Re: [css-align] Clarification on the concept of 'parent' for Self Alignment 'auto' value resolution

On Mon, May 30, 2016 at 5:27 AM, Javier Fernandez <[email protected]> wrote:
> Sorry if this question is seen as a pure implementation detail, but I'm
> finding some problems to properly implement 'auto' value resolution in
> some cases. Both the Flexbox and Box Alignment specification use the
> concept "element's parent" to indicate the Default Alignment
> value to be used for resolving 'auto' values of children's Self
> Alignment properties.
>
> >From the Flexible Box spec:
>
> "On absolutely positioned elements, a value of auto computes to itself.
> On all other elements, a value of auto for align-self computes to the
> value of align-items on the element’s parent, or stretch if the element
> has no parent."
> https://drafts.csswg.org/css-flexbox/#align-items-property
>
> >From the Box Alignment spec:
>
> "The auto keyword computes to the computed value of justify-items on the
> parent (minus any legacy keywords), or normal if the box has no parent."
> https://drafts.csswg.org/css-align/#align-self-property
>
> It may seems like obvious, but, I'd like a direct clarification of what
> "parent" means here. Is it the HTML element's parent or the "parent box"
> generated when elements are laid out ?
>
> My question comes from the fact that some layout models may add an
> anonymous box
> as parent for elements with 'auto' values in their Self Alignment
> properties. There won't
> be a direct CSS rule indicating the Default Alignment value these 'auto'
>  values will be resolved against. It'd depend on the values set, at
> layout time, to those anonymous
> parents.

Anonymous boxes are, as the name suggests, *boxes* - they show up in
the box tree, which is constructed *from* the element tree.  "Parent
element" refers to the element tree; it doesn't care about box-tree
structure.  (And so there's no layout-time determination; this
behavior can instead be resolved at inheritance time.)

~TJ