Re: [css3-regions] flow-into on inline or table-cell

Hi Alex,

On Feb 16, 2012, at 12:10 AM, Alex Mogilevsky wrote:

The spec says ‘flow-into’ applies to anything. If applied to inline elements, what are expectations for flow behavior – are inline elements concatenated in lines, or should there be an anonymous block for each one?


[vh] You are talking of something like:

<div id="d1" style="flow-into:article">...</div>
<p>
Some of my text <span id="s1" style="flow-into: article">is moved to a flow</span>.
</p>

In that case, the flow would be equivalent to:

<div id="article-flow"> <!-- wrapping container -->
    <div id="d1">...</div>
    <span id="s1">...</span>
</div>

In this case, an anonymous block-container should be inserted around the s1 span because of the rules on anonymous block boxes (section 9.2.1.1 of CSS 2.1). However, I think the processing model should be:

a. move elements to the flow, without anticipating if anonymous block boxes are required.
b. process the flow content normally with regards to anonymous boxes. In our example, because there is both inline and block content in the flow, then a wrapping block container is inserted.


Also, if elements in flow are of the kind that normally require fixup (like table-cell), is it expected that fixup is performed separately in the composed flow, or would such elements change effective display type (e.g. table-cell would become a block unless it somehow ends up in a table-row)?


[vh] I would follow the same rule of thumb as described above and do the fixup in the flow content, as in step b. above.

Does that make sense to you?

I have created a new issue in bugzilla for this;

https://www.w3.org/Bugs/Public/show_bug.cgi?id=16002

Cheers,
Vincent