W3C home > Mailing lists > Public > www-style@w3.org > April 2012

Re: [css3-flexbox] "<flex> can also transition to and from a <length>"

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Fri, 6 Apr 2012 09:17:56 -0700
Message-ID: <CAAWBYDB3DkhnY78EXd+yKjABayJTXiTgK+WTZfzsBaQ1d6o=SQ@mail.gmail.com>
To: Daniel Holbert <dholbert@mozilla.com>
Cc: www-style list <www-style@w3.org>
On Fri, Apr 6, 2012 at 12:52 AM, Daniel Holbert <dholbert@mozilla.com> wrote:
> Hi www-style,
>
> TL;DR VERSION:
> The part of the spec that says "<flex> can also transition to and from a
> <length> [...etc...]" isn't really useful right now -- I think it _wants_
> to (but doesn't currently) explain how transitions work between flex
> values that involve 'auto' as the preferred-size.
>
> NON-TL;DR VERSION:
>
> This part of the spec (which I'll call "Chunk A") is good:
>
> [[
> A ‘<flex>’ value is transitionable, by transitioning the preferred size,
> positive flexibility, and negative flexibility independently.
> ]]
>
> ... but the next sentence (which I'll call "Chunk B") needs tweaking, I think:
>
> [[
>  ‘<flex>’ can also transition to and from a <length>, by treating the
> length as if it were a flexible length with a positive and negative
> flexibility of zero and a preferred size of the length.
> ]]
>
> Taken literally, Chunk B _sounds_ like it's describing e.g. this transition:
>  "flex: 1 0 5px"  --->  "flex: 10px"
>        <flex>     --->         <length>
>
> However, in that context, the "flex: 10px" is _already_ a <flex> value
> (according to the definition in Section 7), and it already computes to "0
> 0 10px".  So there's no need to explain that here in a special case.  In
> other words, taken literally, Chunk B is superfluous.
>
> I don't think Chunk B wants to be taken literally, though -- I suspect it
> was _intended_ to describe transitions between e.g. flexible values and
> 'none'.  But note that none computes to '0 0 auto', and there's no reason
> to give that specific value special transition treatment -- if we're
> handling that, we might as well handle transitions between any flex value
> that uses 'auto'.

I believe this is a remnant from when <flex> was the flex() function.
It's definitely superfluous now, given that a bare <length> in 'flex'
computes to "0 0 <length>", as you point out.


> (Presumably, we'd handle 'auto' by substituting in the computed value of
> 'height' or 'width' for the "auto" value, as appropriate, and then
> continuing with the transition as described in Chunk A.  And of course if
> the given element isn't a child of a flexbox, then the meaning of an
> 'auto' preferred size is undefined, so we can't transition.)

Yes, 'auto' in 'flex' should transition iff the value for width/height
is transitionable.  Let's see if I can fix that up...

~TJ
Received on Friday, 6 April 2012 16:18:45 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:52 GMT