Re: [css3-flexbox] flexbox algorithm and writing modes

On Wed, Apr 25, 2012 at 6:16 PM, Alex Mogilevsky <alexmog@microsoft.com> wrote:
> ± From: Tab Atkins Jr. [mailto:jackalmage@gmail.com]
> ± Sent: Wednesday, April 25, 2012 5:41 PM
> ±
> ± > I am not sure about 'start' as fallback. In old spec, 'box-
> ± align:baseline' on block-axis flexbox was treated as 'center', which
> ± may be just as random but should get a little better result (for
> ± example when mixing text and images?).
> ± >
> ± > Also 'center' as fallback avoids the issue of 'start' not always
> ± being same as 'before'.
> ±
> ± Oh, I didn't realize that the old spec fell back to 'center' when
> ± 'baseline' can't work.  Changing that was an oversight on my part,
> ± then.  I'll fix that.
>
> It may be just a bit more complicated. Or easier.
>
> We have a definition for baseline of a flexbox that always works, regardless of content or direction. Similarly, anything with "display-ouside:inline-block" has some kind of baseline, it will align with something when dropped in a line of text.
>
> Then the case of "baseline can't work" can't really happen, ever.
>
> Fallback of "treat 'baseline' as 'center'" made sense in old spec that didn't have per-item alignment and wasn't writing-mode friendly either. Now baseline alignment simply always works...

No, baseline still doesn't work if the item's baseline is
perpendicular to the main axis.  What can it align with?  This is
guaranteed to happen if you use a column flexbox and the items have
the same writing mode as the flexbox.  This is the *normal* case for
column flexboxes.  ^_^

~TJ

Received on Thursday, 26 April 2012 01:24:24 UTC