Re: [css3-flexbox] computing "float" and "clear" on children of a flexbox

On Wed, Jan 25, 2012 at 2:28 PM, Daniel Holbert <dholbert@mozilla.com> wrote:
> On 01/25/2012 11:15 AM, Tab Atkins Jr. wrote:
>>
>> I've just now updated the spec, so that 'float' computes as normal,
>> and changes the computed value of 'display' to block, but otherwise
>> has no effect.  Please check that I've done it correctly!  I also
>> added a line to Example 2 to illustrate this case.
>
>
> Thanks!
>
> I have one slightly-nitpicky question on the new text.  The spec now says:
>> A flexbox creates a new flexbox formatting context for its contents.
>> This is similar to a block formatting context
> [...]
>> Additionally, all of the flexbox items establish new block formatting >
>> contexts for their contents.
>
> I don't think those last two lines are strictly true.  In particular,
> suppose one of the flexbox items is itself a flexbox.  Then, per the first
> two quoted lines, it would establish a FFC for its contents -- *not* a BFC.
>  (though an FFC is "similar to a BFC")
>
> Similarly, a <table> is a flexbox item, but I don't think it establishes a
> BFC for its contents, correct?  Its <td> elements do establish BFCs, but I
> don't think the <table> itself does.
>
> (Actually, I just noticed "ISSUE 3: Figure out the right terms to use here"
> -- maybe the above is what this issue is about?)

Basically, yeah.  I'm not sure exactly how to talk about this.  The
point is just floats outside of a flexbox don't intrude inside of
them, and floats inside of flexbox items don't extrude outside of
them.  Suggestions of better wording definitely welcome.

~TJ