[css-align][css-flexbox] Should 'justify-content: stretch' compute to or behave like 'start'?

On 04/13/2015 12:55 PM, Javier Fernandez wrote:
> On 04/02/2015 04:51 AM, fantasai wrote:
>>
>> Wrt the initial value change; this is expected to be an invisible
>> change to authors.
>>
>> An implementation that implements both specs would use the Box
>> Alignment definitions, which supersede Flexbox's definitions.
>> (I'll clarify that in the module interactions section.)
>
> I understand your point, thanks for clarifying it in the document anyway.
>
> However, I found out another case that perhaps needs further explanations:
>
> * Flexbox:
>    - justify-content
>        + initial: flex-start
>    - align-content
>         + initial: stretch
>
> * Alignment
>    - justify-content
>        + initial: auto -> *Computes* to 'stretch' -> *Behaves* like
> 'flex-start'
>    - align-content
>        + initial: auto -> *Computes* to 'stretch' (Behaves as specified)
>
> So, I understand that we want both properties to have the same computed
> value (stretch) for the initial value ('auto') resolution, but we
> achieve the functionality described in Flexbox using different initial
> values with the *behavior* concept.

This is a good point. I think the question boils down to, should
'justify-content: stretch' compute to 'start' on flex containers,
or should it compute to itself but behave like 'start'? I'm not
sure what the right convention is here.

Note to WG: the Box Alignment spec introduces 'stretch' on the
'justify-content' property, which operates in the main axis.
Flexbox doesn't have this value because it doesn't make sense
for flexbox.

~fantasai

Received on Tuesday, 5 May 2015 21:37:29 UTC