W3C home > Mailing lists > Public > www-style@w3.org > September 2015

Re: [css-break][css-floats] Float pushed to the next fragmentainer, what about its siblings?

From: Brad Kemper <brad.kemper@gmail.com>
Date: Fri, 18 Sep 2015 07:39:03 -0700
Cc: fantasai <fantasai.lists@inkedblade.net>, www-style@w3.org
Message-Id: <00CE22EA-79D6-4DE0-8A5A-C9F4E56D6E1B@gmail.com>
To: Morten Stenshorne <mstensho@opera.com>

> On Sep 17, 2015, at 3:58 AM, Morten Stenshorne <mstensho@opera.com> wrote:
> 
> fantasai <fantasai.lists@inkedblade.net> writes:
> 
>>> On 09/01/2015 08:51 AM, Morten Stenshorne wrote:
>>> When a float is pushed to the next fragmentainer (let's use multicol as
>>> an example, so henceforth I'll use the term "column" instead of
>>> "fragmentainer"), because it's unbreakable, and there's not enough space
>>> remaining in the current column, should consecutive content also be
>>> pushed to the next column, even if there's room for it in the previous
>>> column?
>>> 
>>> Example:
>>> 
>>>     <div style="columns:2; column-fill:auto; height:10em; orphans:1; widows:1;">
>>>         line before float<br>
>>>         <img src="http://home.powertech.no/mstensho/gfx/bilen.jpg" style="float:left; height:10em;">
>>>         line after float, but in which column?<br>
>>>     </div>
>>> 
>>> Looks like I've been assuming that the answer is "yes, push everything
>>> to the next column", seeing how I implemented it in Presto, but I cannot
>>> find anything in the spec suggesting such behavior.
>>> 
>>> And I have a feeling that floats traditionally don't affect sibling
>>> content in such ways; e.g.:
>> 
>> Another example, that's closer to what happens:
>> 
>> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cstyle%3E%0A.float%20{%20float%3A%20left%3B%20width%3A%204em%3B%20height%3A%204em%3B%20border%3A%20solid%3B%20}%0A.clear%20{%20clear%3A%20both%3B%20}%0A%3C%2Fstyle%3E%0A%0A%3Cp%3EFirst%20bit%20of%20text.%20Now%20a%20float%3A%20%3Cspan%20class%3D%27float%27%3E1%3C%2Fspan%3E%20Now%20more%20text.%20Now%20another%20float%3A%20%3Cspan%20class%3D%27float%27%3E2%3C%2Fspan%3E%20Now%20a%20third%20float%20that%20also%20has%20clearance%3A%20%3Cspan%20class%3D%27float%20clear%27%3E3%3C%2Fspan%3E%20And%20here%20is%20the%20text%20after%20the%20cleared%20float.
>> 
>> Here we have a float in the middle of the text that is cleared.
>> It goes below the previous floats, but it doesn't push the
>> content after the float down to come with it.
>> 
>> So I think my conclusion is, no, the float may move, but it
>> does not take its content with it.
> 
> Good, thanks for responding. Could you word this into the spec if
> everyone agrees, then?

I'm not disagreeing, exactly, but I'm not sure. Doesn't this mean that floats can get out of order, if, say, a large left float moves to the next column, but a smaller left float that comes after it stays in the first column? Maybe there should be an exception to help keep floats in the right order. Or maybe only when consecutive floats have the same float value (since floating one left and one right can mess with the visual order anyway).
Received on Friday, 18 September 2015 14:39:32 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:57 UTC