neither height nor width constrained (thus no overflow): foo | foo | foo | bar | baz ------new column row------- qux The reason is because baz will span all remaining columns and the columns are only one block element in height in your example. Next if height is height is constrained at one block element in height (doesn't matter what width is, because current spec forces overflow inline always): foo | foo | foo | bar | baz There is no place to put qux in the above example. It is a degenerate case. It is an error in the design of css-multicol. Should I go on with more examples? It would be better if you fully specified your example first, because I can not see where Alex is getting his layout from in the spec. > (in addition to being weird, it would produce a dependency > loop, as the content in the previous columns could flow differently > when split by and move the column spanner). Okay I can agree with that.
Received on Wednesday, 27 October 2010 01:11:30 UTC