Re: [css3-flexbox] Abspos flex-item positions

On Wed, Aug 1, 2012 at 2:17 AM, Morten Stenshorne <mstensho@opera.com> wrote:
> "Tab Atkins Jr." <jackalmage@gmail.com> writes:
>> Per our action item last week, we've defined the static position of
>> abspos flex items consistently with how they're handled in block and
>> inline flow:
>>   http://dev.w3.org/csswg/css3-flexbox/#abspos-items
>>   http://lists.w3.org/Archives/Public/www-style/2012Jul/0605.html (minutes)
>>
>> The WG just tasked us with defining it according to the resolutions,
>> so please give feedback. In particular, we'd appreciate implementors
>> (Alex, Daniel, Morten) reviewing the proposed text.
>>
>> The solution we came up with is more-or-less Proposal D in the wiki
>> <http://wiki.csswg.org/topics/flex-abspos-placeholders>, with the edge
>> cases fully specified.
>>
>> We ended up not using the concept of a "placeholder" at all here -
>> instead, the abspos item just participates in flex layout through the
>> 'order' step, and then is ignored for the rest of flex layout.  This
>> implies that 'order' applies to the abspos.  We can explicitly
>> disallow this, but based on Brad Kemper's comments
>> <http://lists.w3.org/Archives/Public/www-style/2012Jul/0628.html>, we
>> think the WG should revisit this decision.  (My objection to having
>> 'order' apply during the call was about having it apply *to the
>> placeholder*.  I'm in favor of it applying to the abspos itself.)
>>
>> In summary, all we're asking to do on the call is:
>> 1. See if anyone objects to the current text.
>
> Looks good. One issue:
>
>     if ‘justify-content’ is ‘end’, it is the inner main-end edge of the
>     flex container.
>
> It should say 'flex-end', not 'end'.

Good catch.  I'll fix shortly.

>> 2. Reverse the resolution about 'order' from last week.
>
> As an implementor I mostly dislike that 'order' affect abspos boxes, but
> if authors really love it, who am I to object? :) Besides, letting
> 'order' apply simply makes sense now, since we have "absolutely
> positioned flex items", not "absolutely positioned boxes wrapped inside
> an anonymous flex item" (although I may choose to do the latter in
> Opera's implementation anyway - not that anyone should notice, of
> course).
>
> Just a note: The positioning effect of 'order' only has an effect on
> auto-positioned ("statically positioned") absolutely positioned flex
> items. The painting order modification caused by 'order' applies to all,
> though.

Correct.

~TJ

Received on Wednesday, 1 August 2012 16:01:10 UTC