RE: [css3-box] padding: auto

[Lea Verou:]
> 
> I never argued it's trivial, 
Oh, I did not say you did.

>I argued it's useful. Useful enough to be worth dealing with difficulties 
>like the ones you point out.
Since you're proposing it I do hope you think it's useful enough to be 
specified :) I have no opinion, by the way. Just interested in what it 
means beyond the particular use-case that motivates the proposal.

> 
> Potential collisions with margin:auto are a non-issue, 
Collisions are rarely non-issues in CSS. Since all box models properties 
interact with each other, extending one of them requires thinking through 
these existing interactions. It may be a non-issue but that needs to be
demonstrated. It always helps when a proposal for a new feature tries to 
assess what impact it may/should have on what is already there. And I don't 
mean to say that's trivial work either; but when something as fundamental
as the box model is in play all input on how the new property or value should 
fit helps a lot. 

> it just needs a
> defined priority between the two (I'd suggest padding:auto, but it doesn't
> really matter).
That is one solution though I'm not 100% sure what priority means here; does it 
mean that if both margin:auto and padding:auto are specified then margin:auto 
resolves to 0 and padding:auto resolves what would have been margin space all 
around? If so then wouldn't setting padding:auto on a margin:auto element make 
other elements appear to move apart as margin space is no longer there to collapse?
Is that intentional?

> 
> Interaction with (min|max)-width is tougher, at least when box-sizing:
> border-box or padding-box is in effect. I can think of several reasonable
> things that could happen, but I'm not an implementor and it's one of those
> things where it's easy to suggest something with bad implementability.

You should start with what you think makes sense for authors and explain why
you think it's reasonable. This is definitely the right place to get 
implementability feedback; whatever is easiest to implement may not be what 
authors want...

> 
> I'm still waiting for someone to explain what padding:auto that's already
> defined in css3-box is supposed to do. Bert Bos is the editor, I guess he
> knows?

I don't know. But whomever requests a padding:auto value, the basic questions are 
the same...

Received on Thursday, 19 July 2012 03:34:42 UTC